This document is a thesis that aims to introduce an architectural structure for an earthquake emergency decision support system for local governments in China. It discusses the goals, functions, working processes and structural framework of such a system. The thesis also analyzes necessary key databases and provides suggestions for implementation. The proposed decision support system would provide decision-makers with tools to more efficiently manage emergency response and recovery efforts based on the digitalization of earthquake-related information.
This document is a thesis that aims to introduce an architectural structure for an earthquake emergency decision support system for local governments in China. It discusses the goals, functions, working processes and structural framework of such a system. The thesis also analyzes necessary key databases and provides suggestions for implementation. The proposed decision support system would provide decision-makers with tools to more efficiently manage emergency response and recovery efforts based on the digitalization of earthquake-related information.
This document is a thesis that aims to introduce an architectural structure for an earthquake emergency decision support system for local governments in China. It discusses the goals, functions, working processes and structural framework of such a system. The thesis also analyzes necessary key databases and provides suggestions for implementation. The proposed decision support system would provide decision-makers with tools to more efficiently manage emergency response and recovery efforts based on the digitalization of earthquake-related information.
This document is a thesis that aims to introduce an architectural structure for an earthquake emergency decision support system for local governments in China. It discusses the goals, functions, working processes and structural framework of such a system. The thesis also analyzes necessary key databases and provides suggestions for implementation. The proposed decision support system would provide decision-makers with tools to more efficiently manage emergency response and recovery efforts based on the digitalization of earthquake-related information.
Developing a Decision Support System for Earthquake Emergency
Management
- An Exploratory Study Based on the Wen Chuan Earthquake in China
By: Wei Duan ID i6004999
Master of Science in Public Policy and Human Development Academic Year 2009 2010
Supervised by Mr. Henning Florian (PhD Fellow)
Maastricht, J uly. 2010
II
I
Acknowledgements
The contributions of many people have made this thesis possible. I would like to extend my appreciation especially to the following: First, I want to thank my supervisor, Mr. Henning Florian. He has given me guidance and advice throughout the research. It is highly appreciated that he industriously proofreads the drafts. Especially, as my English is not very good, reading my thesis drafts is a time-consuming work. Indeed, without his guidance and encouragement, I would not be able to put the topic together and finish this job so quickly. I would like to express sincere thanks to all key informants in Pengzhou. Without their help, I could not study this research topic and finish it on time. Equally important thanks are due to Mr. Peihu Wang, who helped me get in touch with most of the interviewers and collect the documents from the government. It has been a pleasure to study at the Maastricht Graduate School of Governance (MGSoG) of Maastricht University as a master student. I would like to thank all the teachers and fellow schoolmates for helping me in academic endeavors and spare time. There are also many good friends outside MGSoG who have given me their good company and support for this whole year. Many thanks to my Chinese friends: Yanlin Mou, Lei Chen, Qixing Yang and Yawei Zhang. Last but not least, my gratitude goes to my family and boyfriend for their unflagging love and care throughout my life. Without your support and encouragement, I could not have had this opportunity to study in Netherlands and finish this one year study successfully. II
Abstract
The paper aims to introduce an architectural structure of an earthquake emergency decision support system for the local government in China. Earthqakes, as major public incidents, can result in serious disasters, which requires immediate and effective government response. However, decisions about earthquake in cities, especially in some undeveloped ones, traditionally rely on the experience and ability of the decision-makers, which inevitably involves too many subjective factors in the process of emergency response. Therefore, it is crucial to strengthen the study on the general design of earthquake emergency management decision support systems so as to effectively support decision making and response efficiency during earthquake emergency management. In this study, we intend to illustrate the goals, functions, working processes, and structural framework of such a decision support system. In addition, we also discuss the necessary key databases, and finally, some suggestions on implementation are given. Based on the digitalization of earthquake related information, the decision support system provides decision-makers with the power to choose schemes to manage and direct the actions to be taken more efficiently in the process of immediate and long term recovery. III
Lists of Acronyms
AHP Analytic Hierarchy Process DSS Decision Support System EDP Electronic Data Processing EOC Emergency Operation Centers FEMIS Federal Emergency Management Information System GIS Geographic Information Systems GPS Global Position System KS Knowledge System LS Language system LAN Local Area Network MIS Management Information System NIMS National Incident Management System NRP National Response Plan PPS Problem-Processing System PS Presentation System RS Remote Sensing
IV
Table of Contents Acknowledgements................................................................................................................ I Abstract.................................................................................................................................II Lists of Acronyms................................................................................................................III Chapter 1 Introduction............................................................................................................1 1.1 Background...................................................................................................................1 1.2 Statement of the Problem..............................................................................................3 1.3 Objective of the study...................................................................................................4 1.4 Research questions........................................................................................................4 1.5 The significance of the study.........................................................................................4 1.6 Methodology.................................................................................................................5 1.6.1 Literature review.....................................................................................................5 1.6.2 Case study..............................................................................................................5 1.7 The organization of the paper........................................................................................8 Chapter 2 Theory review......................................................................................................10 2.1 Emergency management .............................................................................................10 2.1.1 Definition.............................................................................................................10 2.1.2 Process of emergency management.......................................................................10 2.1.3 Decision making in emergency management.........................................................12 2.2 Overview of DSS........................................................................................................13 2.2.1 Definition and characteristics................................................................................13 2.2.2 DSS and MIS........................................................................................................15 2.2.3 The generic architecture of DSS............................................................................17 2.2.4 The types of DSS..................................................................................................19 2.2.5 The application of DSS.........................................................................................21 Chapter 3 Issues of earthquake emergency management DSS...............................................22 3.1 Case introduction and current situation analysis..........................................................22 3.1.1 General introduction.............................................................................................22 3.1.2 Current situation of earthquake emergency management with DSS in China.........24 3.1.3 Earthquake emergency management deficiencies..................................................26 3.2 DSS goal and function analysis...................................................................................29 V
3.2.1 Decision support needs analysis............................................................................29 3.2.2 DSS goal analysis.................................................................................................31 3.2.3 DSS function analysis...........................................................................................32 Chapter 4 Design of earthquake emergency management DSS..............................................35 4.1 DSS business process and data flow analysis...............................................................35 4.1.1 Early warning analysis..........................................................................................35 4.1.2 Classification........................................................................................................38 4.1.3 Emergency plan management ...............................................................................42 4.1.4 Resource management..........................................................................................46 4.1.5 Decision information obtainment..........................................................................51 4.1.6 Decision making performance evaluation..............................................................52 4.2 General architecture of DSS........................................................................................53 4.2.1 Hardware support layer.........................................................................................56 4.2.2 Basic information layer.........................................................................................57 4.2.3 Problem process layer...........................................................................................58 4.2.4 User interface.......................................................................................................62 4.2.5 The relationship of the core components...............................................................63 4.3 Key data bases design.................................................................................................65 4.3.1 Data base design...................................................................................................65 4.3.2 Model base design................................................................................................67 Chapter 5 Implementation....................................................................................................71 5.1 Potential problem........................................................................................................71 5.1.1 From interviews....................................................................................................71 5.1.2 From literature......................................................................................................72 5.2 Suggestion on implementation.....................................................................................73 5.2.1 Organizational environments................................................................................73 5.2.2 People characteristics............................................................................................76 5.2.3 Technical supports................................................................................................76 Chapter 6 Conclusions..........................................................................................................78 6.1 Summary.....................................................................................................................78 6.2 Limitations of study....................................................................................................79 6.3 Policy Recommendations............................................................................................80 6.4 Further research..........................................................................................................81 VI
Chapter 1 Introduction 1.1 Background At 2:28 p.m. on May 12, 2008, the 8.0 magnitude 1 Wen Chuan earthquake in China shocked the world. After the earthquake, governments at all levels took emergency responses aimed at reducing the losses. However, by J uly 23, the official death toll had reached 69,181 with 374,171 people injured, more than 18,498 persons missing 2 and direct economic losses of around 845.1 billion Yuan 3 . It is estimated that there are around 1,300,000 earthquakes each year in the world, among which about 130,000 can be felt by people, and only about 134 of which will cause societal losses 4 . In 1976, the 7.8 magnitude Tangshan earthquake brought about 243,000 deaths, and another earthquake which happened in 1920, in Gansu, China resulted in 234,000 deaths as well 5 . J apan and other developed countries have adopted some effective countermeasures, such as Geographic Information Systems 6 (GIS) and life detection instrument etc., reducing the post-earthquake damage. Thus, for example, the 7.2 magnitude earthquake in Iwate Prefecture J apan in J une 2008 caused only ten people dead and hundreds injured 7 . Large earthquakes (above magnitude 6) are listed in the table 1-1 below. In line with the Comprehensive Emergency Management Framework (Drabek & Hoetmer, 1990), disasters mitigation is an assignment usually with phases like preparation, response and recuperation; therefore, it is important to pay attention that adequate actions should be implemented before, during and after the earthquake. However, it is impossible to wholly reduce the risk of damage. Actually, most developing countries and even many developed countries are far away from relative safety (Morrow, 1999). Local government is not always capable of executing actions to relieve the impact of earthquake. In fact, the trend in many
1 All the magnitudes mentioned in this study are based on the Richter Scale. 2 The Central Peoples Government of the Peoples Republic of China (2008). Casualties of the Wenchuan Earthquake. Retrieved online on 1 August, 2010 from http://www.gov.cn/ldhd/2008-06/24/content_1026042.htm. 3 The Chinese Academy of Social Sciences (2008). Direct Economic Loss of the Wenchuan Earthquake. Retrieved online on 1 August, 2010 from http://www.cass.net.cn/file/20080905196570.html. 4 Infoplease (2007). Frequency of Earthquakes Worldwide. Retrieved online on 1 August, 2010 from http://www.infoplease.com/ipa/A0197837.html. 5 People (2003). The Devastating Earthquakes in the World. Retrieved online on 1 August, 2010 from http://www.people.com.cn/GB/guoji/1029/2274478.html. 6 GIS is any system that captures, stores, analyzes, manages, and presents data that are linked to location. 7 Sina (2008). Iwate earthquake. Retrieved online on 1 August, 2010 from http://news.sina.com.cn/o/2008-06-15/195814021090s.shtml.
2
countries is to deal with the disaster after it happens (Aleskerov, Iseri Say, Toker, Akin, & Altay, 2005). Efficient emergency management is, therefore, necessary to be taken into account in order to develop practices that complement prevention and complete disaster mitigation policies.
Table 1-1 Statistics of large earthquakes in the worldwide since 20th century 8
Time Place Magnitude Loss of Life 27-2-2010 Concepcion, Chile 8.8 Around 800 12-1-2010 Port-au-Prince, Haiti 7.0 Between 217,000 and 230,000 people 30-9-2009 Samoa 8.0 Triggered the tsunami, at least 113 6-4-2009 Italy 6.3 At least 278 people 14-6-2009 Northeast of Japan 7.2 6 2-9-2009 West J ava, Indonesia 7.3 At least 70 people 12-5-1008 Wenchuan, Sichuan, China 8.0 more than 80,000 people 10-2005 South Asian Subcontinent 7.8 86,000 11-2004 Offshore northwest Sumatra, Indonesia 8.7 Triggered the tsunami, nearly 300,000 deaths 08-1999 Northwestern Turkey 7.4 20,000 01-1995 Kobe, Japan 7.3 5,400 06-1990 Northwest Iran 7.7 40,000 12-1988 Soviet Armenia Spiro Tucker 7.0 25,000 09-1985 Mexico 8.1 5,000 09-1978 Tabas, Iran 7.4 25,000 07-1976 Tangshan, China 7.8 242,000 02-1976 Guatemala 7.5 26,000 05-1970 Pacific coast of Peru 7.8 66,800 01-1970 Yunnan, China 7.7 10,000 12-1939 Erzincan, Turkey 8 Around 40,000 01-1939 Chile 8 28,000 05-1935 Quetta, India (now Pakistan) 7.5 50,000 12-1932 Gansu, China 7.6 70,000 09-1923 Kanto, J apan 8.3 At least 140,000 12-1920 Gansu, China 8.5 At least 100,000 01-1915 Aweichanuo, Italy 7.5 29,980 12-1908 Messina, Sicily, Italy 7.5 83,000 08-1906 Valparaiso Chile 8.6 20,000 04-1905 Kashmir border, India 8.6 18,800
8 Data collected from National Earthquake Response Support Service (2010). Earthquake Information. Retrieved online on 1 August, 2010 from http://www.nerss.cn/list.php?fid=40.
3
Earthquake management and mitigation is not an easy task and usually requires a large amount of analysis and resources. In order to reduce the damage and optimize the overall performance of government, the responsible leaders have to make decisions quickly and effectively. Currently, however, decisions are usually made based on personal judgment of leaders or experts, rather than basing their decisions in the different stages of relief efforts on factual evidence. In addition, the information at the first moment just after disaster occurs tends to be confused, imprecise and incomplete. Besides the lack of accurate information, the situation is further complicated by the pressing time. At an emergency moment, time is life and decision makers need to make decision quickly. Such complexity suggests the introduction of a Decision Support System (DSS) a class of information systems that support business and organizational decision-making activities. This system could be used to help analyze the disaster and assist the leaders in making sensible decisions by means of an inference tool capable of offering an assessment of the consequences of almost every combination of adverse phenomena, based on the available information in the limited time. 1.2 Statement of the Problem Presently, the major support the government gains electronically is still largely confined to information systems which do not give active decision support. Natural disasters always have the characteristics such as an ad-hoc nature, unpredictability and destructiveness, and how to make the right decision in a limited time becomes important to the leaders. The DSS would be a powerful tool for the decision makers, as it can help to save time and enhance the effectiveness of decisions, which means that the emergency management would be more efficient and more lives will be saved in an earthquake. This study is mainly focused on the DSS of local government in China. The reasons for choosing local government are listed below: Firstly, earthquakes frequently occur in remote areas, where administrative capability is relatively low, and, applying the DSS can help to enhance governance capacity during the earthquake. Secondly, in China, most of the local governments do not even have a DSS for daily management, let alone for the emergency situation; thus, the need of the DSS is large. Last but not the least, there are 2,862 local governments in China. If the overall capability of the local government is improved, then the Chinese governments emergency response ability will be improved more generally. 4
1.3 Objective of the study The general goal of this research is to formulate an architecture for a computer-based DSS framework which can be used to aid local governments in their decision making process prior to taking any action in earthquake emergency management. Due to the limited scope of this study, it explicitly restricts itself to developing an architecture blueprint for the DSS, rather than engaging in the details of its construction, such as programming. The specific objectives are: To build an architecture of DSS To identify the information processing mechanisms of each module To conceive of potential problems that may occur during the implementation of the DSS and give solutions to avoid them 1.4 Research questions In order to fulfill the main objective of this thesis, the main research question is: What is an ideal architecture for a DSS on earthquake emergency management in China, and how should it be implemented? Furthermore, some sub-questions are posed as follows: - What is the general structure of the DSS? - What is the function of each module in the DSS? - What is the relationship between each module? - What is the information data flow of each module? - What problems might be encountered when implementing the system? - How can potential problems in implementation be solved? 1.5 The significance of the study Referring to historical data and table 1-1, we can see that there is a big difference in life and economic losses between China and other countries. Besides the diversity of earthquakes, population distribution, urban planning and architectural design, the emergency response after the earthquake is another main cause of this difference. Compared with J apan and other developed countries, China's emergency response has shown a certain shortage in terms of technical expertise and management 9 .
9 National earthquake response support service (2010). Characteristics of emergency management in J apan. Retrieved online on 1 August, 2010 from http://www.nerss.cn/bencandy.php?fid=60&id=724
5
A well designed DSS will help decision makers compile useful information from a combination of raw data, documents, which will help to identify and solve relevant problems during the earthquake, make scientific evidence-based decisions and quicken the response of the government. For public emergency management, this study could provide some theoretical foundations for earthquake DSS construction, make linkages between earthquake emergency management and other natural disaster management areas such as hurricanes or snow storms, and provide practical and active support to decision makers. 1.6 Methodology 1.6.1 Literature review The study of earthquake emergency management is characterized by its interdisciplinary nature incorporating elements from emergency management, decision making management, public policy, and the relevant theory of DSS. In order to gain an appreciation of the theoretical base of the research and to understand the latest research trends on earthquake DSS, a great deal of literatures was reviewed (see chapter 2). The general search strategy in this study is by using emergency management and DSS as the main subject areas for the literature review, mainly via internet and library sources, referring to the Chinese and international online information, journals and books. Because of the time limit, once the basic theories were understood and then could be applied to assist the research on DSS architecture construction, it is considered that the search for literatures was relatively complete for this study. 1.6.2 Case study The research is based on an in-depth case study of Wen Chuan earthquake. The main reasons why it is believed that a qualitative research strategy (case study) is appropriate for the research question are: firstly, it is difficult to analyze all the decision making demands of all local governments in China. In addition, this research is an exploratory study, choosing one case to build an initial DSS framework for local government can achieve our research objective. Reasons for choosing Wen Chuan earthquake were, firstly, the 8.0 magnitude Wen Chuan earthquake was the most destructive earthquake since the establishment of New China (1949), its intensity was larger even than the Tangshan earthquake in 1976 (in which 243,000 people were killed); secondly, it is a recent, internationally known disaster in China, which 6
can serve as an example for applying modern technology in the country; in addition, the author comes from the earthquake area, and has the relevant knowledge of this local context; finally, the local government has interest to participate in this research, which improves the data collection process and relevance of research findings. Therefore, relevant first-hand data can be accessed and the interview to relevant decision makers could also be conducted. The research data presented here comes primarily from the following: 1.6.2.1 Documents Analysis At the beginning of the study, various types of documents are collected, such as, Briefings of Earthquake, Earthquake Relief Bulletins, Donate Material Management and Distribution Method, Pengzhou Public Emergency Response Plan, Pengzhou Earthquake Emergency Plan, and some related notice. The main document sources are earthquake relief command headquarter, secretary office and emergency management office in Pengzhou government Related data were collected in order to understand the working organization structure of earthquake emergency team, the emergency working process, the contents of existing emergency plans and the information system construction status. On the other hand, searching data from internet is a helpful method for the analysis as well. Lots of practical emergency data needed in the research came from the national governments official websites. To sum up, the relevant documents from inside-and-outside of the government can help to understand the current earthquake emergency management situation of China. These document and internet resources shed light on elements of the management process, such as, the organization structure of emergency management team, the emergency working process, and the general demand of the decision maker. 1.6.2.2 Semi-structured interview (1) General introduction of interview In order to develop the in-depth case study of Wen Chuan earthquake, a qualitative approach, semi-structured interviews, was taken (Flick, 2009). In the interviews, some open-ended questions were asked to build an understanding of the current working process of earthquake emergency management, how the commander makes decisions, what they would require from the DSS, and some relevant issues on implementation. In brief, the final purpose of these interviews was to support the further analysis of our research questions: what is the architecture of DSS for emergency management and how can this system be implemented. 7
(2) The key informants The key informants selected were the commanders of the Wen Chuan earthquake response forces in the local government of the city of Pengzhou (refer to chapter 3.1.1). The structure of the relief command headquarter 10 (Committee, 2008) is showed in Figure 1-1 below. There are in total eleven sub-groups and thirty two leaders in Pengzhou relief command headquarter. We chose the three top leaders of the relief command headquarter and the leaders of the eleven sub-groups to be our key informants. Each leader is taken to represent a sub-group. For example, the leader from relief supply group can offer information on material allocation decision making.
Figure 1-1 Structure of the relief command headquarter of Pengzhou
(3) Preparedness of interview In order to prevent unintended mistakes occurring during the interview with key informants, the following preparatory steps were taken. Firstly, contact with the key interviewees by email or telephone was made to discuss with them the purpose and relevant issues about the interview (refer to Annex 1) and to make a proper appointment with them. Secondly, be familiar with the question guide (see Annex 2) and the C.V. of each informant before the interview. Third, a tape recorder was prepared to record the interviews in order to get a better analysis afterward. Fourth, the correct location and time (e.g. single quite room, consider the time lag between Netherlands and China) to do the interview was chosen to avoid disturbance and inconvenient. Finally, to make the interview more effective I attempted as much as
10 All the names of relief command headquarter are translated by author 8
possible to focus the discussion specifically on the emergency management during the earthquake in 2008. When conducting the interview, to start, I introduced myself and explained again the purpose of interview. Then, I asked for the permission of recording the interview, and whether the interviewee had any further question before starting the interview. During the interviewing process, I ensured that all the questions were well organized and asked in logic way going through the interview questions topic by topic, and according to the response of the informant, adjusting our interview questions. After finishing the interview, I expressed our appreciation for cooperating with this study. (4) Results of interview Due to the constraints of time and distance, all of the interviews were conducted via telephone. As planned, fourteen commanders were all interviewed: three top leaders of the relief command headquarters, seven sub-group leaders and four vice leaders of the sub-group (see Annex 3). All of the respondents had had experiences with decision-making during the earthquake emergency management. The interviews were semi-structured and lasted between 30 and 80 minutes, with an average of 50 minutes per person. The main topics of the interview encompass questions about Command and control tasks, Information Sources, Demand and Implementation. The full description of the interview is provided in Annex 2. A transcript of each interview was made and utilized for the analysis in the following chapters. 1.7 The organization of the paper In order to demonstrate the research process on DSS, my thesis includes the following sections: Chapter 1 of this paper is a summary of the subject background, research questions, objective, significance, research methods and the general structure. Chapter 2 reviews theoretical literatures written by researchers regarding underlying concepts of emergency management and the overview of DSS. Chapter 3 illustrates the issues of earthquake emergency management DSS. First, based on the analysis of decision making function of Beijing emergency management information system combined with the interviews, this paper discusses the problems existing in current earthquake emergency management, and then it clarifies the decision support needs, goals of the proposed DSS, and concludes six sorts of functions need to be achieved. 9
Chapter 4 discusses the general framework of DSS. This chapter firstly analyzes the general working process and data process of the six sorts of functions in DSS; secondly, this chapter establishes a four-layer general framework of the DSS, discusses each layer in depth and the interaction between different layers; finally, the chapter describes the required databases and model base of DSS. Chapter 5 analyzes potential problems and offers implementation suggestions. The suggestions on implementation consist of three categories, that is, organizational environments, people characteristics and technical supports. Chapter 6 provides a conclusion of this study. This chapter firstly summarizes the research findings; secondly, state out the limitation of this thesis; thirdly, emphasize the concluded policy recommendations; finally, elaborates a research outlook of this topic. The figure below shows the general analysis logic of my thesis.
Figure 1-2 Analysis logic of the paper 10
Chapter 2 Theory review 2.1 Emergency management 2.1.1 Definition In order to study the discipline of Emergency Management (EM), comprehending the definition is vitally important. A simple definition of emergency management is to deal with all risk and risk avoidance. Since risk covers extensive issues, situations and participants, EM is involved in everyones daily lives and should be integrated into daily decisions and not just called on during times of disasters (Haddow, Bullock, & Coppola, 2007). The United States Federal Emergency Management Agency (FEMA) elaborates that emergency management means organized analysis, planning, decision-making, and assignment of available resources to mitigate, prepare for, respond to, and recover from the effects of all hazards and its main aims are to prevent injuries, save lives, and protect property and the environment. 11
2.1.2 Process of emergency management According to a study carried out by National Governors Association in 1970s, the process of emergency management can be generalized into a four-phase model: mitigation, preparedness, response and recovery (Carroll, 2001). The first two phases are considered proactive actions that are taken prior to an emergency while the final two are both reactive actions that are taken after it has occurred. Mitigation is a sustained action to reduce or eliminate risk to people and property from hazards and their effects. It involves the identification of hazard, evaluation of its frequency and severity, estimation of direct and indirect economic and social costs, determination of tolerable risk level, and the identification of risk-reduction opportunities. There are some widely accepted tools for mitigation to reduce risks: hazard identification and mapping, design and construction applications, land-use planning, financial incentives, insurance, and structural controls (Haddow, et al., 2007, pp. 75-77). Preparedness refers to the actual planning, training, placement of resources, mutual-aid agreements across jurisdictions, and other coordination efforts before an emergency strikes.
11 Federal Emergency Management Agency (2010), Fundamentals of Emergency Management. Retrieved online on 1 August, 2010 from http://training.fema.gov/EMIWeb/IS/is230a.asp
11
As a building block of emergency management, preparedness aspires to create a state of readiness to respond to a disaster, crisis or any other type of emergency situation. It operates in a cycle starting from assessment of threats and vulnerability, moving through implementation of enhancements made through identifying shortfalls, and ending with exercising and training for the revamped preparedness (Haddow, et al., 2007, pp. 75-77). Response is the actual activation of emergency plans to meet an emergency. Emergency plan in our research refers to the feasible action guidance which is used to solve the problems raised by earthquakes and guarantee the realization of emergency decision making goal to reduce the deaths and minimize losses on property and environment. After analyzing the current earthquake emergency plans both in Pengzhou and central government of China, the main contents included in the emergency plans are: (1) introduction, including: the purpose, working principles, compilation basis and application scope of the emergency plan; (2) definite the responsibilities and working process of emergency management participating groups or departments; (3) early warning and preventive mechanism, including: information monitoring and reporting, early warning preventive action guide and the early warning level classification; (4) emergency response, including: information sharing and processing rules, communication protection, command and coordination among groups, security protection of victims and emergency workers, impact assessment and resource management; (5) post-disposition, including: reconstruction, social assistance and evaluation. More often than not, the emergency plan is conducted with a system of procedures and protocols. When a disastrous event occurs, the first responders such as firemen, police officials, medical personnel and troops will arrive at the scene with the coordination and direction of emergency managers and government officials. Resources will be committed to the response mission on the basis of emergency management protocols made prior to the event (Haddow, et al., 2007, pp. 75-77). Recovery consists of the actions taken by government to restore order and vital systems in the disaster-impacted zone and to provide assistance by way of temporary housing, food, or medical attention. It often commences a few hours or days after a disaster and can last for months or years. Recovery function encompasses a variety of participants such as government at all levels, individuals and the business community in making decisions about how to conduct the mission so as to restore normal life and ensure a safer future for the zone (Haddow, et al., 2007, pp. 75-77). 12
2.1.3 Decision making in emergency management In this part of the literature review, the main objective is to focus on emergency decision making and to get a general idea of what kinds of information are relevant to design the theoretical architecture of the DSS including how to design our interview questions, which factors will affect the decision making and which of them should be considered in our DSS framework. Decision making and decision theories are popular research topics in almost every discipline from business, social sciences, military, engineering etc. However, the literature about emergency decision making is relatively scarce and is mainly focused on fire emergency. According to Dai Wang & Yang (1994) emergency decision making is defined as addressing some unexpected and consequence-serious problems within a very limited time. Compared with traditional decision making, and building on the definition by Dai Wang & Yang (1994), emergency decision making is characterized by the following factors: The environment of decision making is continuously changing according to the different stage of disaster. Uncertainty is the main characteristic of emergency decision making: the emergencys nature, scale, form and when and where it will take place is all unsure. Decision makers should give quick response, which means they are usually under the pressure of limited time. Information shortage. The information needed for decision making may be inexact or incomplete. There is less normative standard or rule for decision makers to rely on in emergency management. The only normative rules available for them are tactical doctrines of emergency management manuals and guidelines. Danielsson and Ohlsson (1999) interviewed 90 fire chiefs about their own accounts of factors affecting the difficulty level of decisions and described types of especially difficult decisions concerning giving priorities in lifesaving operations and evacuations. This gives us some general clues about what factors can affect the decision making process and helps to construct the interview outline of my research. Sinha (2005) investigated and interviewed 20 firemen to analyze the effect of decision makers experience on the result of decision. One main conclusion is that experience is one of the most influential factors affecting emergency decision making. Education and training are 13
the other two important factors. This research is also useful in shaping our ideas about how to advance the implementation progress of DSS. Donner (2008) used the POET (Population, Organization, Environment and Technology) Model in human ecology to discuss the influence that an association to which decision makers belong, will have upon the individual decision maker during the process of emergency management. Based on this research, we decide that in our simple initial DSS architecture, the association factor would not be included at the first step. In order to adapt to technological development and meet the needs of emergency management, Hernandez and Serrano (2001) demonstrated an application of advanced knowledge models to assist emergency management as an adequate response and described a generic architecture including the knowledge required to deal with emergencies in different problem scenarios. In addition, simulation models are also integrated as part of the knowledge architecture. This research gives a general image of what a simple DSS is as well as its architecture. 2.2 Overview of DSS Before going into the detail of DSS, the concept of information systems (IS) should be introduced. According to Laudon & Laudon (1995), DSS is a sub-category of the IS discipline, and thus, a DSS is actually an information system. Thus, knowledge about what IS is can help to better understand DSS. According to Silver, Markus, & Beath (1995, p. 356), IS are the inclusive super systems, whereas people and procedures-the stuff of business processes-and tools-such as computers and programs-are the subordinated subsystems. This definition is from the view-point of organization, and the main purpose of the system is to improve operational efficiency and effectiveness. DSS and IS share many common characteristics, such as purpose (improve the working efficiency) and approach (computer-based). The biggest difference between the two systems is that DSS is an advanced IS, which is more focused on supporting and improving managerial decision-making. The details of DSS are discussed in the following section. 2.2.1 Definition and characteristics The concept of DSS was first introduced by Morton (1971) in his book Management Decision Systems. This text presents DSS as an intelligence human-computer interactive system, which is based on management science, operation research and behavioral science, applied through the means of computer technology, simulation and information technology to 14
assist the decision making of unstructured or semi-structured decision problems. 12 Sprague and Carlson (1982) also hold the same understanding of DSS. They, too, emphasize that DSS is helping decision makers to solve semi-structured or unstructured problems by taking advantage of data and models. In the past decades, many scholars have given their definitions on the concept of DSS; here three well known ones are listed to provide a general understanding of DSS. Mittra (1986) suggests that DSS are systems using database and mathematical models to help users generate information they need and assist decision-making. In addition, from his point of view, DSS should have the following functions: In order to make decision, users can try What-If functions (a simulation capability of DSS to answer what-if questions) and according to the result of what if questions to test which is a better decision. DSS is constructed by a database management system, a group of mathematical models and a human-computer interactive interface. DSS uses one control module to combine three other modules to answer the decision question, namely, data access module, data transformation module (to retrieve data, generate graphs and reports) and modeling module ( to save mathematical models and use simulation techniques). Keen (1987) asserts that DSS is the combination of decision, support and system, which means that through the development of computer technology on system construction and the gradual expansion of the function of computer support a better decision can be achieved. The support here emphasizes exploring alternatives proactively or even making decision on all alternatives. Burstein & Holsapple (2008) define DSS in terms of the roles it plays in decision processes. They argue that DSS can provide knowledge and knowledge processing capability that is instrumental in making decisions or making sense of decision situations. It can enhance the outcome and efficiency of decision making.
12 A structured problem is when the rules or environment of one problem can be clearly explained by language (mathematical or logic, quantitative or reasoning). The data, process and evaluation of the problem are all determined. If problems cannot be described clearly and can only be judged by experience or intuition, this is called an unstructured problem. In the middle of the continuum are semi-structured decisions. 15
So far, there still is no generally accepted definition of DSS. All the technologies which help to achieve the goal of decision support can be used to construct DSS. The structure of DSS may be totally different because of the different times, different application purpose and different technologies which we use to construct DSS. However, one thing that is common to all DSS is that: DSS must be able to play the role of decision support. Therefore, grasping the basic characteristics of DSS has very important significance. The basic characteristics of DSS can be summarized as follows: DSS is a subcategory of IS. It is a computer-based system. The basic components of DSS include databases, model bases, data processing technique, and information presentation system. DSS is designed for decision makers. The input, output, origin needs and objectives of DSS are all from the decision makers. It has a human-computer communication interface for non-computer professionals to easily operate. It has an emphasis on decision support, but it does not replace the decision-makers in making decision. The payoff is in extending the range and capability of managers decision making processes to help them improve the effectiveness of their decisions. It is user oriented. Because the decision-making process is dynamic, decisions should be made dynamically according to different environment, different users requirements, the understanding of users decision question and the available knowledge. 2.2.2 DSS and MIS Over the years, there have been disputes regarding DSS and Management Information Systems (MIS). Some scholars think that MIS includes DSS, while others suggest that MIS is a sub-concept of DSS. Still others suggest that DSS is a separate category all together. The reason why people cannot distinguish the two systems is because both are sub categories of IS and can help improve the efficiency and effectiveness of an organization. In addition, they can, to some extent, assist in making decisions as well. Therefore, it is necessary and important to understand the difference between the two systems. 2.2.2.1 The definition of MIS According to O'Brien and Marakas (2005), Management Information System (MIS) is regarded as a subset of the overall internal controls procedures in a business, which covers the 16
application of people, documents, technologies, and procedures used by staff members to solve business problems. Academically, the term is commonly used to refer to the group of information management methods tied to the automation or support of human decision making. The functions of MIS are mainly as follows: Provide timely and comprehensive information and data to help achieve the goal of a given decision. Prepare and provide a unified format for the information and data and simplify all statistical work. Use specified mathematical models to analyze the data, and then predict the future based on past situations. According to the requirements, give different reports to different level of managers, with the prospect of helping them to analyze and interpret the reports as soon as possible to make the decision more efficient. 2.2.2.2 The difference between DSS and MIS The relationship between DSS and MIS is an ongoing question of debate. In this study, DSS and MIS are considered related, but not the same. Based on the study of Sprague (1980), the major differences between the two concepts are summarized below: Different user DSS is mostly designed for senior staff to assist them in making decisions whereas MIS is generally used by middle class managers or staff members to serve their daily operations. Different problems to deal with DSS supports semi-structured decision-making. Such decisions are complex, and they cannot always be accurately described. In addition, a large amount of calculations are needed. It is necessary to apply the computer and users participation to achieve satisfactory results. On the other hand, MIS employs back structured decision-making, and such decisions are known, predictable, and often repetitive. Different system structure 17
DSS is always a model driven system and has a model base. DSS is constructed through a combination of multiple models to support decision-making, and its analysis is focus on the needs of policy makers. Meanwhile, MIS is generally a data driven system and has a data base. MIS is an integrated multiple-transaction EDP (Electronic Data Processing) based on the transaction functions (production, marketing, personnel etc.). Its analysis focuses on the overall information needs of the system, and the pattern of output report is fixed. Different pursuit In general, DSS is used to assess validity, that is, whether decisions are effective. However, the pursuit of MIS is efficiency, which needs quick inquiry as well as quick statistical results and reports. In conclusion, DSS is user-oriented and emphasizes support for decision makers from external environmental information, internal comprehensive information and personal experiences etc. However, MIS focuses on the integrity of the information flow inside of the management system, to provide all staff the information needed. MIS cannot meet the requirement of decision makers to assist them in decision making, as it cannot provide the external information or adapt to personal decision making style, which are important elements of decision making. 2.2.3 The generic architecture of DSS This section is the theoretical basis of chapter 4.2. According to Burstein & Holsapple (2008), the generic architecture of DSS consists of four essential parts: Language system (LS), Presentation System (PS), Knowledge System (KS) and Problem-Processing System (PPS). The first three are representational systems, which are used by the PPS (the active component of DSS). The relationship among the four elements can be illustrated by the graph below: 18
Figure 2-1 General architecture for DSS (adapted from Burstein & Holsapple, 2008)
User Theoretically, the user of a DSS could only be the decision maker. However, administrators, DSS developers, data entry persons and facilitators can also be considered as users. LS The LS consists of all information the DSS can accept, including the language for users or models to retrieve data and the language by which users can operate the computer. Decision makers use the LS to describe their problems as the first step to get the final decision. The main function of this system is to accept knowledge, clarify knowledge and recall knowledge. PS The PS consists of all messages the DSS can emit. It consists of two parts: providing knowledge and seeking knowledge. According to the order, the PS first seeks the knowledge it needs to emit. By choosing a PS element to present, it provides users responses that PPS emits. KS The KS is constructed by all knowledge the DSS has stored and retained, including descriptive knowledge, procedural knowledge or reasoning knowledge. All the knowledge is subject to use by PPS. PPS 19
The PSS is the core of DSS, which tries to recognize and solve problems during the decision making process. There are two types of abilities in the PPS, first order ability and second order ability. The first order ability is mainly constructed by five knowledge manipulation abilities, namely, knowledge acquisition, knowledge assimilation, knowledge selection, knowledge generation and knowledge emission. These abilities are primary, front line abilities and contribute to the outcome of a particular decision episode. The second order ability includes the function of coordination, control and measurement, which are concerned with governance of the first order ability within and across decision episodes. The PSS is asked to process a LS element, and this processing requires the PPS to select some portion of the KS contents (first order ability) and then apply second order ability to change the knowledge held by coordination, control and measurement. In other words, some portion of the PSS are covert (strictly internal, yielding assimilations of knowledge into the KS) and others are overt (witnessed by the user via PPS knowledge emission of PS elements). This generic DSS architecture only provides the fundamental common parts of DSS. To fully appreciate any specific DSS, we should know the particular requests making up its LS, the particular responses making up its PS, the particular knowledge representation existing in its KS, and the particular knowledge processing capabilities of its PPS. 2.2.4 The types of DSS Although we learned the common base and some fundamental terms from the generic architecture of DSS, it is necessary to distinguish among different DSS. In the research by Burstein & Holsapple (2008)
, eight types of DSS are introduced. For the purposes of the current study, these can be summarized into four types which provide the overall picture of the development of DSS. 2.2.4.1 Text-oriented DSS In a text-oriented DSS, the LS contains requests concerned with the allowed manipulations as well as user requests for assistance from the DSS. The PS consists of images of stored text that can be emitted. In addition, the KS of text-oriented DSS is made up of electronic documents; each message in the KS is potentially interesting to the decision maker. Finally, the PPS is constructed by some software that can perform various manipulations on contents of any of the stored documents. With a text-oriented DSS, people can easily keep electronic notes, supporting decision makers by electronically keeping track of textually represented knowledge. 20
2.2.4.2 Database-oriented DSS The main differences between text-oriented DSS and database-oriented DSS focus on KS and PPS. In database-oriented DSS, all files that make up of the KS hold information about table structures plus the actual data value contents of each table. The PPS consist of three kinds of software: a database control system, an interactive query processing system and various custom-build processing systems. The database control system consists of capabilities for manipulating table structures and contents (e.g., defining or revising table structures, finding or updating records, and building new tables from existing ones). The query processing system can respond to certain standard types of requests for data retrieval, and the custom-built processing system is developed to meet the specific needs of the DSS user on marketing, production, financial or other application. In all, the data handled by database-oriented DSS tend to be primarily descriptive, rigidly structured and extremely voluminous. 2.2.4.3 Rule-oriented DSS Rule-oriented DSS is constructed by a knowledge management technique that involves representing and processing rules. The logic of decision making is if the premise is true, then the conclusion is valid. Users of rule-oriented DSS can issue requests, and then LS contains requests for advice and requests for explanation. Correspondingly, the PS includes messages presenting advice and explanations. In addition, the KS holds one or more rule sets which can pertain to reasoning about what recommendation or explanation to give a user regarding a particular subject. Finally the PSS has capabilities for creating, revising and deleting state descriptions, and it also has the ability to explain its behavior both during and after conducting the inference. 2.2.4.4 Compound DSS Compound DSS frameworks tend to combine more than one knowledge-management techniques (text, database, or rule management) together to support the decision maker. There are two ways to make compound DSS: Multiple DSSs, each oriented toward a particular technique In this framework, because each DSS has its own LS and PS, multiple staff assistants are needed to operate each system and then decision maker has to translate responses from one DSS into requests to another DSS in order to get the final decision. Single DSS, encompassing multiple techniques 21
This option only needs a staff assistant who is good at multiple knowledge management techniques as there are one LS and one PS for decision maker to learn. The effort required of a decision maker who wants to use results of one technique in the processing of another technique varies depending on the way in which the multiple techniques have been integrated into a single compound DSS. 2.2.5 The application of DSS Based on a detailed literature survey of 1,020 DSS articles published from 1990 to 2003, Arnott & Pervan (2005) revealed that the main application of DSS generally focused on three approaches: analysis based on quantitative models, assessment depending on large data bases and supporting group decision making. Lots of model-oriented DSS research stresses application of individual systems, such as personal DSS; on the other hand, institutional, organizational or ad hoc DSS are mostly data-base oriented DSS. Group DSS always stressed effects on decision process structuring. Eom and some other scholars studied 674 articles on DSS between the year 1971 and 2001 to show the development of DSS application (H. Eom & Lee, 1990; S. Eom & Kim, 2006; SB Eom, Lee, Kim, & Somarajan, 1998). The recent survey suggests that governments have the most DSS applications in the non-corporate area. These government applications are distributed among data-processing acquisitions (Barba-Romero, 2001), online crisis management (Mak et al, 1999), managing the DOE 13 hazardous waste cleanup program, US bridge networks improvement, medium-term economic planning, Chinas coal and electricity planning, railway infrastructure scenarios development, and random sampling-based testing of urine specimens instead of testing all collected specimens (Chari et al, 1998). (S. Eom & Kim, 2006, p. 5)
13 DOE refers to Department of Energy 22
Chapter 3 Issues of earthquake emergency management DSS 3.1 Case introduction and current situation analysis The purpose of this chapter is to identify the main functions a DSS should serve in order to address a Chinese local governments decision support needs in earthquake emergency management. This will be done on the basis of an analysis of the current use of IS for emergency management in Beijing, and a qualitative case study of earthquake emergency management in the city of Pengzhou. In chapter 3.1, the general information of Pengzhou is introduced, and then Beijings emergency management MIS is given as an example to illustrate the current situation of earthquake emergency management with DSS in China. Its main functions for decision support are analyzed, and then the current situation of earthquake emergency management with DSS in China is discussed. Finally, based on information from relevant documents and the current situation analysis, some deficiencies of current earthquake emergency management are discussed. 3.1.1 General introduction Pengzhou, a county-level city in Sichuan province of China, is about 19 kilometers northwest of Chengdu. 14 It had an area of 1,420 square kilometers, 20 towns and a population of 803,400 in 2009. 15 Like other regions around the area, Pengzhou was badly damaged by the 2008 Wen Chuan Earthquake. The epicenter was only 45 kilometers away from the center town of Pengzhou (Tianpeng). Some towns near the mountainous area, such as Longmenshan and Xiaoyudong experienced heavy damage, because the seismic intensity in those areas reached a magnitude of 11, the maximum intensity in the Wen Chuan earthquake. After the earthquake, the Pengzhou government made a quick response and set up the relief command headquarter (figure 1-1) to guide the rescue and coordinate with the upper level government. Official figures stated that 956 people were confirmed dead and 5,775 people injured with 35 listed as missing. The earthquake led a direct economic loss of 239.1 billion Yuan to Pengzhou. 16 Historically, there have been twenty four earthquakes above 4.0 magnitude in
14 Pengzhou Official Website (2010). Overview of Pengzhou. Retrieved online on 6 August, 2010 from http://www.pengzhou.gov.cn/index.asp 15 China Statistics Information (2010). City Statistics Bulletin of Pengzhou. Retrieved online on 6 August, 2010 from http://www.gov.cn/ldhd/2008-06/24/content_1026042.html. 16 Pengzhou Official Website (2009). 2008 Wen Chuan earthquake work summary. Retrieved online on 6 August, 2010 from http://www.chengdu.gov.cn/GovInfoOpens2/detail_allpurpose.jsp?id=VhsI2ihgJ 1kuixtU1J 4Q 23
Pengzhou, including three that were stronger than 5.0 magnitude, and the strongest having a magnitude of 5.4. 17
According to Pengzhou Public Emergency Response Plan (Pengzhou, 2006), different level of earthquake would require related level of government to handle. The Wen Chuan earthquake is so serious that both central government and provincial government are all involved in some decision making processes. While as the study is conducted to build a DSS for the local government of Pengzhou, analyzing the current decision making processes of Pengzhou will be enough to help this research. Respondent 1 said, Decisions need to be made by Pengzhou government mainly depend on the relief command headquarter (Figure 1-1), which was constituted by leaders from different departments of the government. Mayor takes the responsibilities of the command headquarter. Generally, problems are discussed with related sub-groups during the emergency management meeting. If most of the participated people agreed with the proposed schedule or plan, then the decision would be made. Sometimes experts and people from executive organization would attend the decision making meeting, but most time they only provide their suggestions rather than involving decision making (respondent 14). In addition, the current working process of earthquake emergency management in Pengzhou is illustrated in figure 3-1. Once the earthquake happens, emergency management office in Pengzhou government will take some actions firstly, such as, going to the earthquake area, organizing emergency management troops and setting up the temporary relief command headquarter. Secondly, according to the classification standard, the command headquarter would activate relevant emergency plan. Thirdly, according to the emergency plan, the command headquarter would make decisions on conducting the rescue and coordinating the related resources. Fourth, once the earthquake risk is eliminated and approved by the upper level government, the emergency management is finished. Then the reconstruction of affected area would be taken place. Finally, the evaluation of the whole working process should be conducted.
17 China Earthquake Information (2010). Earthquakes above 4.0 Magnitude in Sichuan. Retrieved online on 6 August, 2010 from http://www.csi.ac.cn/sichuan/index080512001.htm 24
Figure 3-1 Working process of earthquake emergency management in Pengzhou 18
3.1.2 Current situation of earthquake emergency management with DSS in China Most cities in China have not yet begun to construct earthquake emergency management DSS. However, some developed cities like Beijing have established emergency management MIS. It is necessary to know the current situation firstly and find the deficiency, then according to the problems to search the solutions; therefore studying an example of current emergency management IS is the foundation for the development of DSS. Once the current situation is well understood, the deficiency of the current system would be well analyzed, and finally a better DSS would be built. Meanwhile, the reason why this study considers the Beijings current emergency management IS to be an MIS, rather than DSS, is summarized below. According to the answers from respondents 2 and 3, the systems construction goal is to integrate information flow inside of the management system (not for decision making). Therefore, its main pursuit is efficiency, not validity. Secondly, this system only contains certain preliminary decision making functions in some of its subsystems, which means it cannot solve complex semi-structured questions and therefore cannot meet the requirement of decision makers to assist them in
18 This figure is directly cited from the Pengzhou Public Emergency Response Plan (General) and translated into English by author. 25
making decisions. Finally, like other MIS, the current structures serve all related staff (not only the decision makers). Respondent 3 explained that Pengzhou had begun to build its office automation system (using computers to digitally create, collect, store, manipulate, and relay office information needed for accomplishing basic tasks) in 2007. However, there is no emergency management DSS system in Pengzhou (Respondent 1). Therefore, studying the situation in Pengzhou cannot tell us a lot about the current usage of IS in Chinese emergency management. Thus, choosing Beijings emergency MIS as an example to illustrate the current situation of using IS for emergency management will assist us in our goal of designing an ideal DSS architecture for Chinese local government. By studying the decision making functions in Beijings current emergency MIS, we can get a rough view on the emergency management DSS construction situation in China. Because Beijing as the capital of China, is the central government of China, and its emergency MIS represents the most advanced technology applied in Chinese emergency management. In addition, as is known, China has a central-governance political system, and the governance is always from up to down. Once a policy or an innovation is well executed in central government, it will promote to the provincial and local government. The most advanced DSS is studied, and then its analyzed deficiencies would be more helpful to guide the construction of the ideal DSS framework. There are four subsystems most concerned with decision making in emergency MIS of Beijing (Li & Ding, 2008). These are emergency command system, emergency technical support system, emergency plan system and emergency logistics system. The decision making functions of the four subsystems are analyzed below: Emergency command system According to Liu & Ren (2009), the emergency command system is made up of three layers. First, the decision making layer, as a unified and integrated leading system, serves Beijing emergency command center to make basic decisions, such as early warning and resource allocation. Second is the system technical support layer, which serves as a comprehensive decision making platform. Its main decision making support functions include collecting first-hand information, compiling emergency plans, and managing various emergency response data bases. Finally, the implementation layer is used to facilitate communications 26
during the emergency, record the emergency materials distribution and release emergency news (Shizhu Liu & Ren, 2009, pp. 6-7). Furthermore, some monitoring equipments can be considered as part of the emergency command system, such as on-site monitoring and on-site rescue command vehicles, which have dynamic monitoring and early warning functions. Emergency technical support system This system includes the network communication subsystem, information database subsystem, emergency assessment subsystem, and rescue subsystem. Among these, the emergency assessment subsystem provides intelligent mathematical models to support the emergency evaluation and some relevant decision making functions (Shizhu Liu & Ren, 2009, p. 8). Emergency plan system Beijing's emergency plan system contains the overall emergency plan and counties sub emergency plans (in total: 35 emergency plans). The construction of an emergency plan system is relatively out of date. Although there are many emergency plans, most of them are not imported into the information system. Furthermore, even though some emergency plans are stored in the system, it still is static and cannot adjust according to changes. Thus, when confronted with disasters, the emergency plans cannot function effectively. Due to this lagging construction of emergency plans, the auxiliary decision support function of the emergency plan system is not achieved (Shizhu Liu & Ren, 2009, pp. 10-11). Emergency logistics system The emergency logistics systems functions include the financial budgeting of emergency system operation, and rescue materials management. The function of rescue equipment and materials management can help decision making about how to distribute resources reasonably and how to ensure an effective resource scheduling (Shizhu Liu & Ren, 2009, p. 14). All in all, based on the study of Liu & Ren (2009), it is easy to see that the current decision support function of emergency MIS in Beijing is relatively insufficient for it cannot well satisfy decision makers needs and help them with decision making. 3.1.3 Earthquake emergency management deficiencies In order to understand why and where a DSS would be a useful improvement for earthquake management in China, we first have to identify where the main deficiencies are in the current system of Chinese earthquake management. 27
Based on the analysis of Beijings emergency management MIS, combining document review and the interviews in Pengzhou, this section will discuss some general deficiencies of earthquake emergency management in China which can be used to guide the goal and function analysis in the next section. Here, the deficiencies are listed as follows: (1) The level of DSS on emergency management in China is low. This argument is based on what we discussed in the previous section and the interview with respondent 1. Since the construction of DSS in China is at an exploratory stage, only some developed cities, like Beijing, have MIS with decision support functions, and their systems are more concerned with post-emergency response and recovery. Furthermore, the allocation and scheduling of emergency resources depend more on experience or static method rather than applying intelligent decision technology (respondent 1 and 4). What is worse, the local government has no DSS, and most of them only have very basic office automation systems. (2) The early warning mechanism is not perfect. According to respondent 2, the standard warning operation processes and rules are not yet established. In addition, because of insufficient integration, analysis and judgment to monitoring data, it is difficult to recognize and report the dangerous factors (which may trigger or worsen an earthquake) earlier, and this influences the effect of early warning analysis. (3) An earthquake classification index system has not been well established. According to Pengzhou earthquake emergency plan, there are some grading standards, but these standards are just a general description, which needs to be further detailed during its implementation to form a sound emergency evaluation index system. Besides, the existing classification system is more like an ex-post evaluation, as its judgment is based solely on the impact and severity of disasters. (4) Lack of classification on the emergency agencies. The vast majority of emergency classification objects are disaster events; very few do evaluation and classification on emergency response agencies. Combing disaster event with response agencies to estimate the emergency response capability is even less common (Respondent 2). However, based on his experience, respondent 2 highly recommends the combination of seismic classification with organization classification to determine emergency plan selection, and he is trying to build organization classification standards. However, the standards are not well developed and the evaluation of emergency agencies is mainly confined to the static assessment. 28
(5) The management of the emergency plan should be improved. On the one hand, the Pengzhou government has not established an evaluation working mechanism for emergency plan (respondent 3). Although there are many emergency plans, most of them are just in file cabinets, and lack necessary training and drilling (respondent 7). When earthquakes happen, the emergency plans may not function effectively. On the other hand, decision makers often make judgment subjectively on choosing emergency plans in the plan database, without scientific analysis (respondent 4). Therefore, it is better to emphasize the support function of DSS on plan assessment and selection. (6) It is difficult to adjust the emergency plan dynamically according to change of the earthquake and environment. After analyzing the overall and specialized emergency plans (e.g. Pengzhou public emergency plan and Pengzhou earthquake emergency plan), it seems that many of the existing emergency plans simply state that: action should be taken to adjust to the change of events and environment, but they do not explain how to make such adjustments. Therefore, it is better for a DSS helps to do so according to the scientific analysis, improving the effect of disaster control. (7) There are conflicts between working procedure and working philosophy among different emergency response groups, which makes the current emergency decision making process more complex and inefficient (respondent 3 and 11). Because of the different nature of various departments, they have formed their own effective way and method of working - the working philosophy and process of decision making in different departments have their own characteristics (respondent 3). During emergency response, all the departments should work together. Thus how to use the DSS to coordinate different departments, reconstruct their decision making process, and improve decision making efficiency become the core issues we need to consider. Once a DSS working process is settled, this problem would be resolved as people would work according to the procedures (respondent 11). (8) There are potential conflicts between information resource integration and data maintenance (respondent 7, 13 and 14). Earthquake emergency management is based on information, and formulating emergency decision making plans is inseparable from the integration and sharing of information. However, if the rights and responsibilities between departments are not clear (respondent 14), this will cause many problems in information collection and exchange. Therefore, it is necessary to establish an information-sharing and obtainment mechanism to ensure more accurate and effective decision making when earthquakes happen (respondent 7). 29
3.2 DSS goal and function analysis In this section, firstly, the decision support needs are analyzed. Then, based on the deficiency and support needs analysis, the DSS construction goal will be discussed. Finally, by summing up the decision support needs, the general functions of DSS would be obtained. 3.2.1 Decision support needs analysis Based on the interviews, decisions that need support are summarized in the table below. As is shown in table 3-1, in order to analyze these more logically, decision support needs have been divided into three stages: pre-earthquake, during earthquake and post-earthquake. In the first stage, the decision needs mainly come from respondent 1, 2 and 3. Their major decision support needs are whether to take some countermeasures or warn people. At this stage, the top commanders are eager to know all precise and timely information, especially earthquake impacts of different towns (such as: damage, loss and casualties) and potential earthquake related events. Once the earthquake happens, emergency management should take place. Decision support needs at this stage focus on emergency plan and resource management. All of the respondents mentioned that plan selection is one of the issues for which they need support. Because of deficient information on earthquake impacts and the selection criteria, when facing the multitude of plans in the archive, decision makers struggle to know which plan would best fit the unfolding situation. Sometimes in serious situation, they may just ignore the plans and make decision by themselves (Respondent 1, 3 and 4). When making decisions to change the plan, I act according to the current situation of earthquake and my working experience. If I can get some help from others that would be perfect (Respondent 7). The other big issue is resource management; with the exception of respondents 12, 13, and 14, almost all the interviewees mentioned this problem. They suggested that the resource requirement evaluation is very difficult in a short time, which results in lots of waste and people complain as some who need resources does not receive them. The distribution scheme is mainly designed by our experience and some basic statistic data (Respondent 4, 5 and 8). Respondent 9 also wants DSS to help decision making regarding information dissemination. Therefore, in emergency response stage, DSS needs to assist decision makers to evaluate earthquake impacts, select and adjust plans, analyze resource need, design resource distribution schemes, and communicate between different groups.
30
Table 3-1 Decision support needs Stage Decisions that need to be supported Detail support needs Pre-earthquake Whether they should take countermeasures? What measures to take? Whether they need warn people?
Monitoring and forecasting earthquake relevant events Analysis earthquake impacts to different towns in the city of Pengzhou Determine potential casualties, damage and loss assessment Earthquake information storage and retrieval issue During earthquake (emergency management / reconstruction) How to respond to the earthquake? Which emergency plan can they choose? How to adjust the emergency plans dynamically according to the development of earthquake? How to distribute resources effectively? (e.g. food, shelter, clothes, health care, Psychological counseling, etc.) Monitoring and forecasting earthquake relevant events Evaluate the possible earthquake impacts Selection criteria issues Emergency plan assessment, selection and adjustment Evaluate the need for resource (e.g. food, temporary shelters) Suitable plans for resource distribution and scheduling Execution of effective mitigation measures for earthquake Earthquake information dissemination Communication and collaboration between different groups Post-earthquake What measures to take during the reconstruction phase? How to distribute resources effectively? (e.g. food, shelter, clothes, health care, Psychological counseling, etc.) How to evaluate the relief performance? Evaluate the need for resource (e.g. long-term rooms, jobs ) Suitable plans for resource distribution and scheduling Performance evaluation
31
In the last stage, the major needs are decisions on reconstruction and performance evaluation. Reconstruction is the major issue during the post-earthquake phase which need decision maker to decide what reconstruction measures they should take, how to allocate resources to satisfy the demands of victims. Every respondent noted the issue of evaluation. One respondent mentioned, We dont have any evaluation standard and method, what we used to evaluate the performance of Wen Chuan earthquake are quite different from group to group (mainly based on experience), and this doesnt help improving earthquake emergency management. 3.2.2 DSS goal analysis Based on the insights gained through the combination of the study of current situation and problems of emergency management MIS in Beijing, the general decision support needs of Pengzhou government and the process of emergency management discussed in chapter two, we can determine a concrete goal of DSS construction: to achieve an overall earthquake emergency process management early warning before earthquake, emergency management decision making during and evaluation afterwards. Early warning is more concerned with Mitigation and Preparedness which are taken prior to an emergency; Response and Recovery refers to decision making during the emergency management; and evaluation means the performance assessment after the whole process of the earthquake. DSS should give full play to early warning. Collecting earthquake monitoring information, and analyzing it, allows an early diagnosis and evaluation about the earthquake and its possible deterioration to be done. Therefore, one basic goal of DSS is to help prediction and taking effective measures in advance. When an earthquake occurs, DSS should work comprehensively and immediately: providing suitable emergency plans to decision makers for reference, utilizing relevant information on rescue equipment, manpower, materials, shelters and medical facilities from pertinent data bases to construct a distribution scheme to allocate resources immediately. At the same time, in the process of emergency, it also needs to adjust emergency plans, in order to realize an effective treatment of a given earthquake. After the earthquake, DSS should be able to assess the effect of emergency decision making proposing the improvements on the deficiencies and inputting new cases or improvements into the emergency plan data base, and then it can optimize emergency plan management, resource allocation. 32
3.2.3 DSS function analysis Function analysis refers to the refining of decision support needs. According to our discussion about decision support needs and DSS goal as well as Chinas national earthquake emergency management plan, ensuring the fulfillment of decision support needs requires the following DSS functions: Function 1: early warning analysis. Based on China national earthquake emergency management plan and the interviews, it is easy to ascertain that warning should be a basic function. If the DSS can predict the possible development trends of an earthquake, analyze the earthquakes impacts on different geographical locations, and determine the damage and casualties, it will help to control and relieve these disasters. Function 2: seismic event classification. The classification mentioned here is different from Richter scale, and the difference between seismic event classification and Richter scale is discussed in chapter 4.1.2.1. According to the internally installed earthquake classification index system and the first-hand information on earthquake, DSS determines the severity of the earthquake automatically. The classification of seismic event is the foundation of earthquake relief. It is the classification that serves as the basis for matching appropriate organizations to certain tasks and choosing emergency plans. Function 3: organization classification. With the same logic as seismic event classification, based on the installed organization classification index system, use the DSS to classify organizations automatically and then match them to related earthquake events. Only by combining the seismic event and organization classification together with the related emergency plan, then the execution might be the most effective one. Organization classification is mainly concerned with the organizations comprehensive support capabilities. Function 4: emergency plan assessment. DSS should be able to support the evaluation of the emergency plan: nowadays, the emergency plan assessment usually takes place through experts and training (Respondent 1 and 14), which is not necessarily that efficient as most of the assessed emergency plans are not used any more. If the emergency plan assessment can be achieved in a DSS and stored in the emergency plan data base, emergency plans can be quickly retrieved and executed in the event of an earthquake. In addition, during the earthquake, the DSS can also provide periodic assessment according to the actual effect. Function 5: emergency plan selection. According to the decision support need analysis, it is clear that how to choose the most appropriate response plan from the emergency plan 33
data-base according to the seismic level and organization capacity is one major problem the DSS needs to solve. It needs to compare the seismic events classification and organization classification with the usage conditions of each emergency plan, and then select the most suitable emergency plan to be the basic response solution. Function 6: emergency plan dynamic adjustment. The system should be able to dynamically adjust the emergency plans objectives, content, and resource allocation, according to the development of earthquake. This can improve the performance of earthquake emergency management. Function 7: resource distribution. From the decision support needs, we can tell that resource management is one major issue during the earthquake response process. This function is defined as working to achieve a more rational distribution of resources based on the dynamic demand of emergency resources and its changing trends. Function 8: resource scheduling. The scheduling of emergency resources is a dynamic multi-stage resource scheduling, which needs to consider the change of resource situation and the effect of previous stages. Hence, we need to study how to carry out the optimal dynamic scheduling of emergency resources in DSS. Function 9: decision information obtainment. Based on the analysis of decision support need, storage and retrieval of earthquake related information and communication between different groups are also a problem area. In addition, the implementation of effective decision making functions relies on the quick obtainment of decision-making information. However, the widespread existence of information shortage among groups seriously affects the sharing of information resources and effective decision-making. Therefore, obtaining decision support information should be one basic function which needs to be achieved. Function 10: decision-making performance evaluation. For the overall process management, it is necessary to study how to establish a related performance assessment system to automatically realize the performance evaluation in DSS. Looking over these ten decision support functions, we find that some decision making functions share the same theoretical background and knowledge. In order to simplify the research, some of the functions will be considered together. As is known, resource management is the major part of emergency plan, in addition, during our interviews, eight informants mentioned that resource management is important decision that they should make. However, all existing emergency plans in Pengzhou do not explain on how to allocate 34
resources, only using one passage to state out which kind of resource they need to provide to people. In order to emphasize the importance of resource management, it is discussed separately - not under the category of plan management. Therefore, these DSS functions are divided into six categories, as shown in the table below. We will then study the working processes and data flow of each function in the next chapter. Table3-2 the function analysis of DSS Num. Function Category 1 Early warning analysis Early warning analysis 2 Seismic event classification 3 Organization classification and matching Classification 4 Emergency plan assessment 5 Emergency plan selection 6 Emergency plan dynamic adjustment Emergency plan management 7 Resource distribution 8 Resource scheduling Resource management 9 Decision information obtainment Decision information obtainment 10 Decision-making performance evaluation Decision-making performance evaluation
35
Chapter 4 Design of earthquake emergency management DSS 4.1 DSS business process and data flow analysis This section will give an in-depth analysis of the working process and data process for each function. The reason why we limit the discussion to only working process and data process is, firstly, our research question is focused on the general architecture of DSS and many functions should be discussed, and accordingly, choosing two important factors of DSS construction to analyze is reasonable; secondly, analyzing the information allocation mechanism is one of our secondary research tasks. 4.1.1 Early warning analysis (1) Early warning analysis working process According to our interview and the Pengzhou Public Emergency Response Plan (Pengzhou, 2006), the working process of DSS early warning function can be illustrated in the figure below:
Figure 4-1 Early warning analysis working process 19
As is shown in the figure, early warning analysis depends on the internal and external environment change before and after the earthquake. Internal environment is the management, technology and equipment level of the emergency management system. External environment
19 In the fingures of all working process analysis, rectangles presentet he general working steps; the column is the representive of databases. 36
mainly concerns other factors which are related to earthquake, such as natural environment, public policies (public security, urban planning and construction) and human factors. Based on monitoring, distinguishing, diagnosing and assessing the internal and external environment factors, signs of emergency can be detected in a timely fashion and then decision makers can take measures to warn people and control the relevant risk. If an earthquake does not happen, safe status will be restored, the warning will be cancelled and the system will go back to warning analysis again. Otherwise, an earthquake occurs, triggers emergency management and some further warning analysis is also needed. Therefore, early warning is not only in the phase of pre-earthquake, but also in emergency management. According to She (2000), early warning analysis function mainly includes four stages, namely, monitoring, distinguishing, diagnosing and assessing. Monitoring is the basis of the early warning systems operation; diagnosing and assessing are the technical analysis processes. There is an orderly causal relationship among them (see figure 4-2). Monitor Warning is primarily based on information from monitoring. Therefore, monitoring is the basis of warning analysis, and it needs to include a sensitive monitoring function which can convert the original information (as historical information or earthquake information) to certain signals for decision makers, which help them get accurate warning information. Distinguish Distinguishing is using an evaluation index system to analyze the monitoring information and recognize the signs and triggers of earthquake and earthquake related events (such as, "barrier lakes" 20 , debris flow) before and during the process of emergency management. The key is to use an appropriate evaluation index to distinguish the signals out of the information. Diagnose Diagnosis is used for analyzing the formation, process and development trends of the observed signs and triggers of earthquake and its related events during a warning or crisis status and for pointing out the most dangerous and harmful hazard-formative factors. It is important to analyze the variation of an earthquake or earthquake related event and its possible chain reaction, judge whether it is the dangerous factor and make an alarm when necessary. Assess
20 Water was shoved from its bed by rock and mud slides set off by the earthquakes. 37
Assessment is a harm evaluation of the adverse consequences of an earthquake or earthquake related event, which mainly focuses on two elements: one analyzing the harm to people and the other the economic loss to city. The appraisal conclusion is the foundation of the control countermeasures. (2) Early warning analysis data process On the basis of a clear warning working processes analysis, we can draw a data flow diagram (Figure 4-2) showing the logic of the corresponding data flow, processing and storage in early warning analysis.
Figure 4-2 Early warning analysis data flow 21
The analysis of early warning data flow is mainly focused around the monitor, distinguish, diagnose and assess stages. Firstly, earthquake monitoring information is input to monitor model. After collating and classifying the monitoring information, monitor model will then deposit some useful information into monitoring information database (D1.1). Secondly, on the basis of D1.1, the information goes into the distinguish process where the information is analyzed to identify the indicators of a potential earthquake or earthquake related events. Thirdly, at the stage of diagnose, some diagnosis on the determined risk factors are conducted by comparing the factors with the historical warning data (D1.3). The greatest risk factors are then identified, which are stated out in a generated warning diagnosis
21 In the fingures of all data flow analysis, rectangles with elliptical angles show the main working process in each function; flags are the representives of databases or reports that input into the DSS or generated from the DSS; rectangle presentets decision maker; and the ellipses are the representive of information sources. 38
report (D1.4). Finally, going to the assess stage, DSS does some evaluation on the potential harms, and then produces a warning evaluation report (D1.5) for decision makers. 4.1.2 Classification The Classification of earthquake emergency management is the recognition of incident level according to the importance of addressing a given problem, including seismic event classification and organization classification. Seismic event classification is achieved by inputting information on causes, nature, influence and harm of earthquakes according to an evaluation index system, which allows the system to classify the earthquake level automatically. This classification is more comprehensive than Richter magnitude scale as it is not only based on Richter magnitude scale, but also takes into consideration the cause, nature, potential influence and harm to people. Organization classification involves a similar input of information into DSS according to the relevant organizations importance, emergency response capacity, resource status, management level, and an installed evaluation index system, which allows the system to classify organizations automatically. We will further discuss these two variants of classification in the following sections. 4.1.2.1 Seismic event classification (1) Seismic event classification working process Nowadays, most of the countries in the world have adopted classification technology in emergency management, such as U.S. federal emergency plan 22 and China national earthquake emergency plan 23 . However, in China, most of the classification is made by individual people. Here, we would like DSS to realize this function automatically. To be specific, in this process we need to first input basic information about earthquakes. After which, DSS will operate immediately by comparing this information with the internally installed classification index system, and the resultant classification will be generated. This is shown in Figure 4-3.
22 Federal Emergency Management Agency (2008). National Response Framwork. Retrieved online on 9 August, 2010 from http://www.fema.gov/emergency/nrf/index.htm. 23 The Central Peoples Government of the Peoples Republic of China (2008). National Public Emergency Plan System. Retrieved online on 1 August, 2010 from http://www.gov.cn/yjgl/2005-08/31/content_27872.htm.
39
Figure 4-3 Seismic event classification working process According to Wu & Liu (2003b), some factors can affect the seismic event classification index system, and they should be taken into consideration when constructing a classification index system database of DSS. Here, I have summarized seven factors which might help to show the difference between this classification method and Richter magnitude scale. Seismic event side Time. When does the earthquake happen? Day or night? Workday or weekend? How long does the earthquake last? Damage level. Such as casualty, economic losses. Coverage. Such as the affected area or population of the earthquake. Diffusion elements. Check whether the earthquake can result in some epidemic disease. Management side Cognitive degree. At current level of scientific knowledge, what is the research situation on earthquake emergency management? The less cognition, the more serious the earthquake. Social influence level. Different levels of earthquake will have different impact on socio-economic development and people's life. The bigger social influence is, the more serious the earthquake is. Resources guarantee level. The higher the level of the resources guarantee is, the bigger the probability of a successful rescue. In brief, the Richter magnitude scale is more technical and is mainly concerned with how to calculate the exact degree of earthquake rather than analyzing the earthquakes effect on people or society. For instance, a very high magnitude earthquake may not be that serious if 40
there are no people living in the affected area. What we want to emphasize here is that earthquake classification should consider more factors to make decisions more precise. (2) Seismic event classification data flow According to the seismic event classification working process, its data flow can be designed as shown in figure 4-4. From D2.1, DSS can get the basic information about an earthquake immediately. Meanwhile, DSS calls the classification index system database (D2.2) and combines the two information sources, which allows DSS to do the analysis and classification automatically. Finally, the classification results are generated and reported to decision makers.
Figure 4-4 Seismic event classification data flow 4.1.2.2 Organization classification and matching Organization classification is an assessment of the emergency response organizations ability to help protect people and property. The aim of this classification is to better understand the protection and problem solving ability of each organization with the emergency response resource. Once an earthquake occurs, the command headquarter can quickly judge the level of the earthquake and then send the corresponding level organization to participate in the relief. (1) Organization classification and matching working process The general logic of organization classification is almost the same as seismic event classification. The classification factors are the major difference, which in this case are more concerned with labor force, rescue equipment, transportation, technical level and administrative ability. Therefore the classification working process will not discussed in detail, according to the aim of organization classification, the organization matching working process can be illustrated as in figure 4-5. 41
The goal of organization classification is to determine the most suitable organization to executing a given part of an emergency plan; therefore how to match the organization with the earthquake and emergency plans needs to be analyzed. According to the analysis above, the working process should be: firstly, according to the seismic event classification information, determine an emergency plan; based on the given plan, the organization capacities needed are identified; then compare the determined organization capacity with the organization classification information to find the most suitable organization for the tasks.
Figure 4-5 Organization matching working process (2) Organization classification and matching data flow Followed by the working process, organization classification and matching data flow is shown in figure 4-6. On the basis of the organization information and the classification index system database (D2.4), organization classification would be done automatically in the DSS. Later, combing the classification result with the Plan selection result (D3.6) - discussed in 4.1.3.2, which determined by the Seismic event classification result (D2.3), the matching result would be generated and submitted to decision makers through DSS.
Figure 4-6 Organization classification and matching data flow 42
4.1.3 Emergency plan management Emergency plan management refers to a whole working process that analyzes monitoring information, predicts the development trend of earthquake and makes corresponding plans. Once the earthquake happens, action should be taken according to these emergency plans. Then based on the terms of the specific situation, emergency plans need to be adjusted to minimize the possible loss. According to Wu & Liu (2003a), three functions which are related to DSS in emergency plan management, will be discussed in detail here. These are emergency plan assessment, emergency plan selection and emergency plan dynamic adjustment. 4.1.3.1 Emergency plan assessment Emergency plan assessment is the evaluation of existing emergency plans rationality and operability. Its goal is to find potential problems and improve the quality of emergency plans. (1) Emergency plan assessment working process Based on J iang (2006), an evaluation module and a computer simulation module need to be integrated in emergency plan assessment process. The working process is shown in Figure 4-7 below.
Figure 4-7 Emergency plan assessment working process There are two ways to assess the emergency plans: one is evaluation, the other is computer simulation. With the same working logic as classification, the evaluation is realized in DSS by using evaluation methods to compare the input emergency plans with the evaluation index and then get the evaluation result automatically. Meanwhile, the computer simulation module can also be used to simulate the implementation of the emergency plans in the DSS and the help 43
to optimize the emergency plans. According to the results of two assessments, a comprehensive evaluation and modification takes place. Finally, the evaluation results and modification suggestion are sent to decision makers. (2) Emergency plan assessment data flow According to the working process discussed above, the emergency plan assessment data flow can be designed as in figure 4-8. On the one hand, after importing the emergency plan information, the general evaluation module will do some assessment by combining the emergency plan information with emergency plan evaluation index system (D3.1), and then the expert evaluation result (D3.2) will be generated. On the other hand, inputting the emergency plan information will also be utilized to simulate the implementation and produce a computer simulation result (D3.3). Based on the results from the general evaluation module and the simulation module, DSS can do a comprehensive evaluation of the emergency plans and generate an evaluation result and some modification suggestions. Finally, decision makers can easily get all the emergency plan evaluation results and suggestions from the DSS to help them make a decision.
Figure 4-8 Emergency plan assessment data flow 4.1.3.2 Emergency plan selection (1) Emergency plan selection working process Emergency plan selection is actually a decision, according to the result of seismic event classification, to choose the most appropriate response plan in the emergency plan base. The working process of emergency plan selection is made up of a comparison between the results 44
of seismic classification with the use condition and scope of each emergency plan and then a choice of the most suitable one. (2) Emergency plan selection data flow The data flow of the emergency plan selection working process can be constructed as in figure 4-9. Firstly, inputting the seismic event classification result into the emergency plan selection module, DSS will call for a relevant emergency plan from the emergency plan database (D3.4). By invoking the proper judgment method (D3.5) and comparing the result of classification with the use conditions and scope of each emergency plan, an emergency plan selection result (D3.6) will be produced. Finally the result will be submitted to decision makers.
Figure 4-9 Emergency plan selection data flow 4.1.3.3 Emergency plan dynamic adjustment Emergency plan dynamic adjustment is part of the process of emergency plan implementation. According to the development of earthquake, this stage assesses the effects of implementation and potential risks and adjusts an emergency plans goals, contents and resource allocation dynamically. (1) Emergency plan dynamic adjustment working process The emergency plan dynamic adjustment working process is expressed in the figure 4-10. Based on early warning information and classification of the seismic event, the proper emergency plan would be selected from the plan database. The selected emergency plan is considered as the initial emergency plan. This process should be repeated periodically to achieve dynamic adjustment. For instance, after a certain time, the implementation of emergency plan needs to be assessed according to the update monitoring information. If the 45
original emergency plan is effective, it could continue to be followed; otherwise, DSS should work into the next step, and according to the development trend of the earthquake, choose a new emergency plans to deal with earthquake. In the dynamic adjustment process, DSS needs to keep making assessment periodically and adjusting the emergency plans until the effects of the earthquake fade away (J ie, Lei, & Hong, 2005).
Figure 4-10 Emergency plan dynamic adjustment working process (2) Emergency plan dynamic adjustment data flow The emergency plan dynamic adjustment data flow is illustrated in the figure below. First, DSS takes the plan selection result (D3.6) as an initial emergency plan. Then periodic assessment of the emergency plan implementation is carried out. By comparing Expected emergency plan implementation result (D3.7) with monitoring information of the actual earthquake situation, DSS can tell whether the emergency plan is well executed or needs to be adapted. The emergency plan implementation effect report 24 (3.8) is produced in this process as well. If the emergency plan is not well executed, then DSS goes to the emergency plan dynamic adjustment module. By calling up the adjusting model (D3.9), the emergency plan can be adjusted and an emergency plan adjustment report (3.10) generated, which will be sent to decision makers to and help them guide the emergency relief process.
24 This report may include the information on whether the emergency plan is well executed, what is the result of carrying out the emergency plan (such as, how many people are rescued), the potential problems on implementation. 46
Figure 4-11 Emergency plan dynamic adjustment data flow 4.1.4 Resource management As is explained in the previous chapter, resource management is the essential component of emergency plan, and resource management function should be under plan management function, the reasons why resource management is studied as a separate part here are, firstly, there is no such resource management countermeasures in the existing Pengzhou emergency plan (need to study in detail); secondly, during our interviews, eight informants mentioned that resource distribution and resource scheduling are important decisions that they should make. As resource management is such important, it is necessary to discuss it separately. Due to the information shortage, decision makers always feel difficult to make decisions on resource management. Therefore, using DSS to assist decision makers would be helpful. Thus, these two major issues - resource distribution and resource scheduling - in resource management need to be discussed in detail. Resource distribution is, according to the right way, placing the right amount and right types of resources to the right place in advance. Resource scheduling is a working process organizing a certain amount of resources and sending them to a specific location in a limited time. 4.1.4.1 Resource distribution Pengzhou Public Emergency Response Plan (Pengzhou, 2006) illustrates that resources the government needs to guarantee and distribute are: communication, transportation, medical treatment, goods and materials, energy (electricity and water), human resource (policemen and soldiers) and capital fund. In this study, it is better to focus resources on medical treatment, goods and materials, and human resource firstly; since they are the most important 47
matters that can save or help victims. Once the DSS is well executed, the resources can then be expanded to the others mentioned in the Pengzhou Public emergency response plan. According to Qi, Chi, Zhao, & Sun (2006), there are three decision making functions which are related to resource distribution in emergency management: resource location, resource coverage and resource allocation. ReVelle & Eiselt (2005) state that resource location is, given certain demands, an analysis of how to site facilities in given places and then minimize the distance between demands and facilities. Four components characterize resource location problems, which are customers presumed to be already located; facilities will be sited; place where customers and facilities are located; and a metric showing the distance and time between customers and facilities. In this research, customers can be considered as the victims, facilities as the resources needed by the victims and the demands refer to the potential victim needs under earthquake emergency. Resource coverage suggests some given standards (mainly concerned with affected people) that need to be satisfied by the facilities and resource. For instance, if 20,000 people are affected in an earthquake, the resource should cover those people. Resource allocation is always related to resource location. According to Rahman & Smith (2000), siting facilities and assigning resource to these facilities are actions taken simultaneously with the aim of optimizing the resource utilization. (1) Resource distribution working process Based on the literatures and Pengzhou earthquake response plan (Pengzhou, 2007), the resource distribution working process can be summarized as shown in figure 4-12. Ordinarily, according to the emergency response mechanism and potential earthquake level, the demand of resource could be roughly calculated. The general demand of resource is inputted into a mathematical model (such as, optimization model or linear programming model) to figure out how to distribute resources. Resource location, coverage and allocation should all be fully considered in this decision making process. In order to improve the generated resource distribution scheme, the evaluation on resource distribution effect needs to be executed. 48
Figure 4-12 Resource distribution working process (2) Resource distribution data flow The data flow of resource distribution is expressed in figure 4-13. First, inputting location related information (location of facilities and targeted people) and coverage requirement (how many people need to be covered) and by invoking the location model 25 (D4.1) and coverage model 26 (D4.3), the solutions of resource location and resource coverage are made and the location (D4.2) and coverage lists 27 (D4.4) are generated. Based on the two lists and imported resource demand information, under the support of mathematical model (D4.5), resource allocation can be achieved and the resource distribution list 28 (D4.6) can be produced. This list then assists decision makers in fulfilling demand effectively.
25 The mathematical model to determine the places to put the resources in order to minimize the distance between victims and resources. 26 The model to figure out the potential resources distribution schemes in order to maximize the coverage of affected people. 27 The location list includes the information on potential location places; the coverage list contains the information on all potential covered areas. 28 The resource distribution list is actually the distribution plan. Decision makers can use the list to allocate resources. 49
Figure 4-13 Resource distribution data flow 4.1.4.2 Resource scheduling Once resource distribution is well planned and executed, it will help to guarantee the resources which are needed during the earthquake. However, resource distribution is only a necessary condition of effective supply. Optimizing resource scheduling is the other important task to which we should pay attention. When earthquake happens, the efficiency of resource scheduling is the fundamental guarantee of emergency management. Based on aid resource demands, the amount of resource and its transport routes need to be determined, and then adjusted according to the earthquake development trend to optimize the resource scheduling plan. (1) Resource scheduling working process According to Barbarosogcaronlus two-stage stochastic programming model for planning first-aid commodities transportation in case of an earthquake (Barbarosogcaronlu & Arda, 2004), resource scheduling working process can be illustrated as in figure 4-14. First, the demand for emergency resources and the situation of resource distribution should be identified. Then optimization model should be used to determine the resource scheduling plan. Thirdly, it is necessary to assess the scheduling scheme periodically to make sure that it is effective. Based on the result of assessment and the earthquake development information, dynamic adjustment of resource scheduling is conducted. 50
Figure 4-14 Resource scheduling working process (2) Resource scheduling data flow According to the working process discussed above, resource scheduling data flow is expressed in figure 4-15. In the first step, it is necessary to obtain the resource demand information from the emergency resource demand list 29 (D4.7). Secondly, using the resource distribution list (D4.6), under the guidance of the resource scheduling model 30
(D4.8), a new resource scheduling scheme 31 (D4.9) would be created. By importing the earthquake development information into the process of resource dynamic adjustment and calling for the adjusting model (D3.10), the resource adjustment scheme (D4.10) would then be generated and eventually this result would feedback to decision makers.
Figure 4-15 Resource scheduling data flow
29 The list illustrates that all the emergency resources which are needed by victims. 30 Mathematical model such as optimization model used to determine the most optimized resource scheduling scheme. 31 Resource scheduling scheme is actually the instruction on how to schedule the resources. 51
4.1.5 Decision information obtainment From the construction of the emergency MIS in Beijing, it can be inferred that the existing system has problems like system segmentation, information islands 32 and inefficient information acquisition from the terminal (such as monitoring equipments). All of which seriously hinder the quick obtainment of decision information. In order to obtain decision information quickly, it is better to extract useful information from a variety of sources and then integrate them effectively. (1) Decision information obtainment working process Based on the interview and Pengzhou public emergency response plan (Pengzhou, 2006), the decision information obtainment working process can be summarized in three steps: firstly, obtain useful decision making related information from a wealth of related channels, such as, professional emergency management IS, monitoring equipments and plan database. Secondly, integrate such scattered information logically to allow this information to be better managed, used and shared by decision makers. The final step is to use some technical means such as a data warehouse 33 to ensure decision makers can get information quickly and accurately. (2) Decision information obtainment data flow According to decision information obtainment working process, its data flow can be expressed as shown in figure 4-16. First, various decision-making information can be obtained from professional emergency management information systems, monitoring equipments, emergency plan database (D3.4), and resource distribution list (D4.6). Then through the information integration module, and with the support of relevant technology, the useful integrated information can be obtained quickly by decision makers.
32 A body of information (i.e. electronic files) that needs to be shared but has no network connection 33 A repository of an organization's electronically stored data, designed to facilitate reporting and analysis 52
Figure 4-16 Decision information obtainment data flow
4.1.6 Decision making performance evaluation
(1) Decision making performance evaluation working process According to Pengzhou earthquake response plan (Pengzhou, 2007), the final step is performance evaluation which will be executed after the whole earthquake, not in the process of emergency management. Feng & Xu (1999) presented an integrated system in which a knowledge-based decision support system is integrated with a multilayer evaluation index system. Based on this idea, the working process of decision making performance evaluation is outlined as shown in figure 4-17 and the specific evaluation steps are as follows. During the performance evaluation process, we need to input some basic information about the result of decision making performance in the phases of emergency management and reconstruction, such as how many people are rescued and whether people are satisfied with the resource allocation or their resettlement. Next, DSS will operate immediately by comparing this information with the internally installed evaluation index system, and then the evaluation result will be produced. 53
Figure 4-17 Decision making performance evaluation working process
(2) Decision making performance evaluation data flow Decision making performance evaluation data flow is illustrated in figure 4-18. Firstly, from D6.1 DSS can get the basic information of decision making performance immediately. Calling upon the decision making evaluation index system database (D6.2), DSS will combine the two information sources utilizing its installed model base and automatically perform the analysis and assessment automatically. Finally, the evaluation results will be generated and reported to decision makers.
Figure 4-19 Decision making performance evaluation data flow
4.2 General architecture of DSS The framework of the earthquake emergency management DSS in our research is designed to be a four-layer structure, which is shown in Figure 4-20. The reason it was designed like this is rooted mainly in relevant literature. First, the rationale of this architecture is based on the generic DSS architecture designed by Burstein & Holsapple (2008), which we discussed in 54
detail in chapter 2.2.3. Furthermore, other relevant research was also utilized in formulating the structure, such as: Chen & Yasuhiko (2000) who introduced a DSS development method - Layer Model Method. In their study, a DSS contained four layers for business decision making. Wang (2005) designed a five-layer emergency management information system. Both of these groups discussed the core content and relationship of each layer as well. Mao & Zhu (2007) built an architecture of DSS for emergency planning and also discussed its hierarchical structure. Previous analysis of the decision support needs in Pengzhou and the DSS function analysis were also considered in the process of framework design, especially for the basic information layer and decision application platform. Generally speaking, this architecture is a compound DSS framework, which is a single DSS and combines more than one knowledge management techniques, such as text, database, or rule management to support decision makers. To be specific, there is one use interface for decision makers to learn, and they can realize different decision making requests through multiple knowledge management techniques. The bottom layer is the hardware support layer, which consists of information acquisition equipment and communication equipment. The upper layer is the basic information layer, which includes basic information, business information, and comprehensive information. The third layer is problem process layer, which contains two parts: the decision support platform and decision application platform. The decision support platform tries to recognize and solve problems, and the decision application platform is used to achieve all the functions of DSS, such as early warning analysis, resource distribution, and emergency plan selection. The top layer is the user interface layer, which can be designed according to different users. We will analyze each layer in detail in the following sections.
55
Figure 4-20 Architecture of DSS 56
4.2.1 Hardware support layer Chen & Xu (2001) state that the physical layer is the platform for information transportation, and is the basis of DSS. Combining this idea with our research, the hardware support layer should be placed in the bottom of the system and includes at least the following three parts: computer network system, communication system and basic hardware devices. (1) Computer network system A computer network is any set of computers or devices connected to each other with the ability to exchange data (Kizza, 2005, p. 3). The computer network system is the foundation of information transmission, information sharing and intelligent decision-making. It should rely on the existing network infrastructure of the government, such as the network of office automation systems. Utilizing the development of network technology and the interoperability of the existing local area network (LAN) 34 and new LAN, the government should build a safe, reliable, advanced, practical and standardized computer network system for DSS. Based on the study of Liu & Zhang (2007), the content of network infrastructure should include at least the following two parts: Emergency department network Emergency department network is any set of computers or devices connected to each other with the ability to exchange data in emergency departments. Its construction mainly consists of three separate parts: a network for public service (devices used in order to exchange data with the public), a network for non-confidential information transportation (devices used for earthquake emergency departments to exchange general data), and a network for confidential transportation (devices are set up to transport the confidential data). This network is mainly designed for the transportation of basic information among different emergency departments. Emergency command network An emergency command network is a group of connected computers or devices with the ability to exchange data in the emergency command center. This network is the base for decision makers to obtain or transfer information and to get the decisions made by DSS automatically. This network is mainly designed for decision makers. Communication system
34 A computer network covering a small physical area, like a home, office, or small groups of buildings, such as a school, or an airport. 57
Based on the definition of Schwartz, Bennett, & Stein (1995), a communication system is a collection of individual communication networks, transmission systems, and data terminal equipment by which users can exchange information. It usually is capable of interconnection and interoperation which forms an integrated whole. The application of many decision making functions depends on the relevant data and signals extracted from the terminals, and then transmitted them in the DSS. (2) Other hardware devices Other hardware devices are the rest of the hardware equipments not used in the computer network system or the communication system, which needs to be configured into the DSS. These could include monitoring equipments, servers, transmission equipment, on-site monitoring devices and on-site rescue command vehicles. 4.2.2 Basic information layer DSS must obtain the comprehensive emergency information quickly and accurately. The basic information layer is designed to provide all the information needed by the decision support platform, such as, seismic events information, emergency response organization information, resource relevant information, environmental and geographic information, expert information and emergency plan information. Based on the general architecture of Burstein & Holsapple (2008), this layer mainly refers to Knowledge System (KS) we discussed in chapter two. According to Mao & Zhu (2007), the information layer is used to manage and maintain all information about the disaster. However, they do not clarify the information categories. The more structured the organization of the information is, the more effective the system will be. By combining the characteristics of different information, information in this layer is divided into the following three categories: (1) Basic information Basic information mainly refers to the public information which would be used in the decision making process, including the spatial and geographic information database, material reserve database, population database, building status database and monitoring equipment database 35 . (2) Business information Business information mainly refers to the professional information resources which can be used in the emergency decision making process. Such information can, not only support decision making, but also provide data support for cross-sectoral applications. Business
35 Explanation of each data base, refer to table 4-1 and table 4-2. 58
information includes incident report database, alarm receipt database, expert database, forecasting warning database, and decision making evaluation index database. (3) Comprehensive information Comprehensive information primarily refers to the information which is generated by analyzing and integrating various types of basic information and business information. It mainly includes data warehouse, model base, knowledge base, emergency plan base, assessment base and statistic base. Data warehouse is a repository of an organization's electronically stored data, designed to facilitate reporting and analysis (Inmon, 1995). According to the information needs of decision making function, data warehouse will extract, integrate and convert relevant data from basic information and business information, and import it in a uniform format into the data warehouse to meet the informational needs of DSS. Model base provides various models which can be used to assist emergency decision making, such as linear programming model, structural equation model, and Bayesian decision model. It the most unique characteristics of DSS, which is also very important for the fulfillment of the decision making function (Fan, 2005). A knowledge base contains a variety of emergency management related expertise knowledge and accumulated practical experiences in previous decision-making processes. An emergency plan base is used to store emergency plans according to the standard format, which clearly presents the conditions, limitations and scopes of each emergency plan. Expert evaluation on emergency plans and emergency implementation effect is stored in assessment base, which assists in achieving information assessment management optimization. A statistic base stores the statistical information by categories. The information will be used in the decision making process to better scientifically support relevant decision. 4.2.3 Problem process layer The problem process layer is the core part of DSS, which is used to realize all decision making requests. It contains two components: a decision support platform and a decision application platform. The decision support platform is the central part of problem process layer, and all problems are solved by this platform. Meanwhile, the decision application platform is designed to classify different decision making needs and then transfer the specific decision demand to decision support platform. The two platforms will be discussed in the following sections. Based on the general architecture of Burstein & Holsapple (2008), this layer mainly refers to Problem Process System (PSS). 59
4.2.3.1 Decision support platform The decision support platform is between the basic information layer and decision application platform. According to the specific requirements, the decision support platform provides various supporting service to help the decision application platform achieve its functions. Based on our DSS architecture, the decision support platform provides four kinds of service. (1) Security support service The main role of security support services is to ensure DSS working safely, stable and efficiently. Users have their own settled right to access and operate the DSS, according to different permission levels. For example, the top leader should have the full rights to operate DSS, both for data inputting and decision making whereas the average worker in an emergency department would only need the right to input data. Based on the study of the National Training Standard for Information System Security (Infosec) Professionals 36 , some main components were identified: identity authentication, permission distribution, access control, data encryption, digital signature, and system maintenance. Identity authentication is used to ascertain the identification of people who use the DSS; permission distribution allocates corresponding permissions to relevant people; access control is the technology that prevents people without authorization from operating the DSS or some key function of DSS; data encryption is the process of converting ordinary information into complicated codes to ensure the safety of certain information; digital signature is a scheme for demonstrating the authenticity of a digital message or document; system maintenance mainly refers to updates for the software and hardware of the DSS. (2) Operation support service Operation support service, according to the needs of the decision application platform, provides assistance technology and service to ensure successful decision making. According to the studies of Liu, Li, & Luo (2006) and Liu & Ren (2009), the operation support service in the DSS consists of GIS, GPS, RS, and simulation platform. GIS is the system that captures, stores, analyzes, manages, and presents data that is linked to a physical location. GPS is a space-based global navigation satellite system that provides reliable location and time information in any weather, at all times and anywhere. RS is stand-off collection done using a variety of devices or equipments to gather information on a
36 National Security Telecommunications and Information Systems Security (1994). National Training Standard for Information Systems Security (Infosec) Professionals, Retrieved online on 6 August, 2010 from http://www.sis.pitt.edu/~jjoshi/courses/IS2150/Fall09/nstissi_4011.pdf.
60
given object or area. GIS, GPS and RS technologies are important elements for establishing DSS, which can meet the needs of observation, information processing and analysis. A simulation platform is used to simulate disasters and measure the corresponding implementation effect of emergency management in a virtual environment. It can be widely used to assess emergency plans and resources distribution. (3) Business support service Business support service refers to different kinds of services which will be used in the process of decision making. Based on the research of Mao & Zhu (2007), business support service should include a data warehouse management system, a model base management system, a knowledge base management system, an emergency plan base management system, an assessment base management system, a statistic base management system and an inference engine. Data warehouse management system is a sub-system to manage and make data warehouse work more efficiency. The goal of setting this management system is to achieve a more effective management of massive amounts of data by Extract Transform and Load (ETL), 37
On-Line Analytical Processing (OLAP), 38 data mining 39 and other services (Inmon, 1996). Model base management system is a sub-system that manages models. It has three main functions, namely, model storage management, model operation management and model portfolio management (Gao, 2005). Model storage management is used to express, query and maintain models; model operation management is used to compile model program, control model operation, and input data into models; model portfolio management is mainly used to combine or connect different models. A knowledge base management system is used to manage a knowledge base, through actions such as querying, browsing, adding, deleting, modifying, and maintaining the data in knowledge base. With the development of data mining technology, such a system can dig out knowledge through a large amount of data in the database. Meanwhile, it can also import new knowledge obtained from the decision making process into the knowledge base and thus enrich the knowledge base.
37 ETL is a process in database usage and especially in data warehousing. It involves: extracting data from outside sources; transforming data to fit operational needs; loading it into the end target (database or data warehouse). 38 OLAP is an approach to swiftly answer multi-dimensional analytical queries. 39 The process of extracting patterns from data 61
The basic function of an emergency plan base management system is the input, storage and output of plans. One critical success factor for emergency decision making is whether the appropriate emergency plan can be activated quickly (Respondent 2). Emergency plans base management system was established for facilitating the management of various emergency plans. An assessment base management system is mainly applied to aggregate various types of assessment information, such as emergency plan assessment information or decision making performance assessment information. Based on the aggregated evaluation information, DSS can optimize decision making on a regular basis, improving the quality of early warning analysis, emergency plan assessment, classification, and decision making performance evaluation. A statistic base management system is used to aggregate statistical information in the statistic base according to the specific decision making request. The aggregated information is used to support the relevant decision making process. An inference engine 40 is one of the most important elements of DSS, which is used to complete the process of reasoning and then to assist in getting to the final result of decision making. Inference engines constructs a reasoning chain from the raw information, to the integrated information, and, after reasoning, to an unknown targeted decision making result. According to the dynamic information during an earthquake and the basic information in the database, DSS should use an inference engine to reason, and then develop decision making proposals. (4) Problem processing service A problem processing service is, firstly obtaining information from data warehouse, model base, knowledge base, emergency plan base, assessment base and statistic base, analyzing the problem and then designing the solution schemes. Problem processing services include problem analysis and problem solving systems. A problem analysis system is used to attain information from relevant databases and then analyze the decision questions. A problem solving system is used to deal with the analyzed problems until the solutions are promoted and then send the analyzed results to the decision application platform.
40 Inference engine is a computer program that tries to derive answers from a knowledge base 62
4.2.3.2 Decision application platform A decision application platform is a core component of DSS, which is mainly used to support the implementation of the decision making functions discussed in chapter 3: early warning analysis, seismic event and organization classification, emergency plan assessment, emergency plan selection, emergency plan dynamic adjustment, resource distribution, resources scheduling, decision information obtainment, and decision making performance evaluation. The reasons why we set this platform are, first, classifying the decision making requests before processing will enhance the efficiency of decision making results; secondly, 10 decision support functions are characteristic of our earthquake emergency management DSS. Thus these functions should be listed to distinguish this particular DSS. A decision application platform is used to connect the decision support platform and the user interface. When obtaining a decision making request from user interface, the request will be distributed into the related function. Utilizing relevant information from the database, model base, knowledge base, emergency plan base, assessment base and statistic base, the request is processed in the decision support platform, and a decision making result will be produced. This result will be fed back into the relevant decision application platform, and then sent to the decision maker automatically. 4.2.4 User interface User interface, as an integral part of the DSS, is the bond that connects people and the DSS system. On the one hand, people import information to DSS or submit their decision needs to the system; on the other hand, the DSS provides solutions or decision support information to people through the interface. Additional information from people through the interface may be required to complete some decision tasks. Based on the architecture of Burstein & Holsapple (2008), our user interface is mainly concerned on the Language System (LS) and Presentation System (PS). According to Gao (2005), the main tasks of the user interfaces includes the following: Control DSS: it allows decision makers to control the system through the interface. Provide decision makers with various forms of interaction with interface: decision makers can access DSS through Web page, client-side system, cell phone, PDA and other channels according to their permission levels. Input and output: decision makers correctly input data and relevant parameters, and the DSS should be able to correctly output the operation results to decision makers. Adaptability: the interface should be able to expand and adapt to changes. 63
4.2.5 The relationship of the core components The previous five sections have discussed the overall architecture of DSS, and the structure and function of each layer. However, the general framework of DSS does not explain how each layer interconnects with others. The following section will discuss how the core components work together to produce effective decision making solutions. 4.2.5.1 Structure of core components Figure 4-21 shows the relationship of the core components in the architecture. Data collecting gathers relevant information from the user interface and transmits the information through the hardware support layer to basic information layer. All kinds of different data bases in figure 4-21 belong to basic information and business information parts in the basic information layer. The data warehouse and data warehouse management system, model base and model base management system, and the other base and its related management system belong to comprehensive information in the basic information layer and business support service in the decision support platform respectively. The inference engine is also part of business support service in the decision support platform. The problem processing service belongs to the decision support platform in problem process layer. Various decision making functions belong to the decision application platform. Finally, the user interface belongs to the user interface layer. 4.2.5.2 Working principle of the core components From bottom to up, the working principles of core components in the overall DSS framework can be summarized as follows: (1) Data collecting gathers valuable information through the hardware support layer and then stores the information in various related databases. The data warehouse extracts, transforms and loads the basic information and business information from relevant databases and then saves them into the data warehouse according to their topics. At the same time, data mining can obtain reliable information from large data sets and store it in the knowledge base further enriching the content of the knowledge base. (Inmon, 1996). (2) According to the specific requests of decision making, the model base and model base management system, knowledge base and knowledge base management system, and other base and its related management system can extract relevant models, emergency plans, knowledge, statistical information and assessment results from related database. Based on the 64
obtained data, the inference engine can conduct a reasoning process, according to the dynamic information about the earthquake.
Figure 4-21 Relationship of core components in the architecture 65
(3) Based on the previous step, the problem processing system makes full use of the available data and the inference engine to analyze and solve decision making problems. (4) The user interface is used to guide users to interact with the system: first, to define the decision problem; then to submit the request and relevant system parameters to DSS; after system operations, the emergency response decision will ultimately be generated and fed back to the decision makers. 4.3 Key data bases design 4.3.1 Data base design DSS is a large complex system, involving broad and diverse kinds of data. In order to achieve system functions effectively and reduce data redundancy, it is better to consider all the databases that a DSS needs before its construction. Based on the interview with respondent 2 and the data flow analysis in chapter 3, the data bases needed in this DSS have been sorted out. Here, we are going to discuss the data bases in two parts: existing data bases and needed data-bases. (1) Existing data bases in Chengdu Based on the interview with respondent 2 and the emergency response plan of Chengdu, the existing data bases in Chengdu were summarized in the table below. They were divided into three stages: before earthquake, during and after. Because of confidentiality issues, we can simply list the name and general content of each data-base. Table 4-1 Existing data bases 41
Num. Name General contents Pre-earthquake 1 Spatial and geographic information database Geographical information on Chengdu and its administrative area 2 Materials reserve database The distribution, quantity and storage related information on food, clothes, tents and other relief supplies; reserve aid manufacturers and their daily production capacity information. 3 Population database Population density, composition and other relevant information, such as, community (settlement) distribution, floating population distribution, school population distribution, and foreigner distribution 4 Building status Distribution information on public places, including
41 The existing data bases were checked with respondent 2. As he said, some of the databases were constructed during or after the 2008 Wen Chuan earthquake, such as data base 6, 7, 8, 10, 12, 13, 14, 18 and 19. Database 11 is used to record all the public emergency incidents, such as, earthquake, flood, and mud rock flow. 66
database their volume, structure, and other related information; information on the current buildings 5 Monitoring equipment database Distribution and attributes information on kinds of monitoring instruments, such as, earthquake measuring instrument, meteorological measurement instrument. 6 Facilities distribution database The distribution information on medical, firemen, police, and army; and their facilities related information 7 Communication facilities distribution database Communication facilities distribution information, such as, communication lines and stations, emergency communication network 8 Transportation power database The distribution, capacity, and other relevant information of transportation teams 9 Rescue equipments database The distribution and quantity information on rescue equipments, such as, crane, excavators 10 Historical statistic database Historical data on previous earthquake, such as, magnitude, casualties, economic loss Emergency management 11 Incident report database Real-time information on the public incidents, recording when, where, reason, response measures and results 12 Alarm receipt database Alarms received on earthquake-triggered social crime during the emergency management, such as rob, murder and other mass events which may result in social unrest 13 On-site information database Record the on-site earthquake information, including video, pictures 14 Released information database Released earthquake related information to the public during emergency management 15 Weather monitoring database Monitoring information on temperature, wind speed, wind direction, and rainfall etc. 16 Seismic monitoring database Monitoring information on earthquake and geology 17 Traffic monitoring database Monitoring information on traffic Post-earthquake 18 Emergency loss database Information on earthquake-triggered losses, such as, casualties, economic losses to the city 19 Rehabilitation database Summary information of rehabilitation information from government aid, social assistance and insurance claims
(2) Needed data bases Mao & Zhu (2007) conducted a specific study on the database design of DSS. Combining his research and data flow analysis in Chapter 3, the needed databases are summed up in the table 67
below. As many of these databases would be used throughout the emergency management process, they are not classified into pre-earthquake, during and post-earthquake. Table 4-2 Needed data bases Num. Name General contents 1 Expert database Record relevant information on experts, such as, name, major field, contact information, and classical earthquake ermergency management cases they have solved 2 Forecasting and warning database Prediction and warning information on weather, earthquake, and other related information 3 Emergency plan assessment database Emergency plan assessment index system and historical evaluation data 4 Seismic event classification index database Information on seismic event classification index system 5 Organization classification index database Information on emergency organization classification index system 6 Emergency plan selection database Historical emergency plan selection information and the conditions of choosing each plan 7 Emergency plan adjustment database Reasons for emergency plan adjustment, and information on how to adjust plans 8 Resource distribution database Relevant information on resource location, coverage and allocation 9 Resource scheduling database Relevant information on the place, quantity and route of resource scheduling 10 Information resource integration database Obtain information from various channels and integrate them 11 Decision making evaluation database Information on decision making evaluation index system and historical evaluation data 12 Earthquake case database Information of earthquake emergency management cases
Based on these databases, the data warehouse will extract the needed data by using certain extraction and conversion tools. The data would be stored separately according to different themes and could be used to better support the decision making process. 4.3.2 Model base design A very promising aspect of a DSS is its ability to integrate data access and decision models (Sprague, 1980). Based on the research of Power & Sharda (2005) and Densham (1991), the model base in our research is designed including a model base and a model base management system. The model base is a static data-base used to store models in a certain structural form. The general types of models included in the model base could be, for instance, algebraic 68
models, forecasting models, optimization models, decision analytic models, simulation models, and differential equation models. The model base management system is used to manage the model base assisting decision makers in constructing, modifying, and applying various models. In order to avoid model repetition and redundancy, the model base is designed to be a shared resource, and all the models can be reused and invoked by different decision making functions. In addition, models can be used together (two or more models are combined to solve one problem) in order to achieve a better solution. Both because research focuses on the overall design of DSS and also due to time limits, we cannot go to every detail of the model base design. Here we are trying only to explain the roles the models play in each category of the decision making function. (1) Early warning analysis There are four models needed in the early warning analysis function: a monitor model, a distinguish model, a diagnose model and a assess model. Firstly, the monitor model is used to collate and classify earthquake monitoring information and then import the monitoring information into the monitoring information database. Secondly, the distinguish model is used to compare monitoring information with the warning evaluation index and distinguish the hazard-formative factors. Thirdly, the diagnose model is used to diagnose the important risk factors by comparing them with historical warning data. Finally, the assess model is used to do a comprehensive evaluation on the potential harm, and produce a warning evaluation report for decision makers. (2) Classification As the working process of seismic event classification is almost the same as organization classification, the model utilized in the two processes is the same. Thus, we will discuss them together. Matching is the only model needed in this working process. This model is used to compare basic information on an earthquake or an organization with the related classification index system. If the conditions of the earthquake or the organization match the related classification index, then the classification result will be generated. (3)Emergency plan management Emergency plan assessment 69
According to the working process of emergency plan assessment, three types of models are needed in emergency plan assessment: matching models, simulation models and decision analytic models. Matching models are applied to compare the emergency plan information with the emergency plan evaluation index, and generate evaluation result. Simulation models are used to simulate implementation and generate a simulation result. Decision analytic models include various analysis methods, such as Analytical Hierarchy Process (AHP) 42 , decision matrix 43 and decision tree 44 (Power & Sharda, 2005). In emergency plan assessment, the decision analytic model is used to do a comprehensive evaluation on emergency plans and generate an evaluation result. Emergency plan selection Similar to classification, emergency plan selection only needs a matching model. It is used to compare seismic event classification results with the use conditions and scope of each emergency plan in the emergency plan database and then produce the emergency plan selection result. Emergency plan dynamic adjustment There are two models needed in this decision making function, the decision analytic model and the adjusting model. The decision analytic model is used to compare the expected emergency plan implementation result with the actual situation to check whether the emergency plan is well executed or to adapt to changes. Once the emergency plan needs to be adjusted, the adjusting model will be activated. It is used to generate emergency plan adjustment reports according to the changes of the earthquakes effects. (4) Resource management Resource distribution Three types of models are needed in the resource distribution function: a location model, a coverage model and a mathematical model. A location model is an operation model, which is used for solving the optimal location selection problem. In the working process of resource distribution, the location model is used to determine resource location by analyzing location
42 AHP is a structured technique for dealing with complex decisions. It provides a comprehensive and rational framework for structuring a decision problem, for representing and quantifying its elements, for relating those elements to overall goals, and for evaluating alternative solutions. 43 A decision matrix is an arrangement of qualitative or quantitative values in rows and columns that allow an analyst to systematically identify, analyze, and rate the strength of relationships between sets of information. 44 A decision tree is a decision support tool that uses a tree-like graph or model of decisions and their possible consequences. 70
related information and sending the result to the location list. A coverage model is used to determine resource coverage under the conditions of the coverage requirement. Mathematical modeling, such as an optimization model or a linear programming model, is used to achieve the resource allocation given the resource location and resource coverage conditions. Resource scheduling A resource scheduling model and an adjusting model will be applied in this decision making function. A resource scheduling model is used to analyze resource distribution information and resource demand information and then generate the resource scheduling scheme. A adjust model is used to adjust the resource scheduling scheme, by analyzing the earthquake development information and then generating a resource adjustment scheme. (5) Decision information obtainment Decision information obtainment is mainly concerned with the data warehouse, data mining and some other technical means for attaining information; therefore, no model is related to this function. (6) Decision making performance evaluation A matching model is needed in the decision making performance evaluation, which is used to compare decision making performance information and a decision making evaluation index, and then generate evaluation results. 71
Chapter 5 Implementation 5.1 Potential problem Because there is no such DSS in Pengzhou, even in Chengdu, our interviewees cannot provide much useful information on DSS implementation. The analysis of implementation is based on both interview and literature review of information system implementation. 5.1.1 From interviews (1) There are no standards or rules on classification and evaluation. According to the interviews, at least three index systems should be fixed before DSS construction: Earthquake classification index system. The establishment of earthquake classification index system is a foundation for achieving the classification function of DSS. According to the Pengzhou earthquake emergency plan, there are some grading standards, but these standards are just general descriptions, which need to be fleshed out in further detail. Besides, the existing classification system is more like an ex-post evaluation: its judgment is based solely on the impact and severity of disasters. Therefore, a better earthquake classification index system (used in the whole process of earthquake management) should be built. Organization classification index system. An organization classification index system is primarily concerned with the organizations comprehensive support capabilities. We do not have any organization classification index now. When the earthquake happens, I always discuss with my colleagues firstly, and then make the decision according to my knowledge on each organizations capability (respondent 1). If there is no such index system, DSS cannot classify organizations automatically. Therefore, a well-built organization classification index system is also a necessary foundation for DSS construction. Decision making evaluation index system. With the same logic of previous index systems, establishing a decision making assessment index system is the basis for the function of decision making performance. (2) Leaders awareness about DSS is inadequate. During the fourteen interviews, only respondent 1, 2 and 3 had some general ideas about DSS. Most of the other respondents had never even heard about the concept of DSS. 72
(3) Most of the respondents held a negative attitude toward the implementation of DSS. We do know this is an advanced technology, and it can help us a lot, but Pengzhou is just a county level government, the IT level and financial budget cannot afford to do this (respondent 3). Almost every interviewee holds the same perspective as respondent 3 on the implementation of DSS. (4) Low dedication of resources. Respondent 1 said, Although the current infrastructure (computer, internet access) is much better than 3 years ago, it is still not well developed in the town level there is only one computer in Bailu town. (5) Low computer literacy. Because the education level of the town level government officers is not that high, most of them cannot operate computer properly. In addition, some elder people (above 50) in the Pengzhou government office cannot use computer either, and sometimes they even refuse to learn (respondent 1). (6) Complex and inefficient decision making process. As mentioned in chapter 3.1.3, there are conflicts between working procedure and working philosophy among different emergency response groups which makes the emergency decision making process inefficient. While the characteristics of an earthquake like, ad-hoc nature, destructiveness and uncertainty will bring huge pressure to decision makers. These decision makers must make quick decisions in order to minimize the damage. Therefore, building a simple and efficient decision making process and embedding it in the DSS has a very great significance. 5.1.2 From literature Since the literatures on DSS implementation is relative scarce, and DSS is an information system, a review of relevant literatures on information system (especially, Enterprise Resource Planning system) implementation also suggests some potential problems with the implementation of DSS: (1) Lack of top management support, data accuracy, and user involvement are potential problems that lead to information system implementation failures (Sum, Ang, & Yeo, 1997). (2) Education, training and understanding of cross-functional working processes are often reported as other important factors in information system implementation (Markus, Axline, Petrie, & Tanis, 2000). (3) Working process changes during the implementation of a DSS system are needed (Al-Mashari & Zairi, 2000). 73
(4) Zhang, Lee, Zhang, & Banerjee (2003) claim that lack of top management support, lack of discipline, resistance, and lack of broad-based inter-organizational commitment are the major factors that slow down the process of information system implementation. (5) When adopting an information system, there is a need to recognize the unique local government context in which the system will be embedded (Moosbruker & Loftin, 1998). (6) Maintaining the software and hardware sustainably has a positive impact on information system implementation success (Zhang, et al., 2003). 5.2 Suggestion on implementation Based on the potential problems analysis, suggestions on implementation are classified into three categories: organizational environments, including top management support, working process setting, government-wide commitment, index systems setting, financial resources and effective project management; people characteristics, including user involvement and education & training; technical supports, including data accuracy and construction of hardware and software. We will discuss each of them in detail below: 5.2.1 Organizational environments (1) Leader support Much research has emphasized that leader support is a necessary component in information system implementation (Burns & Turnipseed, 1991; Duchessi, Schaninger, & Hobbs, 1989). Since earthquake emergency management DSS is an inter-organizational information system, its design, implementation, and operation require the cooperation of all people from all segments of the government. Leader support can help in solving disputes and clarifying doubts. Leaders should create an environment for implementing a DSS system. Their support in DSS implementation can be divided into two main facets: first, providing leadership, and, secondly, giving necessary resources. To implement an information system smoothly and successfully, leaders are required to participate in DSS construction meetings, monitor the implementation efforts, and provide clear directions for the project. Willingness to provide the necessary resources is another indicator of top management commitment to DSS implementation. The implementation could be blocked if some of the critical resources (e.g., people, funds and equipment) are not provided. (2) Financial resources 74
From the interviews, we can see that financial budget is the most serious problem facing the leaders in the Pengzhou government. Here, we would suggest that the financial resources should not only come from government budget, but also from other channels, such as, government bonds, savings from other projects or funds from NGOs or the central government. (3) Working process setting In order to avoid working conflicts between different departments, a simple and efficient working process should be settled upon before constructing a DSS. Under the working process analysis in chapter three, we construct an overall earthquake decision making working process including the following four stages: Information collection and processing Based on the contents of Decision information obtainment in chapter 3.3.5, earthquake related information can be collected comprehensively, and the development trend of the earthquake could also be estimated correctly. These are the foundational elements for the next steps. Problem and goal identification Identifying the problem requirees a comprehensive analysis of the status, properties, development trend and relief conditions of the earthquake related problems. After indentifying the problems, targeted goals would then be created. Emergency plan selection Emergency plans can come directly from the emergency plan base or can be formulized by adapting and adjusting some existing emergency plans. This phase is, under the help of DSS, a comprehensive evaluation of the emergency plans from which a solution for dealing with the earthquke will be chosen according to the seismic event classification and organization classification. Emergency plan implementation The previous three stages are carried out within the DSS. However this step should be achieved by people from different departments. Once the decision is formed, the necessary human, material and financial resources should be assembled, reasonably and effectively allocated, and the emergency plan should be carried out effectively. In addition, during the 75
process of implementation, information feedback and inspection must be strengthened, which would help improve the decision-making plan throughout implementation. (4) Government-wide commitment Since DSS is a government-wide information system that integrates information within and across all- administrative organizations, it is necessary to get support from all segments of the earthquake command center. Every person or organization is partially responsible for the overall system, and key users from different administrative organizations must commit to DSS implementation. Government-wide support should involve two aspects: first, leaders of each administrative organization should support the construction of the DSS; Secondly, as worker within each organization grow to understand the benefits of the system for their work, they will be willing to support the DSS. (5) Index systems setting As was discussed in the potential problems, the seismic event classification index, organization index and decision making evaluation index should be set before the construction of DSS. Once they are established, they need to be imported into the DSS databases, and then they can help achieve the DSS functions on classification and evaluation. (6) Effective project management DSS implementation is a set of complex activities, involving all different departments and often requiring between one and two years of effort (Kivij rvi & Zmud, 1993). Thus, the government could consider DSS construction as a project and use effective project management methods to control the implementation process (Zhang, et al., 2003), which would help avoiding budgetary overrun and ensuring implementation within schedule. According to Sum, C., Ang, J ., & Yeo, L. (1997), five major issues decision makers should take into consideration during the process of DSS construction are: firstly, a formal implementation emergency plan; secondly, a realistic time frame, thirdly, periodic project status meetings, fourth, an effective project leader, and finally, quality and cost control. A formal project implementation plan lists DSS construction activities and commits personnel to those activities. It is important to have a realistic time frame. If the target schedule was too short, the pressure to rush through would result in the implementation being carried out in a haphazard manner. On the other hand, if the implementation is delayed for too long, people 76
may lose patience, which would result in low morale and resistance. Conducting periodic project status meetings where project members report progress and problems is an invaluable means for evaluating the progress of DSS implementation. At the same time, project leaders must pay attention to the quality and cost of the DSS construction project. Selecting the right project leader is also important for the project implementation success. 5.2.2 People characteristics (1) User involvement User involvement refers to the participation of DSS users in the system development and implementation processes. The DSS system is designed to facilitate decision makers jobs. Without their participation, the system cannot be designed to satisfy their needs. According to Zhang, Lee, Zhang, & Banerjee (2003), the following two areas need user involvement: firstly, definition (demand analysis) in DSS requires user involvement; and secondly, user participation is needed in the implementation process of DSS. Involving users in the stage of defining DSS needs can decrease their resistance to the potential system. (2) Education & training Based on the definition of Sum, C., Ang, J ., & Yeo, L. (1997), education and training in this context refers to the process of providing users with the logic and overall concepts of DSS system and system operation. Thus, people can increase their awareness about using DSS and have a better understanding of how their decisions are made and to what extent the DSS system can help them. The main reason for education and training is to increase the consciousness of people who will use DSS as well as their expertise or knowledge level. Three important aspects of the training contents are: 1) logic and concepts of DSS; 2) features of the DSS (how to benefit users); and 3) operation training. Concept and features of DSS training show users why the DSS system is implemented, how the decisions are made and to what extent this system can help decision makers. Operation training helps those decision makers who are computer illiterate to operate the system. 5.2.3 Technical supports (1) Data accuracy Since the databases in DSS are intricately linked to one another, it is important to make sure the input data is accurate. Otherwise, inaccurate data input into one database will adversely affect the functioning of the others. This would result in the final decision which decision 77
makers get from DSS being unhelpful or even misleading. This would reduce the incentive to use the DSS system. Thus data accuracy is a major determinant of DSS system success (Markus, et al., 2000). (2) Construction and interoperability of software and hardware DSS vendors may use different hardware/software platforms, databases, and operation systems which are not compatible with upper-level governments databases and operation systems. In order to consider the future use of the DSS (communicate with upper lever government DSS), a hardware/software requirements analysis should be conducted first to explore potential hardware/software problems. The hardware/software then is selected according to the specific DSS requirements. As time goes by and environment changes, the DSS may not as fully meet the decision makers needs, especially when the government organization or the resource locations has some changes. Thus, to increase the chance of success, the hardware/software should be updated periodically in order to closely fit with decision makers requirements.
78
Chapter 6 Conclusions 6.1 Summary To investigate the architecture of an earthquake emergency management DSS in China local government, this paper introduced research background, research questions, research objective, research significance and research methods. Then relevant theoretical literature about emergency management and DSS were reviewed. After the literature review, documents analysis and interviews, this paper began to discuss the decision making functions of Beijings current emergency MIS and the deficiencies of current emergency management DSS in China. Combining the actual situation of Pengzhou with insights from this literature, decision support needs, objectives, and functions of DSS were defined and summarized ten functions were grouped into six categories. According to the six categories, the working process and data flow of each function were analyzed. After that, the overall framework of the DSS was designed and its components were analyzed. A four-layer general framework of DSS was established, and each layer was discussed in detail as were the interaction between different layers. Finally, the potential problems and possible solutions during the DSS implementation were briefly introduced. To be specific, the main conclusions of this paper include the following: (1) Through the interview and document analysis, a general idea about current situation of the DSS application in China was got. As the level of earthquake emergency management DSS in China is low, there are various opportunities for people to work on this field and develop such emergency management DSS to help deal with not only earthquake but also other nature disasters. (2) The decision support needs of DSS in Pengzhou were defined in chapter 3.2.1. The decision support needs were divided into three stages, which were the foundation of the DSS goal analysis and function analysis. (3) In chapter 3.2.3, 10 major functions of DSS were looked and summarized into six categories. While some decision support functions can be automatized, not every relevant function can be turned into a computer system and the personal judgment of decision makers remains crucial in these areas. (4) Combining the research on emergency management and decision support systems, in chapter 4, the working process and data flow of each function were analyzed, and then the 79
four-layer framework architecture of the DSS was designed. The contents of each layer and the relationships between different layers were analyzed. While the DSS architecture was built, as we mentioned before DSS can only help decision makers, and it cannot replace people, therefore decision makers judgment was needed as well. (5) Contents needed in various databases were summarized and the roles models play in each specific function were discussed. These databases and models were some basic components of the DSS, and should to be update according to the change of decision support needs. (6) Potential problems and solutions during the implementation process of the DSS were also addressed. Potential problems analysis is the precautionary measure for Pengzhou government, in case they can deal with some general problems during the DSS construction process. As new problems always occur during the process of implementation, it is necessary to prepare well the handle these new problems and then record them to help future DSS promotion across China. 6.2 Limitations of study This study has answered the research questions and achieved the research purposes. However, the study has some limitations that should be noted, such as: (1) Due to limited government documentation, the current analysis of the DSS in China is not sufficient, and must rely heavily on the study of the Beijing MIS. (2) The decision support need is mainly determined through the telephone interviews, which may not sufficient. If it is possible, face to face interviews and some field research (such as observe how decision makers make decisions during the earthquake, their working process and what kind of difficult they are facing when making decisions) are necessary as well. (3) The working process and data flow analysis of each decision-making function should be extended in further research. (4) Some parts, such as the model base design, cannot be analyzed in too great of detail, and this analysis could benefit from consultation with an expert or further literature in this area. (5) Because there is no such DSS in Chinese local government, the summary of DSS implementation problems is not enough and needs to be updated during the process of DSS construction. (6) It is difficult to realize this ideal DSS. On one hand, the IT level and peoples consciousness of DSS in local government is too low; on the other, China is a top-down 80
government, and local government has no financial budget to do this kind of project without the support of the central government. 6.3 Policy Recommendations It is possible to draw some policy recommendations based on this research. However, as discussed above, the policy recommendations may be insufficient due to the constraints of knowledge about potential problems. Thus, they need to be amended according to problems that emerge during construction. However, some general recommendations are summarized below: (1) From the view of organizational environments Firstly, DSS construction needs to be supported by leaders as they can create a good implementation environment by providing leadership and giving necessary resources. Secondly, leaders can consider multiple-channels to raise money for DSS construction, such as government bonds, savings from other projects and earmarked funding from NGOs or the central government. Thirdly, the working process of earthquake emergency management should be settled upon before construction. Fourth, it is necessary and important to get government-wide commitment, which means all the segment of the earthquake command center should support the DSS construction. Fifth, the seismic event classification index, organization classification index and decision making evaluation index need to be set before the construction of DSS. Finally, leaders should apply effective project management methods to manage DSS construction process. Five major issues decision makers should take into consideration are a formal implementation plan, a realistic time frame, periodic project status meetings, an effective project leader, and quality and cost control. (2) From the view of people characteristics Firstly, DSS users should participate in the development and implementation processes of DSS construction especially in the stages of identifying DSS definition (demand analysis) needs. Secondly, education and training are necessary. The aim of this education is to increase peoples consciousness about using DSS and offer them a better understanding of how their decision making process functions and to what extent the DSS system can help them. The training contents should at least include: logic and concepts of DSS, features of the DSS (How to benefit users), and operation training. (3) From the view of technical supports 81
Firstly, all input data must be accurate. Since the databases in DSS are linked to one another, inaccurate data will result in wrong decisions and mislead decision makers. Secondly, the hardware and software must be compatible with the upper level governments in order to further the use of DSS. The hardware/software must be updated according to the time and environment changes so that the DSS can continue to closely fit the needs of decision makers. 6.4 Further research During the process of this study, I deeply felt that the research on DSS architecture design is a multidisciplinary topic that bore significance bother practically and theoretically. It is worth continuing and extending this research. Specifically, I believe that the following would be the entry points for the further research: (1) Improving the functions of the DSS, as the current function analysis is only based on the interviews of 14 informants which may be sufficient to fulfill the decision needs during the DSS; (2) As time was limited, the working process and data flow of each function are not discussed in depth and could be further elaborated; (3) Because the classification index system, evaluation index system and working process need to be fixed before DSS construction, an in-depth study on them is necessary; (4) Further research on the data base and model base design focusing on its components, structure, and realization would be useful. (5) Research on the linkage, interaction and integration of each layer/subsystem can help to advance this DSS as well. (6) Last but not the least, the programming required to realize this system is also an important research issue. 82
Reference
Al-Mashari, M., & Zairi, M. ( 2000) . The Effective Application of SAP R/3: A Proposed Model of Best Practice. Logistics Information Management, 13( 3) , 156-166. Aleskerov, F., Iseri Say, A., Toker, A., Akin, H. L., & Altay, G. ( 2005) . A Cluster-Based Decision Support System for Estimating Earthquake Damage and Casualties. Disasters, 29( 3) , 255-276. Arnott, D., & Pervan, G. ( 2005) . A Critical Analysis of Decision Support Systems Research. Journal of Information Technology, 20( 2) , 67-87. Barbarosogcaronlu, G., & Arda, Y. ( 2004) . A Two-Stage Stochastic Programming Framework for Transportation Planning in Disaster Response. Journal of the Operational Research Society, 55( 1) , 43-53. Burns, O., & Turnipseed, D. ( 1991) . Critical Success Factors in Manufacturing Resource Planning Implementation. International Journal of Operations & Production Management, 11( 4) , 5-19. Burstein, F., & Holsapple, C. W. ( 2008) . Handbook on Decision Support Systems 1: Basic Themes. New York, USA: Springer. Carroll, J . ( 2001) . Emergency Management on a Grand Scale: A Bureaucrat's Analysis. Public Administration and Public Policy, 93, 463-480. Chen, X., & Xu, B. ( 2001) . The Application of Decision Support System Based on Layer Model Method. Journal of Central South University of Technology, 32( 1) , 97-100. Chen, X., & Yasuhiko, T. ( 2000) . A New Decision Support System Development Method-Layer Model Method. Systems Engineering, 4. Committee, P. M. ( 2008) . Appointment Notice of Earthquake Relief Command Headquater. Dai, J., Wang, S., & Yang, X. ( 1994) . Computerized Support Systems for Emergency Decision Making. Annals of Operations Research, 51( 7) , 313-325. 83
Danielsson, M., & Ohlsson, K. ( 1999) . Decision Making in Emergency Management: A Survey Study. International Journal of Cognitive Ergonomics, 3( 2) , 91-99. Densham, P. ( 1991) . Spatial Decision Support Systems. Geographical Information Systems: Principles and Applications, 1, 403-412. Donner, W. R. ( 2008) . Decision Making as Community Adaptation: A Case Study of Emergency Managers in Oklahoma. Disasters, 32( 2) , 292-302. Drabek, T., & Hoetmer, G. ( 1990) . Emergency Management: Principles and Practice for Local Government. International City Management Association, Washington, DC. Duchessi, P., Schaninger, C. M., & Hobbs, D. R. ( 1989) . Implementing a Manufacturing Planning and Control Information System. California Management Review, 31( 3) , 75-90. Eom, H., & Lee, S. ( 1990) . A Survey of Decision Support System Applications (1971-April 1988). Interfaces, 20( 3) , 65-79. Eom, S., & Kim, E. ( 2006) . A Survey of Decision Support System Applications (1995-2001). The Journal of the Operational Research Society, 57( 11) , 1264-1278. Eom, S., Lee, S., Kim, E., & Somarajan, C. ( 1998) . A Survey of Decision Support System Applications (1988-1994). Journal of the Operational Research Society, 49( 2) , 109-120. Fan, W. ( 2005) . Public Emergency Information System Architecture Design. Informatization Construction, 9. Feng, S., & Xu, L. ( 1999) . An Intelligent Decision Support System for Fuzzy Comprehensive Evaluation of Urban Development. Expert Systems with Applications, 16( 1) , 21-32. Flick, U. ( 2009) . An Introduction to Qualitative Research. Thousand Oaks, USA: Sage Publications Ltd. Gao, H. ( 2005) . Decision Support System: Theory, Method and Case ( 4 ed. ) . Beijing, China: Tsinghua University Press. Haddow, G., Bullock, J ., & Coppola, D. ( 2007) . Introduction to Emergency Management. Burlington, USA: Butterworth-Heinemann. 84
Hernandez, J ., & Serrano, J. ( 2001) . Knowledge-Based Models for Emergency Management Systems. Expert Systems with Applications, 20( 2) , 173-186. Inmon, W. ( 1995) . Tech topic: What is a data warehouse. Prism Solutions, 1. Inmon, W. ( 1996) . The Data Warehouse and Data Mining. Communications of the ACM, 39( 11) , 49-50. J iang, S. ( 2006) . It is necessary to formulate scientific evaluation system of pre-plan for emergency response and capability for emergency response in urban areas. City and Disaster Reduction 5. J ie, Y., Lei, J., & Hong, C. ( 2005) . Dynamic Game Analysis of Emergency Management. Management Review, 17( 3) , 46-50. Keen, P. ( 1987) . Decision Support Systems: The Next Decade. Decision Support System, 3( 3) , 253. Kivij rvi, H., & Zmud, R. ( 1993) . DSS Implementation Activities, Problem Domain Characteristics and DSS Success. Information Systems 2( 3) , 159-168. Kizza, J. M. ( 2005) . Computer Network Security. New York, USA: Springer. Laudon, K., & Laudon, J. ( 1995) . Management Information Systems: Organization and Technology. Upper Saddle River, NJ, USA: Prentice-Hall, Inc. Li, W., & Ding, M. ( 2008) . The Analysis of Urban Disaster Management in China. Current Construction, 10( 4) , 10-11. Liu, H., Li, L., & Luo, S. ( 2006) . The Study of the City Emergency Management Information System. Computer Knowledge and Technology, 7( 10) . Liu, S., & Ren, Y. ( 2009) . Study on Urban Emergency Management Information System. China Public Security, Academy Edition, 7( 4) , 44-58. Liu, S., & Zhang, Y. ( 2007) . Emergency Decision Supporting System for Community Safety in Medium-Sized City. Journal of Safety and Environment, 7( 2) , 140-143. Mao, Z., & Zhu, B. ( 2007) . Study on The Frame of Decision Support System for Emergency Plan of City. Geomatics and Spatial Information Technology, 30( 2) . 85
Markus, M., Axline, S., Petrie, D., & Tanis, S. ( 2000) . Learning from Adopters Experiences with ERP: Problems Encountered and Success Achieved. Journal of Information Technology, 15( 4) , 245-265. Mittra, S. S. ( 1986) . Decision Support System: Tools and Techniques. New York, USA: Wiley. Moosbruker, J., & Loftin, R. ( 1998) . Business Process Redesign and Organization Development: Enhancing Success by Removing the Barriers. The Journal of Applied Behavioral Science, 34( 3) , 286. Morrow, B. H. ( 1999) . Identifying and Mapping Community Vulnerability. Disasters, 23( 1) , 1-18. Morton, M. ( 1971) . Management Decision Systems: Computer-Based Support for Decision Making. Cambridge, USA: Harvard University Press. O'Brien, J., & Marakas, G. ( 2005) . Management Information Systems. New York, USA: McGraw-Hill. Pengzhou, P. s. G. o. ( 2006) . Pengzhou Public Emergency Response Plan (General). Pengzhou, P. s. G. o. ( 2007) . Pengzhou earthquake emergency plan. Power, D., & Sharda, R. ( 2005) . Model-Driven Decision Support Systems: Concepts and Research Directions. Decision Support Systems, 43( 3) , 1044-1061. Qi, M., Chi, H., Zhao, H., & Sun, Y. ( 2006) . A Research Review on the Public Emergency Management. Management Review, 4. Rahman, S., & Smith, D. ( 2000) . Use of Location-Allocation Models in Health Service Development Planning in Developing Nations. European Journal of Operational Research, 123( 3) , 437-452. ReVelle, C., & Eiselt, H. ( 2005) . Location Analysis: A Synthesis and Survey. European Journal of Operational Research, 165( 1) , 1-19. Schwartz, M., Bennett, W., & Stein, S. ( 1995) . Communication Systems and Techniques. New York, USA: Wiley. She, L. ( 2000) . System Reconstruction for Enterprise Organization: Forewarning Management and System Analysis. International Journal of General Systems, 29( 3) , 385-393. 86
Silver, M. S., Markus, M. L., & Beath, C. M. ( 1995) . The Information Technology Interaction Model: A Foundation for the MBA Core Course. MIS Quarterly, 19( 3) , 361-390. Sinha, R. ( 2005) . Impact of Experience on Decision Making in Emergency Situation. Unpublished C/D Extended Essay, Lulea University of Technology, Sweden. Sprague, R. ( 1980) . A Framework for the Development of Decision Support Systems. MIS Quarterly, 4( 4) , 1-26. Sprague, R., & Carlson, E. ( 1982) . Building Effective Decision Support Systems. New Jersey, USA: Prentice-Hall. Sum, C., Ang, J., & Yeo, L. ( 1997) . Contextual Elements of Critical Success Factors in MRP Implementation. Production and Inventory Management Journal, 38, 77-83. Wang, W. ( 2005) . Public Emergency Information System and Its Technology System. Informatization Construction, 9, 18-20. Wu, Z., & Liu, M. ( 2003a) . Emergency Rescue System and Plan Management of Major Accident. Beijing, China: Metallurgical Industry Press. Wu, Z., & Liu, M. ( 2003b) . Major Accident Emergency Plan Classification System and Its Contents. China Safety Science Journal, 1, 15-18. Zhang, L., Lee, M., Zhang, Z., & Banerjee, P. ( 2003) . Critical Success Factors of Enterprise Resource Planning Systems Implementation Success in China. Paper presented at the 36th Annual Hawaii International Conference on System Sciences (HICSS'03) - Track 8.
87
Annex 1: Research Consent Form 45
STUDY TITLE: Research on the Decision Support System (DSS) of Earthquake Emergency Management - Based on Wen Chuan Earthquake in China PURPOSE: The purpose of the study is to design an architecture of DSS for emergency management to help decision makers in local government and give some suggestion on how to implement the system. PROCEDURES: You will be asked with some open-ended questions during an around one hour telephone interview. COST, RISK, AND DISCOMFORT: There are no costs, risks or discomfort associated with this interview. BENEFITS TO YOU: The final result of the study will be presented back to relevant key informants for further discussions and improvement. It can also be used as the relevant document for the future information system construction of Pengzhou government. CONFIDENTIALITY: Only my supervisor and some relevant professors can get access of these answers. The result will be used only for writing a master thesis at Maastricht University, the Netherlands. If this study will be publishes in a scholarly journal or your comments will to be used as quotes, I will get your written permission. Otherwise, your name and other identifiers will not be used in any publication. REQUEST FOR MORE INFORMATION: You may ask more questions about the study at any time. Please e-mail abcivyduan@gmail.com with any questions or concerns about this study. WITHDRAWAL OF PARTICIPATION: Should you decide at any time during the study that you no longer wish to participate, you may withdraw your consent without prejudice. SIGNATURE: I confirm that the purpose of this research, the study procedures, the possible risks and discomforts as well as benefits have been explained to me. All my questions have been answered. I have read this consent form. My signature below indicates my willingness to participate in this study.
Signature: Date:
45 This is a translated version; in practice the Chinese version was used. 88
Annex 2: Interview question guide 46
Name: .................................................... Telephone: ...................................... Date of Interview: ...................................... Email: .............................................
Command and control tasks 1. What are your tasks/responsibilities in earthquake emergency management? 2. What kind of decisions you made in case of the earthquake? 3. What difficulties you faced when making decision? 4. How might making these decisions be eased? 5. What were your previous experiences with IT in emergency management or decision making? Information Sources 1. What were the relevant information sources/data that you had to use to make decisions in the Wen Chuan earthquake emergency management (or in other emergency management case) 2. How is the information transferred/processed among different decision makers? 3. What kind of shortcomings did you experience with the available information and its processing? 4. What are the possibilities for the improvement of information sources/channels? 5. What are your routines for documentation of earthquake emergency management? (Hint: by hand writing paper or by computer recording) Demand 1. How do you think that computer systems could assist your decision-making in earthquake emergency management? Implementation 1. Have you used another information system? Did you and your staff meet some difficulties while using that system? Was the problem solved? How did you solve the problem?
46 The question guide merely serves as the outline of interview. It is important to conduct the interview more flexibly- interviewer can probe more relevant questions according to the informants response, in order to get as much as information as possible.
89
2. How do you think the idea of constructing a DSS to help you make decision? ( Hint: to test whether they like it or not, and the possibility of construction) 3. If there is a DSS, what are the potential difficulties for you and your staff to implement (construct, build, install) this system? 4. If there is a DSS, what are the potential difficulties for you and your staff to use this system? 5. What do you think can facilitate the implementation of DSS? Other comments and suggestions
90
Annex 3: List of interviewees
Respondent ID Organization Type Function 1 Party Secretary Commander 2 Chief Disciplinary Officer of Chengdu Inspection leader 3 Mayor Vice-commander 4 Relief supply group Leader 5 Victims temporary resettlement group Vice-leader 6 Victims long-term resettlement group Leader 7 Medical treatment and anti-epidemic group Leader 8 Energy and transportation group Leader 9 Public relations group Vice-leader 10 Psychological counseling group Vice-leader 11 Rescue forces coordination group Leader 12 Inspection group Vice-leader 13 Statistics group Leader 14 Secretary group Leader
Note: "Your Bill Includes Arrears Due To Non Payment of Previous Bill. The Payment Towards Arrear Should Be Cleared Immediately To Avoid Disconnection"