CN105116795A

CN105116795A - Distributed control method for automatic manufacturing system with assembly operation

Info

Publication number: CN105116795A
Application number: CN201510435048.XA
Authority: CN
Inventors: 胡核算; 杨妍
Original assignee: Xidian University
Current assignee: Xidian University
Priority date: 2015-07-22
Filing date: 2015-07-22
Publication date: 2015-12-02
Anticipated expiration: 2035-07-22
Also published as: CN105116795B

Abstract

Provided is a distributed control method for an automatic manufacturing system with assembly operation. A change set composed of changes which can be enabled in a state M is detected first, and when detection of all the changes in a TEN is completed, a finally output TCN is a set of changes ensuring deadlock-free running of a system in the state M; any change is excited to obtain a new state, and a set of changes ensuring deadlock-free running of the system in the state is judged; and the steps are repeated, and finally, an emission sequence for ensuring deadlock-free running of the system is obtained. According to the invention, a Petri network is used as a mathematical tool to avoid an exhaustive state space, reduce the complexity of the algorithm and realize efficient control on a huge system. A dynamic, online and distributed control mode is adopted to observe and control the local state, thereby ensuring the deadlock-free performance of the whole system.

Description

Distributed control method for automatic manufacturing system with assembly operation

Technical Field

The invention belongs to the technical field of automatic manufacturing systems, and relates to a distributed control method for an automatic manufacturing system with assembly operation.

Background

In order to quickly respond to market changes and meet the emerging customer needs, scientists developed automated manufacturing systems. Generally, an automated manufacturing system is composed of a plurality of manufacturing processes that concurrently perform respective manufacturing tasks according to respective process recipes. During the execution of these processes, resources (e.g., buffers, material handling equipment, etc.) are often required to assist their completion. In practice, these resources are limited, which makes them necessary to be shared by these processes. However, such a processing environment is prone to deadlock. Once a deadlock occurs, the system will experience production stalls with serious and even catastrophic consequences. To solve the deadlock problem in an automated manufacturing system, it is common practice to introduce a corresponding controller to apply to a given system model, thereby avoiding undesirable behavior.

Over the past decades, a series of supervisory control methods have been developed. These supervisory control methods can be classified into a centralized control method and a distributed control method. In industry, actual automated manufacturing systems are intricate in scale. Each automated manufacturing system is comprised of relatively small scale, local, asynchronous, and interactive subsystems. Its execution is the result of the interaction of these subsystems. It is clear that with the increasing number of these subsystem modules, the problem of "state explosion" will be faced with the centralized control method, with the well-known computational and memory complexity problems. Based on the above problems, more and more researchers are turning their attention to the distributed control method. By adopting a distributed control method, the monitoring task completed by the central controller can be completed by a group of local controllers together. Each local controller receives observation information of the subsystem controlled by the local controller and makes local decisions according to the received local observation information, so that the calculation and storage complexity is greatly reduced.

The existing distributed control methods are summarized and analyzed, and defects exist in the process of wide popularization, and are particularly shown in the following point 1. due to the fact that the design of a controller becomes extremely complex due to an automatic system with an assembly operation, most of the existing distributed control methods are only limited to solving the deadlock problem of an automatic manufacturing system with a flexible path, and the automatic manufacturing system with the assembly operation is ignored. However, the assembly operation gives the system more structural information than a flexible path. The research result based on the automatic manufacturing system with the assembly operation is more general and can be popularized to a more complex automatic manufacturing system. 2. In these distributed control methods, many need to rely on a reachability graph or system architecture to complete the design of the controller. However, whether based on reachability maps or system architectures, it is difficult to apply these methods to large-scale automated manufacturing systems due to computational complexity issues.

Disclosure of Invention

To solve the problems of the prior art, it is an object of the present invention to provide a distributed control method for an automated manufacturing system with assembly operations that greatly reduces the amount of communication between the observer and the process and can be easily extended to more complex automated manufacturing systems.

In order to achieve the purpose, the technical scheme adopted by the invention is as follows:

a distributed control method for an automated manufacturing system having assembly operations, comprising the steps of:

1) given an F-AMG, M is an reachable state at the current time, T_EN＝{t₁，t₂…t_kDenotes the set of transitions that can be enabled in state M, T_CNA set of transitions representing transitions enabling deadlock free operation of the system in state M after the system is controlled; where k represents the number of transitions that can be enabled in state M;

2) initialization m 1, T_EN＝Φ，T_CNPhi is; wherein m is a counting variable;

3) from T_ENIn selecting transition t_mAnd transition t from arrival_mRandomly selecting one Turkan from one operation library, judging the position of the Turkan, if the Turkan is positioned in a parallel process between the shunting operation and the assembling operation, performing the step 4), otherwise, performing the step 5)

4) Under the condition that other tokens are all static, if the selected token can reach one sub-key library place in the parallel process of the selected token from the current position of the selected token, and one token exists in each of the rest parallel processes, and can reach one sub-key library place in the parallel process of the selected token from the current position of the selected token, further judging whether the tokens formed by combining the tokens through assembly operation can reach the nearest global key library place under the condition that other tokens are all static, and if so, changing to t_mCan be enabled, T_CN＝T_CN+{t_mStep 6) is performed, where m is m + 1; otherwise, performing step 6) if m is m + 1;

5) judging whether the selected token can move to the nearest global key library from the current position of the token under the condition that other tokens are static, and if so, changing t_mCan be enabled, and T_CN＝T_CN+{t_mStep 6) is performed, where m is m + 1; otherwise, performing step 6) if m is m + 1;

6) when m is less than or equal to k, performing the step 3);

7) when T is_ENAll transitions in (1) are detected, and T is finally output_CNNamely, the set of transitions which ensure the system to operate without deadlock in the state M; exciting any transition to obtain a new state, and judging a transition set capable of ensuring the system to run without deadlock in the state according to the steps 1) to 6); repeating the steps, and finally obtaining a transmitting sequence which ensures that the system runs without deadlock.

The specific process of the step 4) is as follows:

given aGroup of TorkenThe group of tokens respectively representing n subcomponents of token o in different parallel courses, andindicates the arrival from t_mA selected token in one of the operation libraries,respectively indicate at j₁A, j th₂… … th j_nThe number of the parallel processes is n;

4.1) given the set of local Key librariesWherein z represents the number of local key libraries contained in the set, and respectively indicate at j₁A, j₂… …, j_nSub-key library, N, in parallel processes_z＝{1,2……z}；

4.2) initializing l ═ 1, where l denotes a counting variable;

4.3) from the setIn selecting local key library Proceed to step 4.4)；

4.4) determining whether there are sufficient resources to support the selected Token in case other Tokens are stationaryMove from its current position to jth₁Sub-key library place in parallel processIf so, judging that the selected Token is the right oneAt sub-key libraryAnd in the case of the rest of the other tock, j₂Whether there is one Token in one parallel processCan proceed from its current position to jth₂Sub-key library place in parallel processIf so, judging that the selected Token is the right oneAt sub-key libraryJ th₂Selected Token in parallel courseAt sub-key libraryAnd in the case of the rest of the other tock, j₃Whether or not there is one Token in one parallel processCan proceed from its current position to jth₃Sub-key library place in parallel processIf yes, then judge the j₄A parallel process; repeating the steps until all parallel processes are traversed;

if the above conditions are met, performing step 4.6); otherwise, l ═ l +1, step 4.5) is performed;

4.5) when l is less than or equal to z, performing step 4.3), otherwise, performing step 6) when m is m + 1;

4.6) judging whether the selected token and the token selected in other parallel processes are combined through the assembling operation to advance to the global key library nearest to the token map module in which the token is positioned under the condition that other tokens are all static; if so, transition t_mCan be enabled, and T_CN＝T_CN+{t_mStep 6) is performed, where m is m + 1; otherwise, m is m +1, proceed to step 6).

Compared with the prior art, the invention has the beneficial effects that: the invention avoids exhaustion of state space, reduces algorithm complexity, realizes high-efficiency control of a system with large scale, and adopts a dynamic, online and distributed control mode to observe and control local state, thereby finally ensuring no deadlock of the whole system. Compared with the existing distributed control method, the distributed control method has the advantages that:

1. compared with the original distributed control method provided for the system with the structural characteristic of flexible manufacturing, the method provided by the invention has the advantages that the conclusion generated by the method is more general, and the method is a further extension of the original conclusion, so that the method can be easily expanded into a more complex automatic manufacturing system.

2. The invention adopts a real-time online control mechanism, and a controller is not required to be designed in advance. This means that in case of an emergency, for example, a resource failure, the method of the present invention can dynamically monitor the status, and then adjust the existing resource information in time and feed it back to the corresponding controller to make a correct control decision. Because a complicated controller design flow is avoided, the calculation amount is greatly reduced, and the distributed control method has higher efficiency.

3. The invention carries out deeper extension on the definition of the key library on the basis of the prior art. The invention further discloses a local key library which represents a combination of sub-key libraries of the same type belonging to different parallel processes, and the number of the sub-key libraries contained in the combination is equal to the number of the parallel processes. The invention further extends the concept of a global key library by representing all the process stages between the split operation and the assembly operation as virtual process stages such that each process step in the model of the invention is characterized by sequential execution.

4. The algorithm of the invention does not need to monitor global information, each step is executed only depending on whether the existing resources are enough, except for the process currently being executed, the invention does not need to know the states of other processes, so that the control method of the invention is a distributed control method, and the communication traffic between the observer and the process is greatly reduced.

Drawings

Fig. 1 is a Petri net model with an assembly operation.

Fig. 2 is a simplified diagram of fig. 1.

FIG. 3 is a signature block.

Detailed Description

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The invention provides a distributed control method based on an automatic system with assembly operation, which realizes active distributed control on line by applying a dynamic search method. The individual processes between the split operation and the assemble operation are referred to as parallel processes in the present invention.

First, the library sites in the F-AMG are classified into a key library site and a general library site. According to the definition of the key library in the paper "distributed Supervisory Synthesis for automated manufacturing systems Using Petri nets", if the various working phases of a process are executed sequentially, the present invention defines the working phase with the minimum or maximum resource requirement as the key library. In the F-AMG, the work phases are not completely connected in sequence, as described above, for the parallel process between the flow splitting operation and the assembling operation, the work phases in the same parallel process are connected in sequence, and the work phases between different parallel processes are connected in parallel. Taking F-AMG in FIG. 1 as an example, P₃And P₅Are connected sequentially, and P₃And P₄Are connected in parallel. However, because each parallel process has the characteristic of sequential connection of working phases, the invention extends the definition of the sub-key library.

Definition 1: given an F-AMG, when B (t)_s，t_e) When the following conditions are satisfied, it can be called a signature module: 1) | t_s ^·|≥2，|^·t_s|＝Φ；2)|t_e ^·|＝Φ，|^·t_e| ≧ 2, as shown in FIG. 3, where t_sRepresenting a split operation, t_eRepresenting assembly operations, splitting operations t_sAnd assembly operation t_eThe individual processes in between are referred to as parallel processes. Taking FIG. 3 as an example, t₁₁Representing a split operation, t₁₆Representative AssemblyOperation, t₁₁And t₁₆There are two parallel processes in between, respectively<p₁₁，t₁₂，p₁₂，t₁₃，p₁₃>，<p₁₄，t₁₄，p₁₅，t₁₅，p₁₆>。

Definition 2: given an F-AMG, p is any one of the libraries, and the demand of each resource by the library p can be expressed as a vector a_p＝[a_p(r₁),a_p(r₂)……a_p(r_L)]^TWherein L represents the number of classes of the resource, r₁，r₂，……r_LDenotes various types of resources in F-AMG, a_p(r₁)，a_p(r₂)……a_p(r_L) Respectively representing the p pairs of resources r of the library₁，r₂，……r_LThe required amount of (c).

Definition 3: assuming that there are N parallel processes between the flow-splitting operation and the assembling operation, N belongs to N⁺,Represents a set of all libraries in the ith parallel process, i ∈ N_nThen the key library can be defined as

1.

2.And p is_x＜p＜p_y. And,p_x＜p'＜p_yhas a_p≥a_p'}；

3.And p is_x＜p＜p_y. And,p＜p'＜p_yhas a_p≥a_p'}; wherein p is_x＜p＜p_yDenotes a library p_yBefore the library location p, the library location p is located in the library location p_xBefore; p is a radical of_x、p_yRepresenting two sub-key libraries in the same process.

According to the definition, the sub-key libraries can be divided into three classes, the first class represents the working phase with the minimum requirement on resources, and the F-AMG structure shows that the working phase with the minimum requirement on resources does not need resources (such as the initial library site and the end library site); the second category refers to two sub-key libraries p of the same parallel process_xAnd p_yThere is a repository with a demand ratio p for resources_xAnd p_yIf the resource demand of any library is large, the library is the sub-key library; the third category refers to two sub-key libraries p of the same parallel process_xAnd p_yThere is a repository that has a higher demand for resources than it does for the sub-key repository_yThe requirement of any library in between is large, and the library is also called a sub-key library.

With the definition of the sub-key library, the ordered combination of the sub-key libraries of each parallel process forms the local key library, so that the definition of the local key library is obtained by the method provided by the invention:

definition 4: given a set of sub-keysThen the local key repositoryCan be defined as:

1.

2.

wherein,respectively represent the j th₁A, j th₂… … th j_nA set of all libraries in a parallel process;

as shown in definition 4, the local key library is an ordered combination of n sub-key libraries belonging to different parallel processes and belonging to the same class of sub-key libraries.

When all the libraries between the flow operation and the assembly operation are combined into a virtual library, each process of the F-AMG has the characteristic of working state sequential connection. The specific combination method comprises the following steps: the input arc of each library in each parallel process to the resource library is moved to the transition position representing the assembly operation, the input arc of each library in each parallel process to the resource library is moved to the transition position representing the shunt operation, and the part between the shunt operation and the assembly operation is replaced by a virtual library, so that the simplified F-AMG has the characteristic of sequential connection of working stages. As shown in fig. 2, fig. 2 is a simplified model of the flag module in fig. 1. Taking FIG. 1 as an example, to merge the token map blocks B (t)₂，t₅) In the section between, t is required to be adjusted₃To p₁₅，t₄To p₁₅Is moved to transition t₅A is p₁₆To t₃，p₁₆To t₄，p₁₇To t₃，p₁₇To t₄Is moved to t₂Then p is added₃，t₃，p₅，p₄，t₄，p₆Using virtual places p_fic1Instead, as shown in fig. 2. Thus, the invention obtains the concept of a global library:

definition 5: given the F-AMG, the system is,all operation libraries in the ith process type after F-AMG simplification are represented, and the global key library of the operation libraries can be defined as

1.

2.And p is_x＜p＜p_y. And,has a_p≥a_p'}；

3.And p is_x＜p＜p_y. And,has a_p≥a_p'}；

As can be seen from definition 5, the global key library can be divided into three classes. The first type represents a library place with zero resource demand, and in the F-AMG, represents an initial library place and an end library place; the second type represents that in the simplified F-AMG, any given process is given if any two global key libraries p in the process_xAnd p_yThere is a repository with a demand ratio p for resources_xAnd p_yIf the resource demand of any library is large, the library is defined as a global key library; the third type represents that in the simplified F-AMG, any given process is given if any two global key libraries p in the process_xAnd p_yThere is a repository whose demand for resources is later than that of the global key repository_yAny library in between has a large demand for resources, and the library is defined as the global key library.

The present invention is directed to a distributed control method for an automated manufacturing system having assembly operations, comprising the steps of:

3) from T_ENIn selecting transition t_mAnd transition t from arrival_mOptionally selecting one of the operation libraries, judging the position of the token, and if the token is positioned in a parallel process between the shunting operation and the assembling operation, performing the step 4), otherwise, performing the step 5);

4) at rest in other tomkanIn this case, if the selected token can reach one sub-key library place in the parallel process of the selected token from the current position of the selected token, and one token exists in each of the remaining parallel processes, and can reach one sub-key library place in the parallel process of the selected token from the current position of the selected token, it is further determined whether the tokens formed by combining the tokens through the assembling operation can reach the global key library place closest to the selected token under the condition that other tokens are all static, and if so, transition t is performed_mCan be enabled, T_CN＝T_CN+{t_mStep 6) is performed, where m is m + 1; otherwise, performing step 6) if m is m + 1;

the specific process of the step 4) is as follows:

giving a set of TokenThe group of tokens respectively representing n subcomponents of token o in different parallel courses, andindicates the arrival from t_mThe selected one of the operation libraries,respectively indicate at j₁A, j th₂… … th j_nThe number of the parallel processes is n;

4.2) initializing l ═ 1, where l denotes a counting variable;

4.3) from the setIn selecting local key library Step 4.4) is carried out;

4.4) determining whether there are sufficient resources to support the selected Token in case other Tokens are stationaryMove from its current position to jth₁Sub-key library place in parallel processIf so, judging that the selected Token is the right oneAt sub-key libraryAnd in the case of the rest of the other tock, j₂Whether there is one Token in one parallel processCan proceed from its current position to jth₂Sub-key library place in parallel processIf yes, judging when selectedToken of choiceAt sub-key libraryJ th₂Selected Token in parallel courseAt sub-key libraryAnd in the case of the rest of the other tock, j₃Whether or not there is one Token in one parallel processCan proceed from its current position to jth₃Sub-key library place in parallel processIf yes, judging the jth item by a similar method₄A parallel process; repeating the steps until all parallel processes are traversed;

5) Determining that the selected Token is stationary when all other Tokens are stationaryWhether it can be moved from its current location to the global key store nearest to it, and if so, the transition t_mCan be enabled, and T_CN＝T_CN+{t_mStep 6) is performed, where m is m + 1; otherwise, performing step 6) if m is m + 1;

6) when m is less than or equal to k, performing the step 3);

7) when T is_ENAll transitions in (1) are detected, and T is finally output_CNNamely, the set of transitions which ensure the system to operate without deadlock in the state M; exciting any transition to obtain a new state, and judging a set of transitions capable of ensuring the system to run without deadlock in the state according to the steps 1) to 6); repeating the steps, and finally obtaining a transmitting sequence which ensures that the system runs without deadlock.

The following is a detailed description by way of an example.

Referring to fig. 1, fig. 1 is a Petri net model with an assembly operation, which is called feedbackauthored markedgraph, F-AMG for short. Each F-AMG is composed of multiple process types, denoted as J ═ J_j，j∈N_K}。

Definition 6: an F-AMG is a strongly connected, normal, pure Petri net N ═ (P, T, F, W), where N represents a model with a Petri net structure, P represents a place in the model, T represents a transition in the model, F represents a connection between the transition and the place, and W represents an arc weight of a connection arc between the transition and the place.

1)P＝P_B∪P_E∪P_A∪P_R:

a)P_B,P_E,P_A,P_RRespectively called a start library place, an end library place, an operation library place and a resource library place;

b)andwherein N is_KThe number of process types is {1, 2, …, K }, and, taking fig. 1 as an example, there are two process types, each of which is { p }₁，t₁，p₂，t₂，p₃，t₃，p₄，t₄，p₅，p₆，t₅，p₇，t₆}，{p₈，t₇，p₉，t₈，p₁₀，t₉，p₁₁，t₁₀，p₁₂，p₁₃，t₁₁，p₁₄，t₁₂}；

c) For arbitrary i e N_K,And isj∈N_K,i≠j，

d)P_R＝{r_i,i∈N_L}; wherein N is_LThe number of resource types is {1, 2, …, L }, and 3 resource types are shown in fig. 1 as an example, and are: r is₁(p₁₅)，r₂(p₁₆)，r₃(p₁₇)；

2)For each i ∈ N_K，And any i, j ∈ N_K,i≠j，TransitionRepresenting the beginning of a new manufacturing run, as exemplified in figure 1,

3) for each i ∈ N_KNetwork of networks Is a strongly linked label map, exceptEach loop includesAnd

taking FIG. 1 as an example, there are two processes, J₁＝{p₁，p₂，p₃，p₄，p₅，p₆，p₇}，J₂＝{p₈，p₉，p₁₀，p₁₁，p₁₂，p₁₃，p₁₄}. In process J₁In (1) { p₃，p₅And { p }₄，p₆Are composed of two parallel sub-processes, respectively, where the parallel process { p }₃，p₅The sub-key library corresponding to p₅Parallel process { p₄，p₆The sub-key library corresponding to p₆Thereby obtaining a local key library<p₅,p₆>(ii) a Similarly, in Process J₂Middle, largep₁₀，p₁₂And { p }₁₁，p₁₃Are composed of two parallel sub-processes, respectively, where the parallel process { p }₁₀，p₁₂The sub-key library corresponding to p₁₂Parallel process { p₁₁，p₁₃The corresponding sub-key library is p₁₃Thereby obtaining a local key library<p₁₂,p₁₃>。

To further obtain the global key library, the reduction should be simplified into the form of FIG. 2 according to the aforementioned reduction rules of the present invention, wherein p in FIG. 2_fic1Is { p₃，p₅，p₄，p₆The virtual library generated after simplification, p_fic2Is { p₁₀，p₁₂，p₁₁，p₁₃The virtual library generated after simplification can obtain the simplified process J according to the definition of the global library₁Global key library of { p }₁，p_fic1，p₇}, simplified Process J₂Global key library of { p }₈，p_fic2，p₁₄}。

Since the present invention dynamically and in real time generates event occurrence sequences, it is not difficult to see that the sequences are not unique and different search sequences will result in different event sequences. According to the distributed control method, one transmitting sequence sigma can be obtained, the system can be ensured to operate without deadlock, and sigma is equal to<t₇，t₁，t₈，t₉，t₂，t₃，t₁₀，t₁₁，t₄，t₅，t₇，t₁，t₈，t₉，t₂，t₃，t₁₀，t₁₁，t₄，t₅，t₆，t₁₂>。

The invention has the following advantages:

1. the invention provides a distributed control method for a system with the structural characteristic of assembly operation, which means that compared with the original distributed control method for the system with the structural characteristic of flexible manufacturing, the conclusion generated by the method is more general and is a further extension of the original conclusion, so that the method can be easily expanded to a more complex automatic manufacturing system. For example, a core key point of the implementation of the distributed control method proposed by the present invention is to determine a key library of the system, where the definition of the key library is given in the paper "distributed superior synchronization for automated manufacturing systems using petrinets", however, the definition of the key library is no longer applicable in a system having a structural feature of an assembly operation, because in a system having a structural feature of an assembly operation, the synchronicity between assembly processes is considered. Based on the characteristics, the invention extends the definitions of the sub-key library, the local key library and the global key library, and the definition of the key library given in the paper "distributed superior synthesis for automated manufacturing systems using petrinets" is only a special case of the invention.

3. On the basis of the original paper "distributedSupervisor Synthesis for automated manufacturing systems Using Petrenets", the definition of the key library is extended more deeply. The distributed Supervisory Synthesis for automated manufacturing systems Using PetriNet is a flexible path-based study, each process consisting of processing stages that are executed sequentially and not in parallel, in which the present invention defines a key library as the processing stage that has the smallest or the largest demand for resources. In the model of the invention, however, the machining stages in each machining pass are not performed completely sequentially. When a split operation is encountered, the part being machined is split into several sub-components for machining in different sub-processes until an assembly operation is encountered, and the sub-components are assembled into a new part. For these sub-processes between the splitting operation and the assembling operation, the processing stages in the same parallel process are executed sequentially, and the processing stages in different parallel processes are executed in parallel. Because the processing stages in the parallel process are sequentially executed, the present invention can obtain the definition of the sub-key library, and in the corresponding parallel process, the processing stage with the minimum or maximum resource requirement is called as the sub-key library in the parallel process. The execution of the assembly operation requires that the parallel processes be synchronized, i.e. the assembly operation cannot be executed as long as one of the parallel processes has not yet ended. Based on this, the present invention extends to the definition of the local key library, which represents a combination of sub-key libraries of the same type belonging to different parallel processes, and the number of sub-key libraries included in the combination should be equal to the number of parallel processes. The invention further extends the concept of a global key library by representing all the process stages between the split operation and the assembly operation as virtual process stages such that each process step in the model of the invention is characterized by sequential execution.

The invention applies the Petri network as a mathematical tool, avoids exhaustible state space, reduces the complexity of the algorithm, realizes the high-efficiency control of a system with large scale, adopts a dynamic, online and distributed control mode to observe and control the local state, thereby finally ensuring the no deadlock of the whole system.

Claims

1. A distributed control method for an automated manufacturing system having assembly operations, comprising the steps of:

2) the initialization m is equal to 1 and the initialization is carried out,T_EN＝Φ，T_CNphi is; wherein m is a counting variable;

3) from T_ENIn selecting transition t_mAnd transition t from arrival_mRandomly selecting one token from one operation library, judging the position of the token, and if the token is positioned in a parallel process between the shunting operation and the assembling operation, performing the step 4), otherwise, performing the step 5);

6) when m is less than or equal to k, performing the step 3);

2. The distributed control method for the automatic manufacturing system with the assembling operation as claimed in claim 1, wherein the specific process of the step 4) is as follows:

giving a set of TokenThe group of tokens respectively representing n subcomponents of token o in different parallel courses, andindicates the arrival from t_mA selected token in one of the operation libraries,respectively indicate at j₁A, j th₂… … th j_nThe number of the parallel processes is n;

4.2) initializing l ═ 1, where l denotes a counting variable;

4.3) from the set l_LOCIn selecting local key library Step 4.4) is carried out;