CN102332122A - Optimization method for the layout of urban public bicycle rental points - Google Patents

Abstract

The invention relates to a layout optimization method for urban public bicycle rental stations. The method comprises the following steps of: firstly, establishing basic databases for regional lands, residential structures, trip modes and the like; then dividing traffic zones, and performing trip generation prediction and trip distribution prediction; and finally establishing a two-layer model which consists of an adaptive genetic algorithm and a mode sharing and traffic distribution combined feedback model for different layout schemes, solving, evaluating scheme results, and obtaining an optimal layout scheme after the final convergence of evaluated values to make regional trip cost and the construction cost of public bicycle system facilities the lowest. Compared with the prior art, in the method provided by the invention, the trip needs of residents and the supply of traffic facilities are coordinated, and the method has a wide application range and a high quantification extent, and can provide a scientific technical support for related decision making.

Description

Optimization method for the layout of urban public bicycle rental points

technical field

The invention relates to a layout method of urban public bicycle rental points, in particular to the optimization of the layout of existing public bicycle rental points, and belongs to the field of traffic planning. the

Background technique

As a healthy, energy-saving, and environmentally friendly way of travel, public bicycles bring convenience to short-distance travel on the one hand, and expand the passenger flow of public transportation stations through transfer (B+R) on the other hand. The distance travel mode structure saves road resources. Practice has shown that the successful development of public bicycles is inseparable from correct planning, operation and publicity guidance. The unreasonable planning and layout often lead to an unbalanced allocation of resources at each rental point, resulting in lower service levels, loss of customer sources, and loss of business and ultimately failure to operate. . The existing research on the layout of public bicycle rental points is mainly based on macroscopic and qualitative analysis, mainly focusing on layout principles, feature analysis and scale prediction, and less research on quantitative models. The existing results lack consideration of the feedback between the public bicycle travel demand and the facility supply, and the model needs a large amount of survey data support, and the implementability is not strong. the

Contents of the invention

In order to overcome the deficiencies of the existing methods, the present invention sets up an optimization model for the layout of urban public bicycle rental points based on double-layer planning from the perspectives of residents’ travel needs and transportation facilities supply, researches the model solution, and provides a scientific basis for the decision-making of the layout of public bicycle rental points . the

The concrete scheme that the present invention adopts is as follows:

A method for optimizing the layout of urban public bicycle rental points, comprising the following steps:

1) First establish the database for the lower-level prediction model and the upper-level program evaluation model;

2) Travel generation and travel distribution for demand forecasting;

3) Scheme simulation and evaluation, and obtain the optimal scheme after judging convergence. the

in:

The database establishment process in described step 1) is as follows:

11) Firstly, determine the scope of planning, and select an alternative location for public bicycle rental points according to the current situation and the type of planned land use; secondly, based on the survey data of residents’ travel, analyze the socio-economic indicators, and calibrate the relevant parameters of different travelers, including age, Income structure and the ownership rate, cost and comfort of each travel mode;

12) Clarify the site location and service scope of conventional buses and rail transit in the area, the departure frequency of each public transport line, and the average occupancy rate of public transport entering the area for the first time. the

The trip generation and distribution prediction process in the described step 2) are as follows:

21) Divide the internal and external traffic areas according to the type of land use, and count the population and number of posts in each area;

22) Travel generation is based on the land use structure, using the cross-classification method to predict travelers of different ages and travel purposes, and the prediction time is the morning peak;

23) The travel distribution adopts the double-constraint gravity model, and considers the distribution of external travel separately. the

The scheme simulation, evaluation and final scheme determination process in step 3) are as follows:

31) The initial plan set is the layout plan of n public bicycle rental points defined by the population size randomly generated by the computer, making it the first generation of the population;

32) Carry out mode division and traffic allocation for each scheme according to the solution method of the lower-level model, and the total travel cost obtained from this will be fed back to the upper-level model for evaluation after being corrected by the penalty function;

33) If the evaluation result converges or reaches the maximum algebra of the population, transfer to the next step, otherwise, perform selection, crossover and mutation in the adaptive genetic algorithm for each plan; update the plan set, transfer to the next generation, and return to step 32);

34) Select the solution with the best evaluation result as the final solution. the

On the layout optimization of public bicycle rental points, the present invention adopts a two-layer programming model to model two different target groups of government planning departments and travellers. the

(1) Lower model

The lower model is used to describe the choice of traveler's mode and route on the existing network, that is, mode sharing and traffic allocation. The network used to share traffic distribution in a joint manner is a super-network consisting of a public transport network, a road network, a bicycle network, and a pedestrian network. In this network, each traveler will choose the travel mode and route with the most economical generalized cost. After each allocation, the impedance should be updated according to the road flow and fed back to the mode allocation model. the

In terms of mode sharing, since there are many travel modes and multi-mode transfers are involved, in order to avoid the IIA (correlation independence) characteristics and preference randomness limitations of the simple Logit (logit) model, the present invention adopts mixed Logit (mixed logit) model. Logarithmic) model. the

In terms of traffic allocation, each traveler will choose the mode and route with the lowest travel cost between the origin and destination. When the network reaches equilibrium, travelers cannot reduce travel costs by changing travel modes and routes. The user balance model achieves the theoretical optimum, but in practice, users often make decisions based on random travel costs, so the stochastic user balance model (SUE) is used. the

(2) Upper model

When planning the layout of public bicycle rental points, government departments need to choose the optimal layout plan according to the travel demand characteristics of the area, so as to minimize the total travel cost and facility construction cost. At the same time, according to the principle of the layout of public bicycle rental points, three constraints are established: rental point spacing, density and usage. the

For each layout scheme, the lower-level model obtains the allocated traffic according to the scheme, and feeds back the result to the upper-level model for evaluation, checks whether the constraints are met, and screens and optimizes the scheme set. The specific method is as follows:

For the lower model, multi-mode transfer is considered, and a super network is established for mode division and traffic allocation. The solution method is:

Step 1: Create a network. Establish a super network according to the layout of public bicycle rental points;

Step 2: Calculate the impedance and find the shortest path of OD (origin-destination) pairs for each single mode. A feasible multi-mode shortest path is found based on the result of the single-mode shortest path, and the floyd (Floyd) algorithm is used. the

Step 3: Method division. the

Step 4: Traffic Allocation. The continuous average method is used for allocation, and each iteration of the allocation is calculated using the STOCH (stochastic multi-path traffic allocation) algorithm. the

For the solution of the upper model, a reasonable and effective solution is constructed according to the constraints, and the results are fed back to the lower model. In the present invention, the self-adaptive genetic algorithm is selected to solve the upper model, the layout scheme is represented by discrete binary gene data, and the scheme is further explored through gene selection, crossover and mutation; for schemes that do not meet the constraint conditions, the fitness value penalty is used to reduce its probability of being selected. the

Compared with the prior art, the present invention can plan and evaluate the layout of public bicycles based on the data based on the results of the travel survey of current regional residents or the travel demand forecast in the target year, coordinate the travel demand and facility supply, and coordinate the travel needs of travelers, The interests and needs of the government, the planning department, and the business unit; at the same time, it is supported by a mature four-stage traffic forecasting model, which is highly scientific and applicable. the

Description of drawings

Fig. 1 is a flow chart of the layout optimization method for urban public bicycle rental points;

Figure 2 is a regional abstract road network diagram;

Figure 3 is a diagram of the evolution process;

Figure 4 is a diagram of the optimal layout scheme. the

Detailed ways

As shown in Figure 1, it is a flow chart of the city public bicycle rental point layout optimization method in this specific embodiment, and the method includes the following steps:

Step 1: Data Preparation

Firstly, determine the scope of planning, and select alternative locations for public bicycle rental points according to the current situation and the type of planned land use; secondly, based on the survey data of residents' travel, analyze social and economic indicators, and calibrate the age, income structure and travel modes of different travelers The ownership rate, speed, cost, comfort and other parameters. the

Clarify the station locations and service scopes of conventional buses and rail transit in the area, the departure frequency of each public transport line, and the average occupancy rate of public transport entering the area for the first time. the

Build a super network, including road network, regular bus network, rail transit network, bicycle network and pedestrian network. the

The internal districts are divided according to the land use, and the external traffic districts are divided according to the travel direction and travel distance. Clarify the population and job distribution of each district. the

Assuming that a public bicycle rental point is planned to be built in a residential area in a city with an area of about 3 square kilometers, residential areas, schools, shopping plazas, parks and other main passenger flow distribution points are all candidate locations. There is 1 track line, 2 stations (blocks), 6 conventional bus lines and 26 stations distributed in the area. Composed of traffic and regular bus lines (dotted lines), as shown in Figure 2. the

See Table 1 for the connection relationship between public transportation and external communities. The technical indicators of each travel mode are shown in Table 2. the

Table 1 Connection relationship between public transportation and external communities

Table 2 Technical indicators of travel mode

Considering the impact of transit traffic, it is assumed that the base saturation of the trunk road is 0.6, that of the secondary trunk road is 0.5, and that of the branch road is 0.2. Peak hour travel accounted for 12.5 percent of the day. Referring to the construction and operation of public bicycle systems in Paris, Shanghai, Hangzhou, Wuhan, etc., the minimum distance between public bicycle rental points in residential areas is set to 150m, and the density range is 2 to 4 bikes/km2. A single rental point rents and returns bicycles in the morning and peak hours The lower limit of the sum is 10 vehicles, the construction cost is 200,000 yuan, the vehicle purchase cost is 400 yuan/vehicle, and the service life is 5 years. the

Step Two: Demand Forecasting

Based on the theoretical basis of the four-stage traffic method, the generation and distribution of residents' trips are predicted. Here, it is necessary to divide the travel into regional internal travel and regional external travel according to the planning area, and at the same time consider the impact of transit traffic on the traffic saturation of road sections in the region. The prediction results are shown in Table 3. the

Table 3 Morning peak passenger flow OD

Step 3: Scheme simulation and evaluation

The scheme set is the layout scheme of n public bicycle rental points defined by the random population size generated by the computer. Each scheme is divided into modes and traffic distribution according to the lower model solution method proposed by the present invention, and the total travel cost drawn therefrom is passed through After the penalty function is modified, it is fed back to the upper model for evaluation. According to the evaluation results, the adaptive genetic algorithm is used to solve each plan, and the selection, crossover and mutation of genes are performed to update the plan set and transfer to the next generation. the

The selection of genes adopts the roulette wheel selection strategy (roulette wheel selection), that is, the selection probability of individuals is related to fitness, and individuals are selected at random probability to be retained in the next generation population. The incremental coefficient of the penalty function is 7.5, the initial crossover probability of the gene is 0.9, and the mutation probability is 0.04. the

Use Matlab programming to implement the upper and lower layer algorithms of the bilevel programming model, and apply it to this example. After 100 generations of evolution, the final optimization scheme was obtained. The evolution process is shown in Table 4, and the evolution process is shown in Figure 3. the

Table 4 Evolution process table

evolutionary algebra optimal individual optimal fitness average fitness 1 00111100100001001110 33978 70632 2 11110101000000011111 32075 57104 3 01100001110010011110 31084 50365 4 11110101001100011110 29061 49386 ... ... ... ... 98 00111100011011110000 23967 24114 99 00111100011011110000 23967 23995 100 00111100011011110000 23967 23981

Step 4: Determine the final plan

The simulation and evaluation of the program set are repeated until the evaluation value converges or reaches the maximum population algebra, and the program with the optimal evaluation value is selected as the final program. the

In Figure 3, the optimal fitness of the population represents the fitness obtained by the optimal solution of each generation, and the average fitness of the population represents the degree of evolution. The evolution to the 30th generation is basically stable, and the optimal solution is shown in Figure 4. the

Claims (5)

1. city public bicycles lease point layout optimization method is characterized in that, this method may further comprise the steps:

1) at first sets up the database that is used for lower floor's forecast model and upper strata evaluate alternatives model;

2) carry out the trip generation of demand forecast and the distribution of going on a journey;

3) program simulation and evaluation obtain optimal case after judging convergence.

2. city according to claim 1 public bicycles lease point layout optimization method is characterized in that the database creation process in the described step 1) is following:

11) at first, confirm planned range,, select public bicycles lease point alternate location according to present situation and planned land use type; Secondly, be the basis, analyze socio-economic indicator, demarcate the correlation parameter of different travelers, comprise owning rate speed, expense, the comfort level of age, structure of earnings and each trip mode with the resident trip survey data;

12) site location, the service range of conventional public transport and track traffic in the clear and definite zone, the frequency of dispatching a car of each public transport line

Rate, and the public transport in entering zone is on average taken advantage of the rate of carrying first.

3. city according to claim 1 public bicycles lease point layout optimization method is characterized in that described step 2) in trip generation and forecast of distribution process following:

21) divide inside and outside traffic zone based on the land used type, add up population and post number in each sub-district;

22) trip generates according to using ground structure, adopts the cross division method, the traveler of all ages and classes, trip purpose is predicted respectively predicted time is a morning peak;

23) trip distributes and adopts two constraint Gravity Models, and considers the externally distribution situation of trip separately.

4. city according to claim 1 public bicycles lease point layout optimization method is characterized in that the program simulation in the described step 3), evaluation and final plan deterministic process are following:

31) the initial scheme collection is to produce definition n public bicycles lease point placement scheme of population size down by computer random, and making it is the first generation of population;

32) each scheme is divided and the traffic distribution according to lower floor's model solution method mode of carrying out, the total trip cost that draws thus is through feeding back to layer model evaluation after the penalty correction;

33) if evaluation result restrains or reaches the maximum algebraically of population, be transferred to next step, otherwise, scenarios is carried out selection, intersection and variation in the self-adapted genetic algorithm; The update scheme collection is transferred to the next generation, returns step 32);

34) select the optimum scheme of evaluation result, as final plan.

5. city according to claim 4 public bicycles lease point layout optimization method is characterized in that said lower floor model solution method comprises the steps:

Step 1: create network,, set up supernet according to public bicycles lease point scheme layout;

Step 2: computing impedance is also sought the shortest path of OD to each folk prescription formula, and the result seeks feasible multimode shortest path based on folk prescription formula shortest path, adopts the floyd algorithm;

Step 3: mode is divided;

Step 4: traffic distributes, and adopts straight average method to distribute, and branch timing iteration each time all adopts the STOCH algorithm to calculate.

Images