CN103294823B - Rail transit multi-mode optimal transit transfer inquiring method based on cultural ant colony - Google Patents

Abstract

The invention relates to a multi-mode optimal transfer query method for rail transit based on a cultural ant colony system, comprising the following steps: 1) a central processor receives a query request through a touch screen, and obtains station information from a database according to the query request, and constructs a route Select the model; 2) The central processor executes the cultural ant colony system based on the path selection model, calculates and obtains the optimal rail transit transfer scheme under different optimal goals, and outputs the optimal path; 3) Updates the value of the path selection model, And judge whether the optimization is over, if so, feed back the calculation result to the touch screen, and execute step 4), if not, return to step 2); 4) the touch screen displays the calculation result. Compared with the prior art, the present invention uses the cultural ant colony system to quickly and accurately calculate the optimal rail transit transfer scheme under different optimal objectives, thereby improving the flexibility and efficiency of residents' travel to rail transit transfers.

Description

Multi-mode Optimal Transfer Query Method for Rail Transit Based on Cultural Ant Colony System

technical field

The invention relates to a calculation method for optimal transfer of rail transit, in particular to a multi-mode optimal transfer query method for rail transit based on a cultural ant colony system.

Background technique

The urban rail transit system is one of the links most closely related to the daily life of urban residents, and even determines the lifestyle of urban residents to a certain extent. Therefore, the electronic map products of many cities nowadays focus on the optimal route query of the rail transit network. As its top priority, it is hoped that the electronic map can better meet the needs of users, but the existing query system is not only error-prone, but also inefficient, and it cannot perform multi-modal transfers:

On the one hand, most software developers believe that the optimal path analysis of rail transit network, like other network analysis, should also be based on the shortest path, but the user’s optimal path is not only the shortest path, so it is far from the user’s requirements. big difference;

On the other hand, most users believe that the least number of transfers is the key issue. The least transfer and the shortest path seem to be unified, but they are not. Therefore, how to achieve the unity of the two and propose a realistic and feasible optimized transfer model and algorithm have become urgent problems to be solved.

Contents of the invention

The purpose of the present invention is to overcome the above-mentioned defects in the prior art and provide a multi-mode optimal transfer query method of rail transit based on the cultural ant colony system with fast calculation speed and high precision, which improves the impact of residents' travel on rail transit transfer. multiplied flexibility and efficiency.

The purpose of the present invention can be achieved through the following technical solutions:

A multi-modal optimal transfer query method for rail transit based on a cultural ant colony system, the method comprising the following steps:

1) The central processing unit receives the query request through the touch screen, and obtains site information from the database according to the query request, and builds a path selection model;

2) The central processing unit executes the multi-population cultural ant colony system based on the path selection model, calculates and obtains the optimal rail transit transfer scheme under different optimal goals, and outputs the optimal path;

3) Update the value of the path selection model, and judge whether the optimization is over, if so, feed back the calculation result to the touch screen, and execute step 4), if not, return to step 2);

4) The touch screen displays the calculation results.

The query request includes a starting site and a final site.

The optimal goals include the shortest time, the least transfer and the least distance.

The cultural ant colony system includes the ant colony evolution process in the group space and the knowledge update process in the belief space, and the ant colony evolution process in the group space includes the following steps:

a1) Initialize the pheromone distribution of the group space, and divide the group space into multiple subgroups, and each subgroup adopts the ant colony system with different behaviors for parallel evolution, and obtains the local optimal solution of each subgroup;

a2) The subgroups update their respective local pheromones according to the information interaction strategy based on the learning mechanism;

a3) Update the global optimal solution according to the local optimal solution of each subgroup, and store it in the belief space through the acceptance function;

a4) Update the global pheromone according to the output of the belief space;

a5) judging whether the algorithm termination condition is satisfied, if satisfied, the algorithm is terminated; otherwise, go to step a2);

The knowledge update process of the belief space includes the following steps:

b1) Initialize the belief space;

b2) Receive the current global optimal solution provided by the group space through the receiving function;

b3) Implement 2-OPT operation on the belief space to optimize the belief space;

b4) Output the optimal solution and provide it to step a4) through the influence function.

Each of the subgroups adopts the ant colony system with different behaviors to perform parallel evolution, specifically as follows:

a101) Each subgroup randomly places m ants of different numbers on one of the n sites;

a102) Each subgroup performs state transition according to its own behavior, selects the next node, and updates the local pheromone at the same time, and the behavior includes random, herd, greedy or mixed;

a103) Step a102) is repeated until each ant forms a complete path, that is, each subgroup traverses all nodes respectively to obtain their respective local optimal solutions.

The information interaction strategy based on the learning mechanism is:

Each subgroup performs information interaction with the other two adjacent subgroups, compares the current local optimal solution with the local optimal solutions of the other two adjacent subgroups, and updates its own local optimal solution with a better local optimal solution. Pheromones.

Described acceptance function Accept () is:

Accept()=T

T is a set constant.

The implementation of the 2-OPT operation on the belief space is as follows:

b301) set r ₀ as a given constant in [0, 1], generate a random number r in the range of [0, 1], if r>r ₀ then go to step b4);

b302) If there are nodes c _i and c _j in the current optimal path, where j≥i+2, and

d(c _i , c _i+1 )+d(c _j , c _j+1 )＞d(c _i , c _j )+d(c _i+1 , c _j+1 )

Then replace the edges (c _i , c _j ), (c _i+1 , c _j+1 ) with (c _i , c _i+1 ), (c _j , c _j+1 ), and the path in the line after exchange ( c _j ,..., c _i+1 ) are reversed; otherwise go to step b4).

The influence function Influence() is:

Among them, EndStep is the preset maximum evolution algebra of ant colony system, CurrentStep is the current algebra of ant colony evolution, BaseNum and C are constants.

Compared with the prior art, the present invention has the following advantages:

1. The present invention uses the cultural ant colony system to solve the optimal path. The cultural ant colony system is a new efficient optimization method that incorporates the ant colony system into the cultural algorithm framework. The calculation model includes the group space based on the ant colony system and the The belief space of the current optimal solution, the two spaces have their own groups and evolve independently and in parallel, which improves the speed and accuracy of the algorithm solution;

2. The present invention adopts the ant colony system of multi-population parallel evolution, and each sub-group interacts through the information interaction strategy based on the learning mechanism, which improves the accuracy of the algorithm;

3. The belief space of the cultural ant colony system of the present invention adopts random 2-OPT exchange operation to mutate and optimize the optimal solution, and the evolved solution individuals are used to update the global pheromone of the group space to help guide the evolution process of the group space , so as to achieve the purpose of increasing the diversity of the population, preventing premature maturity, and reducing the calculation cost;

4. The method of the present invention has better accuracy and robustness. Even for large-scale problems, a satisfactory solution with relatively small errors can be obtained with a small number of populations and a short running time.

In a word, the beneficial effects of the present invention are: the present invention designs and implements a new type of intelligent computing method, which can efficiently optimize the design of multi-modal transfers for different residents, and improves the flexibility of residents' travel to rail transit transfers. Efficiency. The invention adapts to the future demand of urban traffic development, and can carry out scientific management on rail transit transfers with increasing scale and complexity in a sustainable manner.

Description of drawings

Fig. 1 is a schematic flow sheet of the present invention;

Fig. 2 is the system frame diagram of multi-population cultural ant colony system of the present invention;

Fig. 3 is a schematic diagram of the framework of multi-population parallel evolution of the present invention;

Fig. 4 is a schematic diagram of the principle of the cultural ant colony system of the present invention.

detailed description

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is carried out on the premise of the technical solution of the present invention, and detailed implementation and specific operation process are given, but the protection scope of the present invention is not limited to the following embodiments.

As shown in Figure 1, a multi-modal optimal transfer query method for rail transit based on the cultural ant colony system, the method includes the following steps:

1) The central processor receives the query request through the touch screen, including the starting site and the final site, and obtains site information from the database according to the query request, and builds a path selection model;

2) The central processing unit executes the multi-population cultural ant colony system based on the path selection model, calculates and obtains the optimal rail transit transfer scheme under different optimal goals (including the shortest time, the least transfer and the least distance, etc.), and outputs the optimal route ;

3) Update the value of the path selection model, and judge whether the optimization is over, if so, feed back the calculation result to the touch screen, and execute step 4), if not, return to step 2);

4) The touch screen displays the calculation results and the route selection model, and the display content includes rail transit transfer information and line diagrams.

As shown in Figures 2-4, the cultural ant colony system includes the ant colony evolution process in the group space and the knowledge update process in the belief space, and the ant colony evolution process in the group space includes the following steps:

a1) Initialize the pheromone distribution of the population space, and divide the population space into multiple subgroups (such as subgroup 1, subgroup 2, subgroup 3, and subgroup 4 in Figure 2 and Figure 3), each subgroup will be different A number of m ants are randomly placed on one of the n sites, and the state transition is performed according to their respective behaviors, the next node is selected, and the local pheromone is updated at the same time. The behaviors include random, herd, greedy Or mixed; until each ant forms a complete path, that is, each subgroup traverses all nodes separately to obtain their own local optimal solutions.

The specific steps for each subgroup to evolve are as follows:

a101) Initialization: t=0, Nc=0, τ _ij (t)=τ ₀ , Δτ _ij (t)=0, m ants are randomly placed on n sites;

a102) set the taboo table index s=1, and add its starting point site in the taboo table respectively, judge whether the taboo table is full, if so, then execute step a104), if not, then s=s+1, execute step a103) ,

a103) Transition probability calculated by each ant Select the next site, and add this site to the taboo table, and update the local pheromones at the same time:

τ _ij ＝(1-ρ)τ _ij +ρτ ₀

Among them: ρ( ₀ <ρ<1) is the local volatilization factor of pheromone; τ0 is the initial pheromone concentration value on each path;

a104) Each ant forms a complete path, that is, traverses all nodes, calculates the travel length L _k traveled by all ants, updates the current optimal solution, and obtains a local optimal solution.

a2) The subgroups update their respective local pheromones according to the information interaction strategy based on the learning mechanism.

As shown in Figure 3, each subgroup is connected in a ring, each subgroup performs information interaction with the other two adjacent subgroups, and compares the current local optimal solution with the local optimal solution of the other two adjacent subgroups , and update its own local pheromone with a better local optimal solution.

a3) Update the global optimal solution according to the local optimal solution of each subgroup, and store it in the belief space through the acceptance function;

Described acceptance function Accept () is:

Accept()=T

T is a set constant, which can be set to 20;

a4) Update the global pheromone according to the output of the belief space:

in: is the global volatilization factor of pheromone,

_Lgb represents the path length of the current global optimal solution (the length of the global optimal path obtained from the beginning of the experiment);

a5) Judging whether the algorithm termination condition is satisfied, if satisfied, the algorithm is terminated; otherwise, all taboo lists are cleared, and then step a103):

The termination condition of the algorithm is to reach the set maximum evolution algebra or the global optimal solution does not change continuously within the set algebra.

The ant colony system used in each subgroup space is a kind of pheromone update on the path using local update rules and global update rules of pheromone, so that the search space of the expanded algorithm and the convergence of the algorithm can be organically unified. As shown in Figure 3, A indicates that each subgroup updates the local pheromone through local cooperation and communication of the optimal solution, and B indicates that each subgroup interacts and provides the global optimal solution to the belief space to update the global pheromone.

The knowledge update process of the belief space includes the following steps:

b1) Initialize the belief space;

b2) Receive the current global optimal solution provided by the group space through the receiving function;

b3) Implement 2-OPT operation on the belief space to optimize the belief space;

The implementation of the 2-OPT operation on the belief space is as follows:

b301) set r ₀ as a given constant in [0, 1], generate a random number r in the range of [0, 1], if r>r ₀ , turn to step b4);

b302) If there are nodes c _i and c _j in the current optimal path, where j≥i+2, and

d(c _i , c _i+1 )+d(c _j , c _j+1 )＞d(c _i , c _j )+d(c _i+1 , c _j+1 )

Then replace the edges (c _i , c _j ), (c _i+1 , c _j+1 ) with (c _i , c _i+1 ), (c _j , c _j+1 ), and the path in the line after exchange ( c _i ,..., c _i+1 ) are reversed; otherwise go to step b4).

b4) Output the optimal solution and provide it to step a4) through the influence function.

The influence function Influence() is:

Among them, EndStep is the preset maximum evolution algebra of the ant colony system, CurrentStep is the current algebra of ant colony evolution, BaseNum and C are constants, set by the user. Usually the value of BaseNum is 30, and the value of C: EndStep is 1:3. In this way, in the initial stage of ant colony evolution, the knowledge solution of belief space has little influence on it, so that it can ensure rapid evolution. In the later stage of ant colony evolution , the influence of the knowledge solution on it gradually increases, making it more able to accept the guidance of the knowledge space, expand the search space at the same time, and have better global search capabilities.

Individuals in the group space form individual experience during the evolution process, and pass the individual experience to the belief space through the function accept(), and the belief space compares and optimizes the received individual experience according to certain behavior rules to form an optimal solution. For the optimal solution found in the evolution process, the belief space adopts the random 2-OPT exchange operation to mutate and optimize the optimal solution, and makes full use of the simple and efficient characteristics of the random 2-OPT algorithm to complete its own mutation. After evolution The solution individuals are used to update the global pheromone of the population space, helping to guide the evolution process of the population space, so as to achieve the purpose of increasing the diversity of the population, preventing premature maturity, and reducing the calculation cost. After forming the group experience, the belief space modifies the behavior rules of individuals in the group space through the influence function, so that the individual space can obtain higher evolutionary efficiency.

Claims (7)

1. a kind of rail transit multi-mode optimal transit transfer querying method based on cultural ant colony, it is characterised in that the method Comprise the following steps：

1) central processing unit receives inquiry request by touch-screen, and site information is obtained from database according to inquiry request, Build path preference pattern；

2) central processing unit performs cultural ant colony on multiple populations based on path Choice Model, calculates and obtains under different optimal objectives Optimal trajectory traffic transfer scheme, export optimal path；

3) numerical value of path Choice Model is updated, and judges whether optimization terminates, if so, then feed back to result of calculation Touch-screen, execution step 4), if it is not, return to step 2)；

4) touch-screen shows result of calculation；

Described cultural ant colony includes the ant colony evolutionary process of group space and the renewal of knowledge process of belief space, described The ant colony evolutionary process of group space comprise the following steps：

A1 the pheromones distribution of group space) is initialized, and group space is divided into into multiple subgroups, each subgroup is respectively adopted not Parallel evolutionary is carried out with the Ant ColonySystem of behavior, the locally optimal solution of each subgroup is obtained；

A2) respective local information element is updated according to the information interactive strategy based on study mechanism between each subgroup；

A3) globally optimal solution is updated according to the locally optimal solution of each subgroup, and by it by receiving function storage to belief space；

A4) global information element renewal is carried out according to the output of belief space；

A5) judge whether to meet algorithm end condition, if meeting, algorithm terminates；Otherwise, a2 is gone to step)；

The renewal of knowledge process of described belief space is comprised the following steps：

B1) belief space is initialized；

B2) the current globally optimal solution that group space is provided is received by receiver function；

B3) implement 2-OPT operations to belief space, optimize belief space；

B4) optimal solution is exported, and step a4 is provided it to by influence function)；

Described each subgroup is respectively adopted the Ant ColonySystem of different behaviors and carries out parallel evolutionary and is specially：

A101) on the website that each subgroup is randomly placed at m ant of varying number in n website；

A102) each subgroup carries out state transfer according to respective behavior, next node is selected, while carrying out local information element Update, described behavior include it is random, comform, greedy or mixing；

A103) repeat step a102), until every ant is respectively formed a fullpath, i.e., each subgroup travels through respectively all sections Point, obtains respective locally optimal solution.

2. a kind of rail transit multi-mode optimal transit transfer issuer based on cultural ant colony according to claim 1 Method, it is characterised in that described inquiry request includes start site and final website.

3. a kind of rail transit multi-mode optimal transit transfer issuer based on cultural ant colony according to claim 1 Method, it is characterised in that described optimal objective includes that the time is most short, transfer is minimum and distance is minimum.

4. a kind of rail transit multi-mode optimal transit transfer issuer based on cultural ant colony according to claim 1 Method, it is characterised in that described is based on the information interactive strategy of study mechanism：

Each subgroup carries out information exchange with other neighbour two subgroups, by current locally optimal solution and neighbour other two The locally optimal solution of individual subgroup is compared, and the local information element for updating itself with more excellent locally optimal solution.

5. a kind of rail transit multi-mode optimal transit transfer issuer based on cultural ant colony according to claim 1 Method, it is characterised in that described function Accept () that receives is：

Accept ()=T

T is the constant of setting.

6. a kind of rail transit multi-mode optimal transit transfer issuer based on cultural ant colony according to claim 1 Method, it is characterised in that described 2-OPT operations of implementing to belief space are specially：

B301) r is set₀For a given constant in [0,1], random number r of [0, a 1] scope is produced, if r>r₀Then Go to step b4)；

B302) if there is node c in current optimal path_i、c_j, wherein j >=i+2, and

d(c_i,c_i+1)+d(c_j,c_j+1)>d(c_i,c_j)+d(c_i+1,c_j+1)

So by side (c_i,c_j)、(c_i+1,c_j+1) replace (c_i,c_i+1)、(c_j,c_j+1), the path (c after exchange in circuit_j,…, c_i+1) be reversed；Otherwise go to step b4).

7. a kind of rail transit multi-mode optimal transit transfer issuer based on cultural ant colony according to claim 1 Method, it is characterised in that described influence function Influence () is：

$Influence\left(\right)=\left\{\begin{array}{cc}BaseNum,& CurrentStep\&le;C\\ BaseNum*\frac{EndStep-CurrentStep}{EndStep-C},& otherwise\end{array}\right.$

Wherein, EndStep is Ant ColonySystem set in advance maximum evolution algebraically, and CurrentStep is that former generation is worked as in ant colony evolution Number, BaseNum and C is constant.