Service design method for public transport oriented MaaS platform

Abstract

The invention relates to a service design method oriented to public traffic guidance, and the method employs a MaaS platform to cooperatively optimize the operation plans of a subway system and a public transportation system, builds a dual-target mixed integer nonlinear programming model on the basis of considering passenger paths, fusing non-public traffic travel modes, and considering the influence of congestion charging, and achieves the optimization of the operation plans of the subway system and the public transportation system. And the non-public traffic sharing rate is used as a part of an optimization target. Furthermore, in order to solve the model, according to the structural characteristics of the model, the original problem is decomposed into two small-scale sub-problems, and a hybrid algorithm based on an adaptive large neighborhood search method and GUROBI is designed, so that the solving efficiency in large-scale cases is remarkably improved, the overall service level and operation efficiency of the public transport system can be improved, and the method is suitable for popularization and application. Sustainable development of a traffic system is promoted, the defects of high cost, high energy consumption and low efficiency in a traditional independent operation mode are overcome, and efficient and high-quality travel service is provided for travelers.

Description

Technical Field

[0001] The present disclosure relates to transportation network planning and scheduling, and more particularly to a service design method for a public transportation-oriented MaaS platform. Background Art

[0002] Under the goal of sustainable transportation development, public transportation, primarily buses and subways, has become a key focus of future transportation system development due to its significant advantages in terms of greenness, efficiency, and cost-effectiveness. Furthermore, with the rise of the MaaS (Mobility as a Service) transportation concept, multimodal door-to-door travel services are the future trend of smart transportation. Coordinating different modes of transportation to provide efficient travel services, particularly coordinating the scheduling of two independently operated systems, bus and subway, has become a key issue that needs to be addressed.

[0003] Existing academic research and engineering practice primarily focus on scheduling optimization within a single system. The limited MaaS-related research typically focuses solely on optimizing public transportation departure frequency or directly assigning specific service parameters. This type of scheduling optimization within a single transportation system cannot effectively address the high costs, high energy consumption, and low efficiency inherent in standalone operation models. Summary of the Invention

[0004] The purpose of this paper is to propose a service design method for a public transport-oriented MaaS platform. By collaboratively optimizing the timetables, vehicle formations, stop plans, and pricing of two public transport systems, and comprehensively considering the impact of passenger routes and non-public transport congestion charges, the method aims to achieve the optimal total passenger travel cost and non-public transport sharing rate for the system, while maximizing the operating revenue of public transport.

[0005] First, the present disclosure proposes a public transportation-oriented service design method, including: based on a multimodal transportation network, establishing a dual-objective mixed integer nonlinear programming model that considers minimizing the total generalized travel cost of passengers and the non-public transportation share on the demand side and maximizing the public transportation operating revenue on the supply side, wherein the generalized travel cost is the weighted sum of waiting time, travel expenses and transfer time; constructing model constraints, including timetable constraints, stop plan constraints, train allocation constraints, passenger path constraints, passenger dynamic evolution constraints, capacity constraints, fare constraints, modular vehicle scheduling constraints and budget constraints; solving the dual-objective mixed integer nonlinear programming model to obtain the optimal public transportation service plan.

[0006] In this technical solution, the multimodal transportation network includes a non-public transportation system, a subway system, and a bus system based on modular vehicles. This involves leveraging the MaaS platform to collaboratively optimize the operational plans of the subway and bus systems. A dual-objective mixed-integer nonlinear programming model is developed, taking into account passenger routes, integrating non-public transportation modes, and the impact of congestion pricing. The non-public transportation share ratio is incorporated into the optimization objective. By solving the model to obtain decision variables for schedule-related variables, passenger route variables, vehicle formation variables, stop plan variables, fare variables, and modular vehicle rebalancing scheduling variables, the overall service level and operational efficiency of the public transportation system are improved, addressing the high costs, high energy consumption, and low efficiency of the traditional independent operation model, and providing travelers with efficient and high-quality travel services.

[0007] In a second aspect, the present disclosure provides a computer-readable storage medium storing a computer program that can be loaded by a processor and execute any method of the present disclosure.

[0008] The beneficial technical effects of this disclosure include: Based on a sustainable multimodal transportation network comprising a subway system, a bus system based on modular vehicles, and non-public transportation modes, the multimodal transportation network is optimized by considering passenger travel costs along different routes and the costs and revenues of operators, thereby reducing the operating costs of public transportation vehicles, improving the utilization efficiency of the multimodal transportation network, and maximizing the satisfaction of passengers' dynamic travel needs. Furthermore, a solution method for the established dual-objective mixed-integer nonlinear programming model is proposed, which significantly improves the solution efficiency. This method is therefore beneficial for optimizing the overall service of the public transportation system based on real-time traffic conditions (such as holidays) to meet the travel needs of travelers. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0010] Figure 1 A schematic diagram of the operation of a public transportation-oriented MaaS platform.

[0011] Figure 2 A flowchart for solving two sub-problems. DETAILED DESCRIPTION

[0012] Regarding the issue of how to coordinate different modes of transportation to provide efficient travel services, especially the issue of coordinating the scheduling of two independently operated systems, namely, buses and subways, this disclosure takes the perspective of a public transportation-oriented MaaS platform, while taking into account the impact of non-public transportation modes, and designs a multimodal service optimization model that includes buses, subways, and private cars to provide the system with optimal, efficient, and high-quality travel services. In the designed multimodal service optimization model, dynamic passenger travel demand, time-varying vehicle capacity constraints, and passenger travel paths that include multiple transportation modes are taken into account. In addition, infrastructure constraints and a modular vehicle rebalancing scheduling process are also considered. That is, based on a multimodal transportation environment, this disclosure proposes a method for collaborative optimization of the operations of different public transportation systems.

[0013] The following, combined with the accompanying drawings, provides a clear and complete description of how the technical solution of this case is implemented. Obviously, the described implementation methods are only part of the implementation methods of this case, not all of the implementation methods. Based on the implementation methods of this case, all other implementation methods obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0014] Transit-oriented MaaS platforms collaborate with metro and bus companies and have decision-making authority over timetables, vehicle composition, stop schedules, and pricing for both the metro system and the modular bus system. The platform comprehensively evaluates passenger travel costs, the share of non-public transportation, and public transportation operating revenue, collaboratively optimizing the scheduling of both systems and recommending the optimal route for passengers. Modular vehicles are vehicles that can be flexibly combined and separated based on demand, such as the intelligent transportation system Next Future Transportation being piloted in Dubai.

[0015] See also Figure 1 , the multimodal transportation network includes non-public transportation system, subway system and bus system based on modular vehicles. The subway system is composed of stations and subway lines. The bus system is composed of bus stations, stations and bus lines. Non-public transportation includes private cars, and non-public transportation lines are composed of stations and non-public transportation lines. Among them, there are transfer channels between the bus system and the subway system. The stations of non-public transportation are the union of the stations of the subway system and the bus system. For passengers with the same origin-destination, there may be multiple feasible public transportation paths and non-public transportation paths, such as Figure 1 The blue path, red path, and dashed path shown in (b) of FIG. 3 are shown, and each path generally has a different travel cost. Travel costs are not limited to including waiting time costs, transfer time costs, and travel fares. This disclosure only uses these three costs to constitute travel costs as an example.

[0016] It should be noted that the non-public transportation routes in the figure only represent the connectivity between two locations and have nothing to do with the actual road structure.

[0017] Denote the multimodal transportation network as ,in Indicates the subway system, represents a public transportation system based on modular vehicles, Represents a non-public transportation system; the subway and bus station sets are and , the line sets are and , non-public transportation stations are , the line is ;Remember the service set on each line is A service is a trip from the starting station to the ending station, and the number of services is the number of trips. For example, a bus service or subway service is a trip from the starting station to the ending station along the subway train or bus route. The number of trips from the starting station to the ending station is the number of services.

[0018] The set of discrete time periods is denoted as , where the length of each time period is . Let the dynamic passenger flow demand be , that is, in the time period From the station Depart for the station The number of passengers. The path set is , where each path A subroutine sequence Each sub-trip is recorded as , indicating that the line Services from the station Go to the station .

[0019] By abstracting and expressing the multimodal transportation network and the discretized planning time domain using mathematical modeling, it is convenient to characterize the spatial characteristics of the physical lines, the spatiotemporal status of vehicles and passengers, and the characteristics of the travel paths, providing a basis for subsequent optimization.

[0020] Based on the above sets, parameters and index symbols, the decision variables required for the model are established, including timetable-related variables, passenger path variables, vehicle formation variables, stop plan variables, fare variables and modular vehicle rebalancing scheduling variables. Specifically, the variables involved in the service design optimization model include 、 、 、 、 、 、 、 、 , meanings refer to Table 1 below. Indicates public transportation routes. Indicates the station, , Indicates the index of the service, For the line The service collection on represents the passenger path, Represents a discrete time period.

[0021] Table 1

[0022] The above lines On Service Use grouping as Train instructions, for example, if you select line On Service Use grouping as The train is 1, otherwise is 0.

[0023] These decision variables can be used to quantitatively describe the interaction between the operation of subway and bus systems and the dynamic evolution of passenger traffic, and quantitatively characterize the effectiveness of operational plans from both the supply and demand sides, laying the foundation for optimizing service design for public transport-oriented MaaS platforms.

[0024] By deeply analyzing the coupling relationship between passengers on different paths and subway trains or buses in a multimodal transportation network, as well as the coupling relationship between modular vehicles and stations, and taking into account the safe operation conditions, pricing standards, and operating budgets of the subway and bus systems, a series of constraints are constructed, including timetable constraints, stop plan constraints, train assignment constraints, passenger path constraints, passenger dynamic evolution constraints, capacity constraints, fare constraints, modular vehicle scheduling constraints, and budget constraints, as shown below.

[0025] (1) Schedule constraints Introducing parameters Indicates line Services at the station Running time between the downstream station and the station; introduce the parameter Indicates the line when the stop occurs The stop time of the service at each station; the introduction of parameters and Respectively represent lines The minimum and maximum departure intervals between two adjacent services are as follows:

[0026] (2) Stop plan constraints Introducing parameters and Respectively represent lines The subway system adopts a stop-at-every-station operation mode, and the bus system can skip at intermediate stations except the first and last stations. The specific form of the constraints is as follows:

[0027] (3) Group allocation constraints Introducing parameters It represents the maximum number of modular units that a modular vehicle can operate safely. The specific form of the constraint is as follows:

[0028] (4) Passenger path constraints Introducing parameters Indicates origin-destination The corresponding path set and constraint conditions are as follows:

[0029] (5) Passenger dynamic evolution constraints Introducing variables , and In the time period line Services on At the station Indicator variables for service arrival, departure, and stop times; introducing variables , and Respectively indicate the time period line The station on Arrive, get off and wait and the travel path is The number of passengers; introduce the parameter Indicates line and lines At the station The transfer travel time, the specific form of the constraints is as follows:

[0030] (6) Capacity constraints Introducing variables Indicates line On Service Leaving the station The number of passengers on board at the time; the introduction of parameters Indicates the marshaling in the subway system The capacity of the train; introduce the parameter represents the capacity of a modular unit in the bus system. The specific form of the constraints is as follows:

[0031] (7) Fare constraints Introducing parameters and , which represent the lines The minimum and maximum fares per unit distance; introduce variables Indicates the path Travel fare; introduce parameter Indicates line Get on the bus station and the station The specific form of the constraint conditions is as follows:

[0032] (8) Modular vehicle scheduling constraints Introducing variables and Respectively indicate the time period line Services on At the station 0-1 indicator variables for arrival and departure times; introduce variables , and Respectively indicate stations In the time period Number of modular units held, arrived, and departed; introduction of parameters Indicates the station The maximum number of modular units to be retained; introduction parameters Indicates the station To the station The empty vehicle travel time; introduce the parameter Indicates service station A collection of lines; introducing parameters , the specific form of the constraints is as follows:

[0033] (9) Budget constraints Introducing parameters Indicates subway lines The upper group is The operating cost of the train; the introduction of parameters and Respectively indicate stations and the station Along the route service distance and empty distance; introduce parameters Represents the cost per unit distance of modular units; introduces parameters and They represent the operating budgets of the subway system and the bus system respectively. The specific forms of the constraints are as follows:

[0034] Next, we construct a dual-objective mixed-integer nonlinear programming model for service design optimization, taking into account the demand-side total generalized travel cost of passengers and the supply-side public transportation operating revenue, combined with the non-public transportation share. The generalized travel cost is the weighted sum of waiting time, travel expenses, and transfer time.

[0035]

[0036] Equations (33)-(34) constitute the objective function of the dual-objective mixed integer nonlinear programming model, and (1) (32) constitute the constraints of the objective function, namely:

[0037] By introducing a public transport-oriented MaaS platform, this model aims to collaboratively optimize the operation plans and passenger routes of the subway and bus systems from a system perspective, so as to achieve optimal passenger travel in the system, improve the overall service level and operational efficiency, and then guide passengers to choose public transportation and promote the sustainable development of the transportation system.

[0038] pass The constraint method transforms the above dual-objective model into a single-objective model, and then introduces auxiliary variables to further process the model into a form containing only quadratic nonlinear constraints, so that the transformed model can be directly solved using the optimization solver.

[0039] Specifically, the parameters are introduced Convert the maximization objective into constraints, and the equivalent model is specifically expressed as follows:

[0040] Constraints (16) and (17) are cubic nonlinear constraints of three variables multiplied together. Auxiliary variables are defined for constraints (16) and (17) respectively. and , the introduced auxiliary variables are used to process constraints (16) and (17) into quadratic constraints that can be handled by the optimization solver without adding too many variables and constraints.

[0041] The corresponding constraints of auxiliary variable definition and transformation are expressed as follows:

[0042] In summary, the original problem can be reconstructed into a form that can be directly processed by the following solver:

[0043] By analyzing the structural characteristics of the above problem and the mathematical properties of the optimization model, the reconstructed original problem (39) is decomposed into two simpler sub-problems, so that a hybrid algorithm can be designed based on the adaptive large neighborhood search method and the GUROBI solver to achieve efficient solution of large-scale problems.

[0044] The first sub-problem is denoted as SP-1, which determines the timetable and stop plan of the subway system and bus system under constraints (1)-(5), that is, the decision variables , and The second subproblem is denoted as SP-2. It uses the solution of subproblem SP-1 as input to optimize the vehicle formation, pricing, passenger routing, and rebalancing scheduling of modular units in the two systems. The specific expression is as follows:

[0045] Based on the above model decomposition, an adaptive large neighborhood search method was developed to find a feasible timetable and stop plan for subproblem SP-1. This method is used as input. Subproblem SP-2 is an optimization solver that can efficiently solve a linear programming model. The optimal objective value of subproblem SP-2 is used to evaluate the feasible solution of subproblem SP-1, thereby guiding the adaptive large neighborhood search method to find a better timetable and stop plan. The search process is iterative until the termination condition is met.

[0046] The framework of the aforementioned solution method is as follows Figure 2 shown.

[0047] The above-mentioned adaptive large-area search method specifically includes: S1. Initialization and Data Input: Based on a uniform departure interval and stop-at-every-station operation model, generate an initial feasible timetable and stop plan. Use this initial timetable and stop plan as the currently feasible and optimal solutions to the first subproblem. Design the destruction operator for arrival time, the destruction operator for stop plan, and the repair operator for arrival time, and initialize the weights and scores for each operator.

[0048] S2. An adaptive large neighborhood search algorithm is used to solve the first subproblem: Based on the current feasible solution to the first subproblem, a roulette wheel method is used to select operators according to operator weights to find new feasible timetables and stop plans for the subway and bus systems, thereby obtaining candidate feasible solutions to the first subproblem. S3. Solve the second sub-problem using the optimization solver: The candidate feasible solution of the first sub-problem is used as the input of the second sub-problem and solved using the optimization solver to obtain the current optimal public transportation service plan and the optimal objective value; S4. Update operator scores and weights, current optimal solution, and current feasible solution: Based on the optimal target values ​​corresponding to the current feasible solution, candidate feasible solutions, and current optimal solution, calculate and update the scores and weights of each operator, update the current optimal solution, and select the current feasible solution for the next search based on the current feasible solution update strategy; When the termination condition is not met, return to step S2; otherwise, output the public transportation service plan corresponding to the current optimal solution.

[0049] The above arrival time destruction operator is configured to randomly select a certain number of services and randomly advance or delay their first station arrival time by a discrete time period, and the number of selected services decreases during the iteration process.

[0050] The destruction operator of the above stop plan is configured to randomly select a certain number of bus services and intermediate stations and change whether they stop or not. Similarly, the selected number decreases during the iteration process.

[0051] The above repair operator is configured such that the candidate solution generated by the destruction operator may become an infeasible solution due to violation of constraint (3). Therefore, the departure interval of two adjacent services is adjusted until it satisfies the maximum and minimum departure interval constraints while ensuring that all services are scheduled within the planning time domain.

[0052] definition and Operators The initial score and weight are 0 and 1 respectively. The score is updated based on the quality of the candidate feasible solution. The better the quality, the higher the score. is replaced by the actual operator.

[0053] The update rule of the operator score is: given a set of score change parameters , whose values ​​range from large to small. If the objective function value corresponding to the obtained candidate feasible solution is better than the objective function value corresponding to the current optimal solution, the score of the operator used increases. ,Right now If the objective function value of the candidate feasible solution is worse than the objective function value of the current optimal solution, but better than the objective function value of the current feasible solution, the score of the operator used is increased. ,Right now If the objective function value of the candidate feasible solution is worse than the objective function value of the current feasible solution, but the candidate feasible solution is selected as the current feasible solution for the next search, the score of the operator used increases ,Right now ; Otherwise, the score of the operator used is increased ,Right now .

[0054] The weight update formula is as follows:

[0055] in, is the weight update coefficient. During the search process, the roulette wheel method is used to select operators based on the operator weights to generate new feasible solutions.

[0056] Current feasible solution update strategy: To avoid falling into the local optimal solution, the current feasible solution for the next search process is selected similarly to the simulated annealing algorithm. That is, when the optimal target value corresponding to the candidate feasible solution is better than the optimal target value corresponding to the current feasible solution, the candidate feasible solution is directly accepted as the current feasible solution for the next search process; otherwise, the Metropolis-Hastings criterion is used to determine whether to accept the candidate feasible solution, and the acceptance probability is ,in represents the difference between the optimal objective value corresponding to the candidate feasible solution and the current feasible solution, is the system temperature parameter, which decreases with the increase of iteration number.

[0057] The termination conditions of the algorithm include: (i) reaching the maximum number of iterations (ii) continuous The current optimal target value of the iteration remains unchanged.

[0058] Finally, a set of numerical experiments are conducted to compare the performance of the proposed hybrid algorithm and the GUROBI solver, verifying the effectiveness of the proposed algorithm (see Table 2).

[0059] Table 2

[0060] based on Figure 1For the multimodal network in (a), the planning horizon consists of 60 discrete time periods, each of 1 minute in length. Within this horizon, 10 services are scheduled on each subway line, and 6 services are scheduled on each bus line. For the subway service, two train formations with capacities of 400 and 650 are considered. The operating costs of executing a service are 1000 and 1500, respectively. The minimum and maximum fares per unit distance are 2 and 3, respectively, and the operating budget is 30,000. For the bus system based on modular vehicles, the capacity of the modular unit is 30, and the operating budget is 30,000. In addition, the solution time of the GUROBI solver is limited to 7200 seconds, and the maximum number of iterations of the hybrid algorithm and the number of consecutive iterations during which the current optimal objective value remains unchanged are set to 1000 and 300, respectively.

[0061] Based on the above experimental settings, Right-hand side of the constraint The values ​​of 110,000, 130,000, 150,000, 170,000, and 190,000 are used for solution, and the solution results of the hybrid algorithm and the GUROBI solver are shown in Table 1. The results show that the proposed hybrid algorithm can find Pareto solutions that are significantly better than those of the GUROBI solver, while reducing the computational time by an average of 87.498%.

[0062] From the above, it can be seen that the purpose of the present disclosure is to use the MaaS platform to collaboratively optimize the detailed operation plans of the subway system and the bus system. To this end, a dual-objective mixed integer nonlinear programming model is established based on the consideration of passenger paths, the integration of non-public transportation modes represented by private cars, and the consideration of the impact of congestion charges, and the non-public transportation share rate is used as part of the optimization target. Furthermore, in order to solve the model, the original problem is decomposed into two smaller sub-problems according to the structural characteristics of the model, and a hybrid algorithm based on the adaptive large neighborhood search method and GUROBI is designed, which significantly improves the solution efficiency in large-scale cases. The technical solution disclosed in this disclosure can be used to improve the overall service level and operational efficiency of the public transportation system, promote the sustainable development of the transportation system, solve the shortcomings of high cost, high energy consumption and low efficiency under the traditional independent operation model, and provide travelers with efficient and high-quality travel services.

[0063] Through the description of the above embodiments, those skilled in the art can clearly understand that according to the method disclosed herein, a corresponding system can be implemented.

[0064] Through the above description of the embodiments, those skilled in the art will clearly understand that the disclosed methods and systems can be implemented using software plus necessary general-purpose hardware. Of course, they can also be implemented using dedicated hardware, including application-specific integrated circuits, dedicated CPUs, dedicated memories, and dedicated components. Generally speaking, any function performed by a computer program can be easily implemented using corresponding hardware. Moreover, the specific hardware structures used to implement the same function can also be diverse, such as analog circuits, digital circuits, or dedicated circuits. However, for the present disclosure, software implementation is often the preferred embodiment.

[0065] Although the embodiments of the present disclosure have been described above with reference to the accompanying drawings, the present disclosure is not limited to the specific embodiments and application areas described above. The specific embodiments described above are merely illustrative and instructive, and not restrictive. A person of ordinary skill in the art, guided by this specification and without departing from the scope of protection of the claims of the present disclosure, may devise various other forms, all of which fall within the scope of protection of the present disclosure.

Claims

1. A public transportation-oriented service design method, characterized in that: include: Based on a multimodal transportation network, a dual-objective mixed-integer nonlinear programming model is established, which considers minimizing the total generalized travel cost of passengers and the non-public transportation share on the demand side, and maximizing the public transportation operating revenue on the supply side. The generalized travel cost is the weighted sum of waiting time, travel expenses, and transfer time. Construct model constraints, including timetable constraints, stop plan constraints, train assignment constraints, passenger path constraints, passenger dynamic evolution constraints, capacity constraints, fare constraints, modular vehicle scheduling constraints, and budget constraints; Solve the bi-objective mixed-integer nonlinear programming model to obtain the optimal public transportation service plan.

2. The method according to claim 1, characterized in that Solve the bi-objective mixed integer nonlinear programming model, including: The constraint method transforms the dual-objective mixed integer nonlinear programming model into a single-objective model, and introduces auxiliary variables to further process the model into a form containing only quadratic nonlinear constraints, so that the transformed model can be directly solved using the optimization solver.

3. The method according to claim 2, characterized in that pass The constraint method transforms the bi-objective mixed integer nonlinear programming model into a single-objective model, including: introducing parameters Convert the maximization objective into constraints.

4. The method according to claim 2, characterized in that The auxiliary variables further transform the model into a form that only contains quadratic nonlinear constraints, specifically: The following constraints on the number of passengers in the passenger dynamics evolution constraint are: By introducing auxiliary variables and This is treated as a quadratic constraint for the optimization solver: In the above formula: variable , and Respectively indicate the time period line The station on Arrive, get off and wait and the travel path is the number of passengers; For the time period line Services on At the station An indicator variable for service stop time; Indicates line Services on At the station A 0-1 variable indicating whether to stop at the station; and represent the subway line set and bus line set respectively, is a combination of discrete time periods, t is the time period, For the line The station set on, i, j is The station in For the line The service collection on Index of services, is a set of paths, each travel path A subroutine sequence composition, For sub-trip, it means passing through the line Services from the station Go to the station .

5. The method according to claim 3, characterized in that The single-objective model is decomposed into the first sub-problem and the second sub-problem. The objective functions of the first sub-problem and the second sub-problem are the same, and both have the same objective function. The constraints converted from the maximization objective are different in that the first sub-problem also has timetable constraints and stop plan constraints, while the constraints of the second sub-problem also include train formation allocation constraints, passenger path constraints, passenger dynamic evolution constraints, capacity constraints, fare constraints, modular vehicle scheduling constraints and budget constraints.

6. The method according to claim 5, characterized in that After decomposing the single-objective model into two sub-problems, the solution methods include: S1. Use the initial timetable and stop plan as the current feasible solution and the current optimal solution to the first subproblem. Design the destruction operator for the arrival time, the destruction operator for the stop plan, and the repair operator, and initialize the weight and score of each operator. S2. Based on the current feasible solution to the first subproblem, a roulette wheel method is used to select operators according to operator weights to find new feasible timetables and stop plans for the subway system and bus system, thereby obtaining candidate feasible solutions to the first subproblem. S3. Use the candidate feasible solution of the first sub-problem as the input of the second sub-problem to obtain the current optimal public transportation service plan and the optimal target value; S4. Calculate and update the scores and weights of each operator based on the optimal target values ​​corresponding to the current feasible solution, candidate feasible solutions, and the current optimal solution, update the current optimal solution, and select the current feasible solution for the next search; When the termination condition is not met, return to step S2; otherwise, output the public transportation service plan corresponding to the current optimal solution.

7. The method according to claim 7, characterized in that The initial timetable and stop plan are generated based on the operation mode of uniform departure interval and stop at every station.

8. The method according to claim 7, characterized in that The operator score is calculated as follows: Given a set of score change parameters , whose values ​​are arranged from large to small; If the objective function value corresponding to the obtained candidate feasible solution is better than the objective function value corresponding to the current optimal solution, the score of the operator used increases ,Right now ; If the objective function value corresponding to the obtained candidate feasible solution is worse than the objective function value corresponding to the current optimal solution, but better than the objective function value corresponding to the current feasible solution, the score of the operator used increases ,Right now ; If the objective function value corresponding to the obtained candidate feasible solution is worse than the objective function value corresponding to the current feasible solution, but the candidate feasible solution is selected as the current feasible solution for the next search, the score of the operator used increases ,Right now ; Otherwise, the score of the used operator is increased ,Right now .

9. The method according to claim 7, characterized in that The weights are calculated as follows: in, For operator The score, For operator The weight of is the weight update coefficient.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 9.

Images