Modular automatic driving connection bus scheduling optimization method based on NSGA-II

Abstract

The invention relates to the technical field of automatic driving vehicle scheduling, and discloses a modular automatic driving connection bus scheduling optimization method based on NSGA-II, and the method comprises the steps: obtaining passenger order information, vehicle path information and vehicle connection information; calculating the operation cost of the modularized automatic driving connection bus according to the passenger order information and the vehicle path information; calculating the minimum fleet scale of the modularized automatic driving connection bus according to the vehicle connection information; according to the operation cost and the minimum fleet scale of the modularized automatic driving connection bus, constructing a scheduling optimization model considering the passenger capacity and the bus number connection constraint; the scheduling optimization model comprises multiple decisions of passenger distribution, timetable formulation, vehicle scheduling, vehicle marshalling and driving paths; and solving the scheduling optimization model to obtain a modular automatic driving connection bus scheduling optimization result. According to the invention, the operation efficiency and the service quality of the connection system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic driving vehicle scheduling, in particular to a modular automatic driving shuttle bus scheduling optimization method based on NSGA-II. BACKGROUND

[0002] With the continuous advancement of urbanization and the increasing travel demand of the public, urban transportation is facing problems such as traffic congestion, traffic energy consumption, and traffic environmental pollution. In order to solve these problems, the development of large-capacity public transportation modes, including buses, subways, and high-speed rails, is actively promoted to improve urban travel efficiency, relieve traffic pressure, and improve air quality. However, in the process of developing large-capacity public transportation modes, there is still a problem of insufficient travel connection in urban fringe areas. How to optimize the connection between rail transit and other transportation modes, especially the connection between rail transit and buses, is a problem of practical significance. The connection between rail transit and buses has the highest proportion and is an important consideration factor in overall planning and design. Effectively solving the problem of urban traffic connection and improving the convenience of passengers after leaving the rail transit station is an urgent need to optimize the efficiency of "door-to-door" travel chain and improve the overall operation efficiency of urban transportation. Although urban rail transit networks are developing rapidly, in low-density urban fringe areas, residents face difficulties in "first and last kilometer travel", resulting in poor public transportation experience.

[0003] Modular autonomous vehicles (MAVs) as a new type of transportation tool can effectively provide services for passengers who need short-distance connection. First, MAVs can form vehicle platoons with different vehicle capacities by coupling and decoupling vehicles. This feature can effectively solve the fluctuation of passenger demand at different times. When the passenger demand level is high, MAVs form vehicle platoons with high passenger capacity by coupling to provide high transport capacity, and when the passenger demand level is low, MAVs form vehicles with low passenger capacity by decoupling to save transport capacity. This way improves the utilization rate of vehicles and also provides higher quality transportation services for passengers - passengers no longer need to wait for the full load rate of vehicles.

[0004] The flexibility of MAV also brings certain challenges to its scheduling. Vehicle scheduling is an important link in the organization of feeder bus transport, and its purpose is to reasonably arrange the operation of vehicles to reduce the use of vehicles, operating costs and passenger costs. For systems that use traditional fixed-capacity buses, the scheduling problem mainly focuses on departure frequency, line allocation and personnel arrangement; while for MAV systems with dynamic reconstruction capabilities, scheduling not only needs to consider the running path and service time of the vehicle, but also needs to consider the combination of modules, collaborative scheduling and passenger allocation and other issues. This greatly increases the dimension and solution difficulty of the scheduling optimization model. SUMMARY

[0005] In view of the above problems in the prior art, the application provides a modular autonomous driving feeder bus scheduling optimization method based on NSGA-II, to solve the application of modular autonomous driving buses in the feeder rail transit scene, and improve the operation efficiency and service quality of feeder services.

[0006] In order to achieve the above application purposes, the technical scheme adopted by the application is:

[0007] A modular autonomous driving feeder bus scheduling optimization method based on NSGA-II, comprising the following steps:

[0008] Obtain passenger order information, vehicle path information and vehicle connection information;

[0009] Calculate the operating cost of the modular autonomous driving feeder bus according to the passenger order information and the vehicle path information;

[0010] Calculate the minimum vehicle fleet size of the modular autonomous driving feeder bus according to the vehicle connection information;

[0011] According to the operating cost and the minimum vehicle fleet size of the modular autonomous driving feeder bus, a scheduling optimization model considering the passenger capacity and train connection constraints is constructed; the scheduling optimization model includes multiple decisions of passenger allocation, timetable making, vehicle scheduling, vehicle grouping and driving path;

[0012] Solve the scheduling optimization model to obtain the scheduling optimization result of the modular autonomous driving feeder bus.

[0013] Further, the calculation of the operating cost of the modular autonomous driving feeder bus according to the passenger order information and the vehicle path information is specifically:

[0014] ;

[0015] Wherein, is the operating cost of the modular autonomous driving feeder bus; is the passenger unit waiting time cost, for a passenger order set, for an index sequence number of the passenger order; for a vehicle route set, for an index sequence number of the vehicle route; for a station set, for an index sequence number of the station; for a vehicle marshalling type set, for an index sequence number of the marshalling type; for an order contains a passenger quantity; for an order whether to be assigned to a path 0-1 decision variable; for a line departure time; for an order expected boarding time; for a passenger unit in-transit time cost; for a vehicle executing a route arrival time at a station ; for an order off-site; for a vehicle of marshalling type unit time running cost; for a path whether to be executed by a vehicle marshalling vehicle 0-1 decision variable; for a path whether to pass through a road section 0-1 decision variable; for a vehicle running time between stations and station ; for a fixed departure cost.

[0016] Further, the calculation method of the unit time running cost of the vehicle of marshalling type is:

[0017] ;

[0018] wherein, is an operating cost coefficient, is a passenger capacity of a single vehicle, is a scale effect coefficient.

[0019] Further, the calculation of the minimum fleet size of the modular autonomous shuttle bus according to the vehicle connection information is:

[0020] ;

[0021] wherein, is the minimum fleet size, is the execution route is followed by the execution route the number of vehicles.

[0022] Further, according to the operating cost of the modular automatic driving feeder bus and the minimum fleet size, a scheduling optimization model considering the passenger carrying capacity and the train connection constraint is constructed, specifically:

[0023] ;

[0024] s.t.;

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] ;

[0030] ;

[0031] ;

[0032] ;

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] ;

[0039] ;

[0040] ;

[0041] wherein, is the path whether it passes through the link arc 0-1 decision variable, taking value 1 if passing, otherwise taking value 0; denotes a vehicle route passing through stations from depot 0 denotes a vehicle route passing through stations back to depot 0, denotes a vehicle route passing through stations in sequence departure station of passenger p , denotes a route arrival station of passenger p, , is a positive integer, is the earliest departure time of route, is the latest departure time of route, is the number of vehicles executing route after executing route , is a 0-1 variable indicating whether route meets the connection condition with route , is the departure time of route .

[0042] Further, the NSGA-II genetic method with variable domain search strategy is used to solve the scheduling optimization model, including:

[0043] constructing passenger allocation chromosome and departure time chromosome;

[0044] generating a random permutation sequence containing all passenger numbers, and splitting the random permutation sequence into multiple sub-lists corresponding to vehicle trips;

[0045] calculating the fitness function corresponding to the objective function of the scheduling optimization model;

[0046] randomly selecting one route from each parent as a crossover fragment, moving the crossover fragment to the head of the other parent, checking and deleting repeated genes;

[0047] adopting variable domain search strategy for timetable search and path search;

[0048] adopting elite strategy to select the parent population of the next generation, and stopping iteration when the maximum number of iterations is reached;

[0049] inputting the Pareto front solution in the current population as the final Pareto solution set.

[0050] Further, the fitness function corresponding to the objective function of the scheduling optimization model is specifically:

[0051] ;

[0052] ;

[0053] wherein, is the fitness function value of the individual on the first objective function, is the first objective function value of the individual , is the penalty value corresponding to the violation of the passenger capacity constraint, is the 0-1 variable of the individual whether the passenger capacity constraint is violated, is the fitness function value of the individual on the second objective function, is the second objective function value of the individual , is the minimum fleet size of the individual .

[0054] Further, the calculation method of the minimum fleet size of the individual is:

[0055] Obtain the individual and the operation period, and initialize the arrival and departure function of the vehicle;

[0056] Traverse all routes in the individual, calculate the vehicle marshalling, departure time and arrival time, and update the arrival and departure function of the vehicle at the departure time;

[0057] Initialize the inverse difference function, and calculate the inverse difference function at each time point according to the updated arrival and departure function of the vehicle at the departure time;

[0058] Calculate the minimum fleet size of the individual according to the inverse difference function at each time point.

[0059] Further, the timetable search using the variable field search strategy includes:

[0060] Calculate the fitness function value of the current individual;

[0061] Randomly select a route in the current individual, calculate the earliest departure time of the selected route, and randomly select a time within the set range of the earliest departure time as the departure time of the selected route, and calculate the fitness function of the new individual;

[0062] Determine whether the new individual dominates the current individual; if yes, update the current individual with the new individual; otherwise, keep the current individual;

[0063] determining whether the maximum iteration number is reached; if yes, outputting the updated individual; otherwise, selecting the next individual to continue iteration.

[0064] Further, the path search adopting the variable field search strategy comprises:

[0065] calculating the fitness function value of the current individual;

[0066] randomly selecting a route in the current individual, exchanging the positions of two passenger orders in the selected route, and calculating the fitness function of the new individual;

[0067] determining whether the new individual dominates the current individual; if yes, updating the current individual with the new individual; otherwise, keeping the current individual;

[0068] determining whether the maximum iteration number is reached; if yes, outputting the updated individual; otherwise, selecting the next individual to continue iteration.

[0069] The present application has the following beneficial effects:

[0070] The present application constructs a multiple complex decision model comprising passenger allocation, vehicle path and timetable and vehicle scheduling, generates an optimal modular automatic driving bus scheduling scheme, thereby improving the operation efficiency and service quality of the transfer system; and adopts the NSGA-II genetic method with variable field search strategy, realizes efficient solution of the vehicle scheduling problem of the transfer bus system. BRIEF DESCRIPTION OF DRAWINGS

[0071] Figure 1 A flowchart of a modular automatic driving transfer bus scheduling optimization method based on NSGA-II in the embodiment of the present application;

[0072] Figure 2 A schematic diagram of a Pareto solution set obtained by solving the optimization model in the embodiment of the present application. DETAILED DESCRIPTION

[0073] The specific embodiments of the present application are described below to facilitate those skilled in the art to understand the present application, but it should be clear that the present application is not limited to the scope of the specific embodiments, and for those skilled in the art, it is obvious that various changes are within the spirit and scope of the present application defined and determined by the appended claims, and all the inventions utilizing the concept of the present application are within the scope of protection.

[0074] The application comprehensively considers the stop site selection, passenger distribution, vehicle path planning and vehicle scheduling in the connection operation scene, and proposes a modular automatic driving connection bus scheduling optimization method based on NSGA-II (Non-dominated Sorting Genetic Algorithm II). The application establishes a double-objective optimization model capable of simultaneously deciding the passenger distribution, vehicle path and timetable and vehicle scheduling, and designs an NSGA-II algorithm with a variable field search strategy, so as to realize efficient solution to the vehicle scheduling problem of the connection bus system.

[0075] The application proposes a modular automatic driving connection bus scheduling optimization method based on NSGA-II, so as to solve the multiple complex decision problems of passenger distribution, vehicle path and timetable and vehicle scheduling in the modular automatic driving vehicle scheduling problem, generate an optimal modular automatic driving bus scheduling scheme, and thus improve the response speed and service quality of the connection system.

[0076] As shown in Figure 1 The application embodiment provides a modular automatic driving connection bus scheduling optimization method based on NSGA-II, which comprises the following steps S1 to S5:

[0077] S1, obtaining passenger order information, vehicle path information and vehicle connection information;

[0078] In an optional embodiment of the application, the step S1 collects the departure time and destination information of the passenger order and the vehicle connection information through an intelligent mobile phone APP, a vehicle-mounted terminal device or a city traffic platform and the like.

[0079] S2, calculating the operation cost of the modular automatic driving connection bus according to the passenger order information and the vehicle path information;

[0080] In an optional embodiment of the application, the step S2 of calculating the operation cost of the modular automatic driving connection bus according to the passenger order information and the vehicle path information is specifically:

[0081] ;

[0082] Among them, is the operation cost of the modular automatic driving connection bus; is the passenger unit waiting time cost, is a passenger order set, is a passenger order index serial number; is a vehicle route set, is a vehicle route index serial number; is a station set, is a station index serial number; a set of vehicle formation types, an index number for the formation type; an order a number of passengers included; an order whether assigned to a path a 0-1 decision variable, an order assigned to a path , otherwise 0; a departure time of a line ; an expected pick-up time of an order ; a unit in-transit time cost of a passenger; a time for a vehicle executing a route to arrive at a site ; a drop-off site of an order ; a unit time running cost of a vehicle of a formation type ; a 0-1 decision variable for whether a path is executed by a vehicle of a formation type , a path is executed by a vehicle of a formation type , otherwise 0; a 0-1 decision variable for whether a path passes through a road segment , a path passes through a road segment , otherwise 0; a running time of a vehicle between a site and a site ; a fixed departure cost.

[0083] wherein the unit time running cost of a vehicle of a formation type is calculated as:

[0084] ;

[0085] wherein, is an operation cost coefficient, is a passenger carrying capacity of a single vehicle, is a scale effect coefficient.

[0086] S3, calculating a minimum fleet size of the modular autonomous shuttle bus according to vehicle connection information;

[0087] In an optional embodiment of the present application, step S3 calculates the minimum fleet size of the modular autonomous feeder bus according to the vehicle connection information calculation module, specifically:

[0088] ;

[0089] wherein, is the minimum fleet size, is the number of vehicles performing the route followed by the route .

[0090] S4, according to the operating cost of the modular autonomous feeder bus and the minimum fleet size, a scheduling optimization model considering the passenger capacity and the train connection constraint is constructed; the scheduling optimization model includes multiple decisions of passenger allocation, timetable making, vehicle scheduling, vehicle grouping and driving path;

[0091] In an optional embodiment of the present application, step S4 constructs a scheduling optimization model considering the passenger capacity and the train connection constraint according to the operating cost of the modular autonomous feeder bus and the minimum fleet size, the objective of the model is to minimize the fleet size and the operating cost, specifically:

[0092] ;

[0093] s.t.;

[0094] ;

[0095] ;

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] ;

[0101] ;

[0102] ;

[0103] ;

[0104] ;

[0105] ;

[0106] ;

[0107] ;

[0108] ;

[0109] ;

[0110] wherein, is a 0-1 decision variable of whether the path passes through the link arc , and takes the value 1 if it passes through, otherwise takes the value 0; denotes the vehicle path passing through the station from the depot 0, denotes the vehicle path passing through the station back to the depot 0, denotes the vehicle path passing through the station in turn, denotes the departure station of the passenger p, denotes the time of the path arriving at the station , , is a very large positive integer, is the earliest departure time of the route, is the latest departure time of the route, is the number of vehicles executing the route and then executing the route , is a 0-1 variable of whether the route satisfies the connection condition with the route , is the departure time of the route .

[0111] S5, solve the scheduling optimization model to obtain a modular automatic driving shuttle bus scheduling optimization result.

[0112] In an optional embodiment of the present application, step S5 proposes a NSGA-II with variable field search strategy to solve the multi-objective optimization model proposed in step S4, including:

[0113] Construct a passenger allocation chromosome and a departure time chromosome;

[0114] generating a random permutation sequence containing all passenger numbers, and splitting the random permutation sequence into a plurality of sub-lists corresponding to vehicle trips;

[0115] calculating a fitness function corresponding to the objective function of the scheduling optimization model;

[0116] randomly selecting one route from each of the two parents as a crossover fragment, moving the crossover fragment to the head of the other parent, checking for repeated genes and deleting them;

[0117] adopting a variable field search strategy for timetable search and path search;

[0118] adopting an elite strategy to select the parent population of the next generation, and stopping iteration when the maximum number of iterations is reached;

[0119] inputting the Pareto front solution in the current population as the final Pareto solution set.

[0120] The parameter settings of the embodiment are as follows: the population size is 100, the maximum number of iterations is 1000, and the mutation probability is 0.2. The specific solving process is as follows:

[0121] (1) Chromosome encoding and decoding. Integer encoding is adopted for chromosome encoding, including two sub-chromosomes: passenger allocation chromosome and departure time chromosome. The passenger allocation chromosome is represented in the form of list nested list, and the inner sub-list represents the passenger allocation result. Each gene on the departure time chromosome corresponds to the departure time of each route. Decoding is achieved by replacing the passenger number with its corresponding station.

[0122] (2) Population initialization. First, a random permutation sequence containing all passenger numbers is generated. Then, the sequence is split into a plurality of sub-lists, each corresponding to the trip arrangement of a vehicle.

[0123] (3) Fitness function. The fitness function corresponding to the objective function of the optimization model in step S4 is calculated as follows:

[0124] ;

[0125] wherein, is the fitness function value of individual on the first objective function, is the first objective function value of individual , is the penalty value corresponding to the violation of the passenger capacity constraint, is a 0-1 variable indicating whether individual violates the passenger capacity constraint, and individual violates the passenger capacity constraint is 1, otherwise 0.

[0126] Objective function value The corresponding fitness function can be directly expressed by the minimum fleet size, calculated by the inverse difference function, i.e.:

[0127] ;

[0128] Where, is the individual The fitness function value on the 2nd objective function, is the 2nd objective function value of individual , is the minimum fleet size of individual .

[0129] Where individual The corresponding minimum fleet size is calculated by an inverse difference function-based method:

[0130] 1. Input individual , operating period .

[0131] 2. Initialize the arrival and departure functions of modular vehicles For each time point : .

[0132] 3. Traverse all routes in individual : Calculate vehicle marshalling , departure time , arrival time . And update , . .

[0133] 4. Initialize the inverse difference function For each time point : .

[0134] 5. Calculate the inverse difference function For each time point : .

[0135] 6. Calculate the minimum fleet size required by individual .

[0136] 7. Output: The minimum fleet size required by individual .

[0137] ​​(4) Crossover and Mutation. The partial match crossover operator is used in the crossover operator. Specifically, 1. A route is randomly selected from each parent as a crossover fragment. 2. The crossover fragment is moved to the head of the other parent. 3. Check for duplicate genes and delete. The mutation operator contains three kinds: 1) Exchange mutation operator: randomly select two genes to exchange. 2) Merge mutation operator: randomly select two routes in the chromosome to merge to form a route. 3) Split mutation operator: randomly select a route in the chromosome to split it into two routes.

[0138] (5) Variable Neighborhood Search Strategy

[0139] 1) Timetable search operator, the specific steps are:

[0140] 1. Input individual .

[0141] 2. Calculate the two fitness function values of the current individual , .

[0142] 3. Randomly select a route in the current individual , calculate its earliest departure time and randomly select a time as the departure time of the route , the changed individual is , and the corresponding fitness function is , .

[0143] 4. If the individual dominates the individual , update . Determine whether the maximum number of iterations is reached.

[0144] 5. Output the updated individual .

[0145] 2) Path search operator, the specific steps are:

[0146] 1. Input individual .

[0147] 2. Calculate the two fitness function values of the current individual , .

[0148] 3. Randomly select a route in the current individual , exchange the route the location of the two passenger orders in the set of candidate locations, the changed individual order is , and the corresponding fitness function is , .

[0149] 4. If the individual order dominates the individual order , update . Determine whether the maximum number of iterations is reached.

[0150] 5. Output the updated individual order .

[0151] (6) The elite strategy is used to select the parent population of the next generation. When the maximum number of iterations is reached, the iteration is stopped.

[0152] (7) The Pareto front solution in the current population is input as the final Pareto solution set, as shown in Figure 2 .

[0153] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and a combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate a means for implementing the functions specified in the flow Figure 1 flow or multiple flows and / or blocks Figure 1 of the flowcharts and / or block diagrams.

[0154] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction means, which implements the functions specified in the flow Figure 1 flow or multiple flows and / or blocks Figure 1 of the flowcharts and / or block diagrams.

[0155] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide steps for implementing the functions specified in the flow Figure 1 flow or multiple flows and / or blocks Figure 1 of the flowcharts and / or block diagrams.

[0156] The principles and implementation manners of the present application are described by using specific examples in the present application. The above examples are only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges can be changed. In summary, the content of the present description should not be understood as a limitation on the present application.

[0157] Those skilled in the art will understand that the examples described herein are for the purpose of understanding the principles of the present application and should be understood as not limiting the scope of protection of the present application. Those skilled in the art can make various other specific modifications and combinations according to the technical inspiration disclosed in the present application without departing from the essence of the present application, and these modifications and combinations are still within the scope of protection of the present application.

Claims

1. A modular autonomous driving shuttle bus scheduling optimization method based on NSGA-Ⅱ, characterized by: The following steps are involved: Obtain passenger order information, vehicle route information, and vehicle connection information; Calculate the operating cost of modular autonomous shuttle buses based on passenger order information and vehicle route information; Calculate the minimum fleet size of modular autonomous shuttle buses based on vehicle connection information; Based on the operating costs and minimum fleet size of modular autonomous shuttle buses, a scheduling optimization model was constructed that takes into account passenger capacity and bus connection constraints. The scheduling optimization model includes multiple decisions on passenger allocation, timetable development, vehicle scheduling, vehicle grouping, and driving routes. The scheduling optimization model is solved to obtain the scheduling optimization results of modular autonomous shuttle buses.

2. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 1 is characterized in that: The operating cost of the modular autonomous shuttle bus is calculated based on passenger order information and vehicle path information: ； in, The operating costs of modular autonomous shuttle buses; is the unit waiting time cost of the passenger, For passenger order collection, Passenger order index number; is the vehicle route set, The vehicle route index number; Gather at the station, is the index number of the station; is a collection of vehicle group types, The index number of the group type; For orders Number of passengers included; For orders Is it assigned to a path? 0-1 decision variables in ; For the line departure time; For orders Expected boarding time; is the passenger's unit travel time cost; To execute the route Vehicles arriving at the station time; For orders The alighting point; For group type The unit time operating cost of the vehicle; For path Whether the vehicle is organized into The 0-1 decision variables executed by the vehicle; For path Whether it passes through the road section 0-1 decision variables; For vehicles at the site With site The running time between is the fixed dispatch cost.

3. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 2 is characterized in that: The group type is The unit time operation cost of a vehicle is calculated as follows: ； in, is the operating cost coefficient, is the passenger capacity of a single vehicle, is the scale effect coefficient.

4. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 3 is characterized in that: The minimum fleet size of modular autonomous shuttle buses calculated based on vehicle connection information is: ； in, is the minimum fleet size, To execute the route Then execute the route The number of vehicles.

5. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 4 is characterized in that: Based on the operating cost and minimum fleet size of modular autonomous shuttle buses, a scheduling optimization model considering passenger capacity and bus connection constraints is constructed. Specifically: ； st ; ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； in, For path Whether it passes through the road arc The 0-1 decision variable takes the value 1 if it passes, otherwise it takes the value 0; Represents the vehicle path Depart from parking lot 0 and pass through the station , Represents the vehicle path Passing the site Back to parking lot 0, Represents the vehicle path Passed through the sites Passenger p's departure station , Indicates the path Arrival at the site time, , is a positive integer, is the earliest departure time of the route, The latest departure time for the route. To execute the route Then execute the route The number of vehicles, For route With route A 0-1 variable indicating whether the connection condition is met, For route Departure time.

6. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 5 is characterized in that: The scheduling optimization model is solved using the NSGA-Ⅱ genetic method with a variable domain search strategy, including: Construct passenger allocation chromosome and departure time chromosome; Generate a random permutation sequence containing all passenger numbers and split the random permutation sequence into multiple sublists corresponding to vehicle trips; Calculate the fitness function corresponding to the objective function of the scheduling optimization model; Randomly select one route from each of the two parents as a crossover segment, move the crossover segment to the head of the other parent, check for duplicate genes and delete them; Adopting variable domain search strategy for timetable search and route search; Use the elite strategy to select the parent population of the next generation, and stop the iteration when the maximum number of iterations is reached; Input the Pareto front solutions in the current population as the final Pareto solution set.

7. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 6 is characterized in that: The fitness function corresponding to the objective function of the calculation scheduling optimization model is specifically: ； ； in, For individuals The fitness function value on the first objective function, For individuals The first objective function value of is the penalty value corresponding to violating the passenger capacity constraint, For individuals A 0-1 variable indicating whether the passenger capacity constraint is violated, For individuals The fitness function value on the second objective function, For individuals The second objective function value of For individuals Minimum fleet size.

8. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 7 is characterized in that: The minimum fleet size for an individual is calculated as: Get the individual and operating period, and initialize the vehicle arrival and departure functions; Traverse all routes in the individual, calculate vehicle grouping, departure time and arrival time, and update the arrival and departure function of the vehicle at the departure time; Initialize the inverse function and calculate the inverse function at each time point based on the updated arrival and departure function of the vehicle at the departure time; The minimum fleet size of an individual is calculated based on the inverse function at each time point.

9. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 8 is characterized in that: Timetable search using variable domain search strategy includes: Calculate the fitness function value of the current individual; Randomly select a route from the current individual, calculate the earliest departure time of the selected route, and randomly select a time within the set range of the earliest departure time as the departure time of the selected route, and calculate the fitness function of the new individual; Determine whether the new individual dominates the current individual; if so, update the current individual with the new individual; otherwise, keep the current individual; Determine whether the maximum number of iterations has been reached; if so, output the updated individual; otherwise, select the next individual to continue iteration.

10. The modular autonomous driving shuttle bus scheduling optimization method based on NSGA-II according to claim 9 is characterized in that: The path search using the variable domain search strategy includes: Calculate the fitness function value of the current individual; Randomly select a route from the current individual, exchange the positions of the two passenger orders in the selected route, and calculate the fitness function of the new individual; Determine whether the new individual dominates the current individual; if so, update the current individual with the new individual; otherwise, keep the current individual; Determine whether the maximum number of iterations has been reached; if so, output the updated individual; otherwise, select the next individual to continue iteration.

Images