A timetable optimization method considering passengers' travel choices

By constructing a generalized fee function and Weibull distribution model for passenger travel, combining multi-objective hybrid planning and non-dominant sorting evolution algorithm, the multi-mode timetable optimization problem in the existing technology is solved, and intercity traffic coordination and travel efficiency are improved to meet passengers. need.

CN116229733BActive Publication Date: 2024-11-15BEIJING JIAOTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310211221.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-03-07
Publication Date
2024-11-15
Estimated Expiration
2043-03-07

AI Technical Summary

Technical Problem

The existing multi-mode schedule optimization model is difficult to fully consider the actual passenger travel needs, and the multi-objective solution algorithm is difficult to effectively and stably solve the multi-mode schedule, resulting in low intercity traffic coordination and low travel efficiency.

Method used

Build a generalized fee function for passenger travel, use the discrete selection model based on Weibull distribution, establish a multi-objective hybrid planning model, and use the multi-objective non-dominant sorting evolution algorithm and pseudo-weight method to solve iteratively, obtain the non-dominant optimal solution set to satisfy passengers The timetable optimization results of requirements.

Benefits of technology

The convergence, stability and diversity of multi-mode schedule optimization have been improved, passenger travel convenience and intercity traffic coordination have been improved, and theoretical and technical support has been provided to balance intercity passenger demand and capacity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116229733B_ABST
    Figure CN116229733B_ABST
Patent Text Reader

Abstract

The present invention provides a timetable optimization method that takes into account the travel choices of passengers. The method includes: constructing a passenger travel generalized cost function that includes the parameters of on-board time, waiting time, walking time, and travel cost, and characterizing the travel choice behavior of passengers in multi-modal transportation modes; using the passenger generalized cost function as input, constructing a discrete choice model of passenger travel modes, and using the discrete choice model to obtain the passenger flow data of each path; constructing a multi-objective hybrid planning model based on the passenger flow data of each path; using a multi-objective non-dominated sorting evolutionary algorithm to iteratively solve the multi-objective hybrid planning model to obtain a non-dominated optimal solution set, and using a multi-criteria decision-making method according to the operator's decision preferences to obtain a timetable optimization result that best meets the needs of passengers. The method of the present invention solves the problem of lack of coordination among multiple modes of transportation and low passenger travel efficiency caused by low inter-city transportation service frequency and different frequencies of multiple modes.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of inter-city multi-modal transportation intelligent planning and decision-making technology, and in particular to a timetable optimization method considering passenger travel options. Background Art

[0002] Urban agglomerations require a high degree of transportation integration, but the dispersion of stations, uncoordinated transport capacity, and diversity of passenger choices have reduced the degree of integration. Intercity multimodal travel still has problems of low coordination and low travel efficiency. Passengers departing at the same time have different experiences. Some face empty trains, while others face overcrowded vehicles, which brings great troubles to the daily operation of intercity transportation and intercity passenger travel. At the same time, in intercity travel, passengers rarely change their travel methods, but more often adjust their itineraries and departure times. Therefore, in order to achieve the goal of reducing passenger travel inconvenience while fine-tuning the timetable to achieve the goal of balancing supply and demand, it is of great research significance to consider passenger travel choices to coordinate and optimize the multimodal timetable.

[0003] At present, multimodal timetable optimization is mainly a single-objective optimization study. For example, minimizing passenger travel time and optimizing multiple objective functions at the same time are less studied, which means that most studies will always focus on one aspect and ignore another, which may increase passenger travel costs while reducing passenger travel time. At the same time, in timetable optimization research, most studies are only based on the single topic of timetable optimization, ignoring the influence relationship between passenger behavior and timetable. Therefore, it is difficult to coordinate the optimization of multimodal timetables using the existing multimodal timetable optimization model. Of course, the multi-objective multimodal timetable optimization problem is an NP-hard problem, which is difficult to solve and is difficult to solve with ordinary genetic algorithms. Summary of the invention

[0004] The embodiments of the present invention provide a timetable optimization method taking into account passenger travel choices, so as to solve the problem that the existing timetable optimization model is difficult to fully consider the actual timetable optimization, and the problem that the existing multi-objective solution algorithm is difficult to effectively and stably solve the multi-mode timetable.

[0005] In order to achieve the above object, the present invention adopts the following technical scheme.

[0006] A timetable optimization method considering passenger travel choices, comprising:

[0007] Construct a generalized cost function for passenger travel that includes the time in the car, waiting time, walking time and travel cost parameters, and use the generalized cost function for passenger travel to characterize the travel choice behavior of passengers in multi-modal transportation modes;

[0008] Taking the passenger generalized cost function as input, a discrete choice model of passenger travel mode based on Weibull distribution is constructed, and the passengers are allocated to each path by using the discrete choice model to obtain the passenger flow data of each path;

[0009] Based on the passenger flow data of each route, a multi-objective hybrid planning model is constructed with the goals of minimizing passenger waiting time, minimizing the total adjustment quantity and time of multi-modal trains, and minimizing passenger travel costs.

[0010] Iteratively solving the multi-objective hybrid programming model using a multi-objective non-dominated sorting evolutionary algorithm to obtain a non-dominated optimal solution set, the non-dominated optimal solution set including schedule adjustment schemes corresponding to various objective function values ​​of the multi-objective hybrid programming model;

[0011] By using a multi-criteria decision-making method and according to the decision preferences of the operator, a timetable optimization result that best meets the passenger needs is obtained in the non-dominated optimal solution set.

[0012] Preferably, the construction includes a passenger travel generalized cost function including the parameters of travel time, waiting time, walking time and travel cost, and the passenger travel generalized cost function is used to characterize the travel choice behavior of passengers in multi-modal transportation modes, including:

[0013] Based on the five multi-modal transportation modes of high-speed rail, motor train, ordinary rail, bus and private car, the travel cost components corresponding to each transportation mode are defined respectively, and the travel cost components include the time in the car, waiting time, walking time and travel cost;

[0014] The calculation formula of time value VOT is:

[0015] VOT=I / (W×D×60)

[0016] Through the time value VOT, time expenses are converted into costs to achieve the unification of generalized expenses, where I is the average annual income of residents, W is the statutory working hours per day, and D is the statutory working days;

[0017] Constructing a generalized cost function for passenger travel for:

[0018]

[0019] in, is the time in the vehicle for different modes m, is the waiting time of different modes m, is the walking time of different modes m, p m is the travel cost of mode m, θ1, θ2, θ3, θ4 are the set weight parameters.

[0020] Preferably, the passenger generalized cost function is used as input to construct a discrete choice model of passenger travel mode based on Weibull distribution, and the passengers are allocated to each path by using the discrete choice model to obtain the passenger flow data of each path, including:

[0021] The passenger generalized cost function is used as the input of discrete choice model to quantitatively study the impact between inter-city multi-modal timetable adjustment and passenger travel behavior;

[0022] The discrete choice model Multinominal weibit model of passenger travel mode based on Weibull distribution is constructed as follows:

[0023]

[0024] in, is the partial derivative of the joint generating function, is the independent Weibull distribution random error term of OD for mode m between ij, is the passenger perceived cost of mode m, is the average perceived cost, ζ ij for β ij is the shape parameter∈(0,∞), M ij is the set of all ways between OD pairs ij, and IJ is the set of all OD (Origin to Destination) pairs.

[0025] By solving the above discrete choice model, the passenger flow sharing rate of each path is obtained. The sum of the passenger flow sharing rates of all paths is 1. The total passenger flow is distributed to each path according to the calculated passenger flow sharing rate. The sum of the passenger flow on all paths is the total passenger flow.

[0026] Preferably, the multi-objective hybrid planning model based on the passenger flow data of each path is constructed with the objectives of minimizing the passenger waiting time, minimizing the total adjustment quantity and time of multi-mode trains, and minimizing the passenger travel cost, including:

[0027] In the multi-objective hybrid planning model, multiple objective functions at the passenger demand level and the transportation mode operation service level are set based on the passenger flow data of each path. The objective functions include the objective function of minimizing the passenger waiting time, the objective function of minimizing the total number of train schedule adjustments, the objective function of minimizing the total adjustment time, and the objective function of minimizing the passenger travel cost. Specifically,

[0028] Passenger waiting time minimization objective function:

[0029]

[0030] The objective function of minimizing the number of train schedule adjustments is:

[0031]

[0032] The total adjustment time is minimized to minimize the objective function:

[0033]

[0034] The objective function to minimize the total travel cost of passengers is:

[0035]

[0036] in, Q is the waiting time of passenger j between trips i. i is the adjustment status of train number i, which is 1 if adjusted, otherwise 0, i.e., the adjustment time is 0min. i Arrival after is the departure time of train number i after adjustment, C ij is the fare of train number i. i Arrival before To adjust the departure time of the previous train number i, I is the set of train number i, J is the set of passenger j, is the total waiting time of passengers, Q total Adjust the total number of train schedules. is the total adjustment time of all trains, C total is the total travel cost for all passengers;

[0037] The constraints of the multi-objective hybrid programming model are set, wherein the constraints include departure interval constraints, path feasibility constraints, time adjustment range constraints, capacity constraints, on-board status constraints, time window constraints and passenger selection constraints.

[0038] Preferably, the multi-objective hybrid programming model is iteratively solved by adopting a multi-objective non-dominated sorting evolutionary algorithm to obtain a non-dominated optimal solution set, and the non-dominated optimal solution set includes schedule adjustment schemes corresponding to various objective function values ​​of the multi-objective hybrid programming model, including:

[0039] The hybrid model is iteratively solved based on an optimized multi-objective non-dominated sorting genetic algorithm to obtain a non-dominated optimal solution set, including the following steps:

[0040] S41, population initialization;

[0041] S42, Generate offspring: Generate offspring solutions based on evolutionary crossover, selection, and mutation operators in the multi-objective evolutionary algorithm MOEA;

[0042] S43, solution selection: non-dominated sorting, and calculate the congestion degree, select the critical layer part of the solution into the solution set to select the offspring;

[0043] A non-dominated optimal solution set is obtained, which includes the values ​​of the objective function of minimizing the passenger waiting time, the objective function of minimizing the total number of train schedule adjustments, the objective function of minimizing the total adjustment time, and the objective function of minimizing the passenger travel cost, as well as the timetable adjustment plan corresponding to each objective function value.

[0044] Preferably, the method of obtaining a timetable optimization result that best meets passenger needs in the non-dominated optimal solution set according to the operator's decision preference by using a multi-criteria decision method includes:

[0045] The pseudo-weight method is selected as the decision method of the multi-criteria decision-making method, and the pseudo-weight vector of each non-inferior solution in the non-dominated optimal solution set is calculated and searched, and the maximum value f of each solution in the non-inferior solution set is searched. i max and the minimum value f i min , the weight w of the non-inferior solution i-th objective function is calculated by the following formula i :

[0046]

[0047] Where, M is the number of targets;

[0048] The decision maker inputs preference information and selects the non-inferior solution corresponding to the weight vector closest to the decision maker's preference through the decision function, which is defined as:

[0049]

[0050] Among them, d i is the decision function value of the ith solution in the solution set, w m is the i-th pseudo weight vector; is the weight vector preferred by the i-th decision maker;

[0051] The timetable optimization result that best meets the passenger needs is obtained according to the selected non-inferior solution, and the timetable optimization result includes: the passenger waiting time minimization objective function, the total adjustment amount of the train schedule minimization objective function, the total adjustment time minimization objective function and the passenger travel cost minimization objective function value, the departure time adjustment amount of each train in the multi-modal transportation mode, the multi-modal time-varying passenger flow sharing rate and the optimization model evaluation index value.

[0052] It can be seen from the technical solutions provided by the above-mentioned embodiments of the present invention that the method of the present invention fully considers the characteristics of various modes of transportation and the mutual influence between passengers and modes of transportation, constructs and solves the model, and effectively improves the convergence, stability, and diversity comprehensive performance of multi-mode timetable optimization, thereby providing passengers with the greatest convenience with the smallest adjustment, improving the efficiency and coordination of inter-city travel, and providing theoretical and technical support for the balance between inter-city passenger demand and capacity.

[0053] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0055] Figure 1 A schematic flow chart of a timetable optimization method taking into account passenger travel choices provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0057] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items.

[0058] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0059] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings, and each embodiment does not constitute a limitation on the embodiments of the present invention.

[0060] A processing flow chart of a timetable optimization method considering passenger travel choices provided by an embodiment of the present invention is as follows: Figure 1 As shown, the processing steps include the following:

[0061] Step S10, constructing a passenger travel generalized cost function including the time in the car, waiting time, walking time and travel cost parameters, and using the passenger travel generalized cost function to accurately characterize the travel choice behavior of passengers in multi-modal transportation modes;

[0062] Step S20: Taking the passenger generalized cost function as input, constructing a discrete choice model of passenger travel mode based on Weibull distribution, and using the discrete choice model to allocate passengers to various paths to obtain passenger flow data of each path;

[0063] Step S30: construct a multi-objective hybrid planning model based on the passenger flow data of each path, with the objectives of minimizing passenger waiting time, minimizing the total adjustment quantity and time of multi-modal trains, and minimizing passenger travel costs; and optimize the multi-modal timetable based on the passenger flow data of each path.

[0064] Step S40, adopting a multi-objective non-dominated sorting evolutionary algorithm (FHR-NSGA-III) to iteratively solve the multi-objective hybrid programming model to obtain a non-dominated optimal solution set, which includes four objective function values ​​and corresponding schedule adjustment schemes.

[0065] Step S50: obtaining a timetable optimization result that best meets passenger needs from the non-dominated optimal solution set according to the operator's decision preference by using a multi-criteria decision-making method such as a pseudo-weight method.

[0066] Furthermore, the multimodal transportation modes include high-speed rail, motor train, conventional rail, bus and private car; the optimization objectives in the multimodal timetable include minimizing passenger waiting time, minimizing the total number of train schedule adjustments, minimizing the total adjustment time, and minimizing passenger travel costs.

[0067] Furthermore, constructing the multi-modal passenger travel generalized cost function in step S10 includes the following steps:

[0068] According to the five multi-modal transportation modes of high-speed rail, motor train, ordinary rail, bus and private car, the corresponding travel cost components are defined respectively, namely, the time on the train, waiting time, walking time, travel cost, namely ticket price or fuel cost. The generalized cost function considering various influencing factors can better simulate the perceived cost of passengers' travel;

[0069] The time cost is converted into cost through the time value VOT = I / (W×D×60), realizing the unification of generalized cost. Among them, I is the average annual income of residents, W is the statutory working hours per day, and D is the statutory working days;

[0070] Constructing a generalized cost function for passenger travel for:

[0071]

[0072] in, is the time in the vehicle for different modes m, is the waiting time of different modes m, is the walking time of different modes m, p m is the travel cost of mode m, θ1, θ2, θ3, θ4 are weight parameters, which are 0.53, 0.17, 0.11, 0.19 respectively.

[0073] Furthermore, the construction of the discrete choice model of the passenger travel mode in step S20 includes:

[0074] The passenger generalized cost function is used as the input of discrete choice model to quantitatively study the impact between inter-city multi-modal timetable adjustment and passenger travel behavior;

[0075] The discrete choice model Multinominal weibit model of passenger travel mode based on Weibull distribution is constructed as follows:

[0076]

[0077] in, is the partial derivative of the joint generating function, is the independent Weibull distribution random error term of OD for mode m between ij, is the passenger perceived cost of mode m, is the average perceived cost, ζ ij for β ij is the shape parameter∈(0,∞), M ij is the set of all ways between OD pairs ij, and IJ is the set of all OD (Origin to Destination) pairs.

[0078] By solving the above discrete choice model, the passenger flow sharing rate of each path is obtained. The sum of the passenger flow sharing rates of all paths is 1. The total passenger flow is distributed to each path according to the calculated passenger flow sharing rate. The sum of the passenger flow on all paths is the total passenger flow.

[0079] Furthermore, the multi-objective hybrid planning model based on the passenger flow data of each path in step S30 is constructed with the objectives of minimizing the passenger waiting time, minimizing the total adjustment quantity and time of multi-mode trains, and minimizing the passenger travel cost, including:

[0080] The above multi-objective hybrid planning model adjusts the departure time of each mode and takes the changes in passenger mode selection into account in the timetable optimization model, so as to minimize the total adjustment at the operation level and shorten the total waiting time of passengers at the service level. Before modeling, multiple assumptions are made for the timetable optimization problem to be studied considering passenger travel choices, so as to reduce errors and determine decision variables while describing the timetable optimization problem;

[0081] In the multi-objective hybrid planning model, multiple objective functions are set at the passenger demand level and the transportation mode operation service level, including the objective function of minimizing the passenger waiting time, the objective function of minimizing the total number of train schedule adjustments, the objective function of minimizing the total adjustment time, and the objective function of minimizing the passenger travel cost. Specifically:

[0082] Passenger waiting time minimization objective function:

[0083]

[0084] The objective function of minimizing the number of train schedule adjustments is:

[0085]

[0086] The total adjustment time is minimized to minimize the objective function:

[0087]

[0088] The objective function to minimize the total travel cost of passengers is:

[0089]

[0090] in, Q is the waiting time of passenger j between trips i. i is the adjustment status of train number i, which is 1 if adjusted, otherwise 0, i.e., the adjustment time is 0min. i Arrival after is the departure time of train number i after adjustment, C ij is the fare of train number i. i Arrival before To adjust the departure time of the previous train number i, I is the set of train number i, J is the set of passenger j, is the total waiting time of passengers, Q total Adjust the total number of train schedules. is the total adjustment time of all trains, C total is the total travel cost for all passengers.

[0091] The constraints of the above multi-objective hybrid planning model are set in consideration of the actual situation, and the constraints include departure interval constraints, path feasibility constraints, time adjustment range constraints, capacity constraints, on-board status constraints, time window constraints and passenger selection constraints.

[0092] The multi-modal timetable is the solution result of the above multi-objective hybrid planning model. The adjustment amount of each train in the timetable is the decision variable of the hybrid planning model, which is reflected in the solution set. As long as the model is solved, the multi-modal timetable optimization solution can be obtained. The purpose of optimizing the multi-modal timetable is to make passengers travel more convenient, such as reducing costs and time. Therefore, passenger flow data provides data support for the optimization of the timetable. The first and fourth optimization goals set by the model are closely related to passenger flow.

[0093] The solution of the above multi-objective hybrid programming model is an optimal solution set, including four optimization target values ​​and corresponding schedule optimization schemes of the same number as the population. Each solution corresponds to an optimization scheme, and there is no objective distinction between the optimization schemes. In the multi-objective decision-making set later, the user's subjective factors need to be brought into the decision-making process, and a unique optimal scheme that best satisfies the user's preferences is selected.

[0094] Furthermore, the multi-objective non-dominated sorting genetic algorithm in step S40 includes the following steps:

[0095] ① Population initialization; generation of offspring:

[0096] ② Generate offspring solutions based on the evolutionary crossover, selection, and mutation operators commonly used in the classic MOEA (Multi-objective Evolutionary Algorithm);

[0097] ③Solution selection: non-dominated sorting, and calculate the congestion degree, select some solutions in the critical layer to enter the solution set to select the offspring.

[0098] Furthermore, in step S50, a multi-criteria decision-making method such as a pseudo-weight method is used to optimize the multi-modal timetable in the non-dominated optimal solution set according to the operator's decision preference, so as to obtain a timetable optimization result that best meets the passenger needs.

[0099] The multi-criteria decision-making method is:

[0100] Multi-criteria decision making is a decision to choose from a set of conflicting and incommensurable solutions. The pseudo-weight method is used as the decision-making method to calculate and search for the pseudo-weight vector of each non-inferior solution, and select the solution corresponding to the weight vector closest to the decision maker's preference;

[0101] Calculate the pseudo weight vector. Search for the maximum value f of each solution in the non-inferior solution set i max and the minimum value f i min , the weight w of the non-inferior solution i-th objective function is calculated by the following formula i :

[0102]

[0103] Where, M is the number of targets;

[0104] The decision maker inputs preference information and selects the closest non-inferior solution. The decision function is defined as:

[0105]

[0106] Among them, d i is the decision function value of the ith solution in the solution set, w m is the i-th pseudo weight vector; is the weight vector preferred by the i-th decision maker.

[0107] The timetable optimization results include: passenger waiting time minimization objective function, total train schedule adjustment amount minimization objective function, total adjustment time minimization objective function and passenger travel cost minimization objective function value, multi-modal transportation mode each train departure time adjustment amount, multi-modal time-varying passenger flow sharing rate, and optimization model evaluation index value.

[0108] The multimodal timetable is part of the optimization result, which includes the departure time adjustments of each train of multimodal transportation modes.

[0109] Example 1

[0110] A specific embodiment 1 of the present invention discloses a timetable optimization method considering passenger travel choices, such as Figure 1 As shown, including:

[0111] S1. Construct a generalized cost function for passenger travel that includes on-board time, waiting time, walking time, and travel cost parameters to accurately characterize passenger travel choice behavior.

[0112] During implementation, the corresponding travel cost components are defined according to the five multimodal transportation modes, namely high-speed rail, motor train, conventional rail, bus and private car, namely, on-board time, waiting time, walking time, travel cost, i.e. ticket price or fuel fee. The generalized cost function that takes into account various influencing factors can better simulate the passengers' perceived travel cost.

[0113] Specifically, the time cost is converted into cost through the time value VOT = I / (W×D×60), realizing the unification of generalized cost. Among them, I is the average annual income of residents, W is the statutory working hours per day, and D is the statutory working days.

[0114] Constructing a generalized cost function based on uniform cost for:

[0115]

[0116] in, is the time in the vehicle for different modes m, is the waiting time of different modes m, is the walking time of different modes m, p m is the travel cost of mode m, θ1, θ2, θ3, θ4 are weight parameters, which are 0.53, 0.17, 0.11, 0.19 respectively.

[0117] S2. Using the generalized cost function as input, a discrete choice model of passenger travel modes based on Weibull distribution is constructed, passengers are assigned to roads, and passenger flow data of each path is obtained. The intercity travel modes selected by passengers include high-speed rail, EMU, conventional rail, bus and private car.

[0118] When implementing, the discrete choice model generally uses the Logit model, but because the variance of the utility function is fixed in the Logit model, the perceived error of each alternative remains unchanged, which is inconsistent with the actual situation. The Weibit model changes the utility function from addition to multiplication, which solves the heterogeneity problem in the Logit model to a certain extent.

[0119] Specifically, the generalized cost function is used as the input of the discrete choice model to quantitatively study the impact of inter-city multimodal schedule adjustment on passenger travel behavior. The Multinominal weibit model of passenger travel mode based on Weibull distribution is:

[0120]

[0121] in, is the partial derivative of the joint generating function, is the independent Weibull distribution random error term of OD for mode m between ij, is the passenger perceived cost of mode m, is the average perceived cost, ζ ij for β ij is the shape parameter ∈(0,∞), M ij is the set of all ways between OD pairs ij, and IJ is the set of all OD pairs.

[0122] S3. Construct a multi-objective hybrid planning model with the goals of minimizing passenger waiting time, minimizing the total adjustment quantity and time of multi-modal trains, and minimizing passenger travel costs, and optimize the multi-modal timetable based on the passenger flow data.

[0123] S31. During implementation, multiple assumptions are made for the timetable optimization problem to be studied considering passenger travel choices before modeling, so as to reduce errors and determine decision variables while describing the timetable optimization problem. Specifically:

[0124] The results of passenger mode selection are determined according to the probabilities obtained from the discrete choice model, and the sum of the results is 1;

[0125] The route is fixed, and passengers who choose a certain mode of transportation will take the next train of the same mode of transportation;

[0126] Trains depart and arrive on time according to schedule;

[0127] A station can serve multiple trips at the same time. A track or lane can serve multiple trains.

[0128] S32, setting multiple objective functions at the passenger demand level and the transportation mode operation service level, wherein the optimization objectives include minimizing the passenger waiting time, minimizing the total number of train schedule adjustments, minimizing the total adjustment time, and minimizing the passenger travel cost. Specifically:

[0129] Minimize passenger waiting time:

[0130]

[0131] Minimize the number of train schedule adjustments:

[0132]

[0133] Minimize total adjustment time:

[0134]

[0135] Minimize the total travel cost of passengers:

[0136]

[0137] in, Q is the waiting time of passenger j between trips i. i The adjustment status of train number i is 1 if it is adjusted, otherwise it is 0. i Arrival after is the departure time of train number i after adjustment. ij is the fare for train number i.

[0138] S33. Considering the actual situation, set constraints, including departure interval constraints, path feasibility constraints, time adjustment range constraints, capacity constraints, on-board status constraints, time window constraints, and passenger selection constraints. Specifically:

[0139] Constraints on the departure interval of trains on the same track:

[0140]

[0141] Constraints on the same route departure interval:

[0142]

[0143] Specifically, the track allocation of stations should be reasonably set to allow overtaking. Necessary arrival and departure intervals should be set not only between adjacent trains on the same track, but also between stations to prevent trains from being dispatched in a short time.

[0144] Path feasibility constraints:

[0145]

[0146] Adjust the threshold constraints at all times:

[0147]

[0148] Train capacity constraints:

[0149]

[0150] Passenger in-car status constraints:

[0151]

[0152] If passenger j is on train i at time t, b ijt =1, otherwise 0.

[0153] Passenger boarding time window constraints:

[0154]

[0155] Waiting time constraints:

[0156]

[0157]

[0158] Passenger travel choice constraints:

[0159]

[0160] Among them, T i stm is the minimum departure interval between adjacent trains on the same track, T i am is the maximum time adjustment value of train i, M i is the capacity of train i, T i wm is the maximum waiting time of train i, L i Walking distance to get on the bus.

[0161] S4, iteratively solving the hybrid model based on the optimized multi-objective non-dominated sorting genetic algorithm (FHR-NSGA-III algorithm) to obtain a non-dominated optimal solution set. The following steps are included:

[0162] S41, population initialization;

[0163] S42, Generate offspring: Generate offspring solutions based on the evolutionary crossover, selection, and mutation operators commonly used in classic MOEA;

[0164] S43, solution selection: non-dominated sorting, and calculate the congestion degree, select the critical layer part of the solution to enter the solution set to select the offspring. This step is the optimization process of the existing multi-objective non-dominated sorting genetic algorithm.

[0165] Specifically, the hypervolume index is the area or volume between the Pareto solution set and the coordinate axis. It is a comprehensive index to measure the convergence and uniformity of the solution set. The larger the hypervolume value, the better the algorithm effect and the higher the quality of the solution set. For multi-objective evolutionary algorithms based on indicators, reducing complexity is a key issue.

[0166] In implementation, for a solved two-objective problem, the solution set is S = {a, b, c, d, e}, five points are distributed on the Pareto front, and the reference set is R = {r}. The hypervolume of the solution set is defined as HV(S, R), which is the volume enclosed by R and a, b, c, d, e. The hypervolume contribution of a single solution, such as solution b, is defined as HV att (b,S,R), the formula is:

[0167] HV att (b,S,R)=HV(S,R)-HV(Sb,R)

[0168] Specifically, define a bad solution function:

[0169] W = nondominated-worse(b,Sb)

[0170] Specifically, w1, w2, w3 and w4 are inferior solutions among ba, bc, bd and be respectively. W = nondominated({w1, w2, w3, w4}) = {w1, w3}, where w2 and w4 are dominated by w1 and w3 respectively. HV att The calculation method of (b,S,R) is:

[0171] HV att (b,S,R)=HV(b,R)-HV(W,R)

[0172] This optimization calculation method is combined with an indicator-based multi-objective evolutionary algorithm, defined as the FH-MOEA algorithm, but the calculation results of this algorithm have strong convergence and poor distribution, while the solutions obtained by searching the other reference point-based R-NSGA-III algorithm are uniformly distributed. In order to improve the diversity and convergence of the solution set, a FHR-NSGA-III algorithm is proposed, which combines FH-MOEA with R-NSGAIIII. This algorithm can obtain a population with better convergence, which is then used as the initial population to further iterate and optimize its distribution and diversity.

[0173] S5. Based on the pseudo-weight method, a multi-criteria decision-making method is used to obtain a multi-modal schedule that best meets passenger needs in the solution set according to the operator's decision preferences.

[0174] Multi-criteria decision making (MCDM) is a decision making method for choosing between conflicting and incommensurable solutions. Pseudo-weight methods calculate pseudo-weight vectors for each non-dominated solution and select the solution corresponding to the weight vector that is closest to the decision maker's preference. The decision maker's preferences are subjective and need to be adjusted according to their needs. Pseudo-weight methods can find the corresponding optimal solution based on preferences.

[0175] Specifically, the pseudo weight vector is calculated. Search for the maximum value f of each solution in the non-inferior solution set i max and the minimum value f i min , the weight w of the non-inferior solution i-th objective function is calculated by the following formula i :

[0176]

[0177] Where, M is the number of targets;

[0178] The decision maker inputs preference information and selects the closest non-inferior solution. Decision function definition:

[0179]

[0180] Among them, d i is the decision function value of the ith solution in the solution set, w m is the i-th pseudo weight vector; is the weight vector preferred by the i-th decision maker.

[0181] Example 2

[0182] In order to verify the validity of the model in Example 1 of the present invention, the specific embodiment 2 of the present invention uses a Lenovo PC for experiment. Core (TM) i7-10700CPU @ 2.90GHz, RAM is 16GB. The operating system is a 64-bit operating system Windows 10Enterprise Edition based on an x64 processor. The software version is PyCharm 2020.3.2, and the program operation is based on the pymoo platform. In order to test the efficiency and comprehensive performance of the proposed algorithm, this embodiment conducts experiments on the inter-city comprehensive transportation network in three different urban agglomerations in Beijing and Zhangjiakou, Chengdu and Chongqing, Guangzhou and Qingyuan, and also selects HV-MOEA, AGE-MOEA, AGE2-MOEA, SMS-EMOA, REVA, C-TAEA, NSGA-II and R-NSGAIII for comparative analysis. All algorithms are run independently 10 times under the same conditions. Among them, the initial population size is 40, the number of iterations is 2000, simulated binary crossover and polynomial mutation are used for real and int types, and semi-uniform crossover and bit flip mutation are used for bool types, with a crossover rate of 0.8 and a mutation rate of 0.2.

[0183] The experimental results show that:

[0184] (1) The efficiency of the FHR-NSGA-III algorithm in obtaining feasible solutions is extremely excellent among all algorithms. Although its comprehensive index is not the best, compared with the situation where some algorithms perform well on a certain road network but poorly on another road network, the algorithm proposed in this embodiment performs well and stably on the three road networks, showing better comprehensive performance. This means that the algorithm is fast, excellent and stable under the three data sets.

[0185] (2) Due to the instability of computer performance, each algorithm was set to run 10 times, and the running time of each algorithm had a small fluctuation. However, except for the long running time of C-TAEA and RVEA, the running time of other algorithms was basically the same. In terms of average running time, NSGA-II, AGE-MOEA, AGE2-MOEA and FHR-NSGA-III have the best performance.

[0186] (3) FHR-NSGA-III is not the best in terms of solution quality, especially in the Beijing-Zhangjiakou and Chengdu-Chongqing networks, where its solution is similar to that of SMS-EMOA, AGE-MOEA, AGE2-MOEA, and C-TAEA, but its victory lies in stability. It can solve stably under three different data sets and shows good convergence and distribution. In short, FHR-NSGA-III is the best choice to reduce the total waiting time and travel costs of passengers under smaller time adjustments.

[0187] In summary, FHR-NSGA-III performs better in multi-mode schedule collaborative optimization.

[0188] In summary, an embodiment of the present invention provides a timetable optimization method that takes into account passenger travel choices. First, a generalized cost function for passenger travel that includes on-board time, waiting time, walking time, and travel cost parameters is constructed to accurately characterize passenger travel choice behavior. Secondly, taking the generalized cost function as input, a discrete choice model for passenger travel modes based on Weibull distribution is constructed, and passengers are assigned to roads to obtain passenger flow data for each path. Then, a multi-objective hybrid planning model is constructed with the goals of minimizing passenger waiting time, minimizing the total adjustment number and time of multi-modal trains, and minimizing passenger travel costs, and the multi-modal timetable is optimized based on the passenger flow data. Finally, based on the fusion multi-objective evolutionary algorithm (FHR-NSGA-III) and the pseudo-weight method, a multi-modal timetable that best meets passenger needs is obtained according to the operator's decision preferences.

[0189] The method of the present invention fully considers the characteristics of various modes of transportation and the mutual influence between passengers and modes of transportation, constructs and solves the model, effectively improves the convergence, stability, and diversity comprehensive performance of multi-mode timetable optimization, provides the greatest convenience for passengers with the smallest adjustment, improves the efficiency and coordination of inter-city travel, and provides theoretical and technical support for the balance between inter-city passenger demand and capacity.

[0190] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0191] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention or certain parts of the embodiments.

[0192] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0193] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for optimizing a timetable taking into account passenger travel choices, characterized in that: include: Construct a generalized cost function for passenger travel that includes the time in the car, waiting time, walking time and travel cost parameters, and use the generalized cost function for passenger travel to characterize the travel choice behavior of passengers in multi-modal transportation modes; Taking the passenger generalized cost function as input, a discrete choice model of passenger travel mode based on Weibull distribution is constructed, and the passengers are allocated to each path by using the discrete choice model to obtain the passenger flow data of each path; Based on the passenger flow data of each route, a multi-objective hybrid planning model is constructed with the goals of minimizing passenger waiting time, minimizing the total adjustment quantity and time of multi-modal trains, and minimizing passenger travel costs. Iteratively solving the multi-objective hybrid programming model using a multi-objective non-dominated sorting evolutionary algorithm to obtain a non-dominated optimal solution set, the non-dominated optimal solution set including schedule adjustment schemes corresponding to various objective function values ​​of the multi-objective hybrid programming model; Obtaining a timetable optimization result that best meets passenger demand from the non-dominated optimal solution set according to the operator's decision preferences through a multi-criteria decision method; The multi-objective hybrid planning model based on the passenger flow data of each path is constructed with the objectives of minimizing the waiting time of passengers, minimizing the total adjustment quantity and time of multi-mode trains, and minimizing the travel cost of passengers, including: In the multi-objective hybrid planning model, multiple objective functions at the passenger demand level and the transportation mode operation service level are set based on the passenger flow data of each path. The objective functions include the objective function of minimizing the passenger waiting time, the objective function of minimizing the total number of train schedule adjustments, the objective function of minimizing the total adjustment time, and the objective function of minimizing the passenger travel cost. Specifically, Passenger waiting time minimization objective function: The objective function of minimizing the number of train schedule adjustments is: The total adjustment time is minimized to minimize the objective function: The objective function to minimize the total travel cost of passengers is: in, is the waiting time of passenger j between trips i, Q i is the adjustment status of train number i, which is 1 if adjusted, otherwise 0, that is, when the adjustment time is 0min, T i Arrivalafter is the departure time of train number i after adjustment, C ij is the fare of train number i, T i Arrivalbefore To adjust the departure time of the previous train number i, I is the set of train number i, J is the set of passenger j, is the total waiting time of passengers, Q total Adjust the total number of train schedules. is the total adjustment time of all trains, C total is the total travel cost for all passengers; The constraints of the multi-objective hybrid programming model are set, wherein the constraints include departure interval constraints, path feasibility constraints, time adjustment range constraints, capacity constraints, on-board status constraints, time window constraints and passenger selection constraints.

2. The method according to claim 1, characterized in that: The construction includes a passenger travel generalized cost function including the parameters of travel time, waiting time, walking time and travel cost, and uses the passenger travel generalized cost function to characterize the travel choice behavior of passengers in multi-modal transportation modes, including: Based on the five multi-modal transportation modes of high-speed rail, motor train, ordinary rail, bus and private car, the travel cost components corresponding to each transportation mode are defined respectively, and the travel cost components include the time in the car, waiting time, walking time and travel cost; The calculation formula of time value VOT is: VOT=I / (W×D×60) Through the time value VOT, time expenses are converted into costs to achieve the unification of generalized expenses, where I is the average annual income of residents, W is the statutory working hours per day, and D is the statutory working days; Constructing a generalized cost function for passenger travel for: in, is the time in the vehicle for different modes m, is the waiting time of different modes m, is the walking time of different modes m, p m is the travel cost of mode m, θ1, θ2, θ3, θ4 are the set weight parameters.

3. The method according to claim 2, characterized in that The method of constructing a discrete choice model of passenger travel mode based on Weibull distribution by taking the passenger generalized cost function as input, allocating passengers to various paths by using the discrete choice model, and obtaining passenger flow data of each path includes: The passenger generalized cost function is used as the input of discrete choice model to quantitatively study the impact between inter-city multi-modal timetable adjustment and passenger travel behavior; The discrete choice model Multinominal weibit model of passenger travel mode based on Weibull distribution is constructed as follows: in, is the partial derivative of the joint generating function, is the independent Weibull distribution random error term of OD for mode m between ij, is the passenger perceived cost of mode m, is the average perceived cost, ζ ij is the position parameter ∈ β ij is the shape parameter∈(0,∞), M ij is the set of all ways between OD pairs ij, and IJ is the set of all OD pairs from the starting point to the end point; By solving the above discrete choice model, the passenger flow sharing rate of each path is obtained. The sum of the passenger flow sharing rates of all paths is 1. The total passenger flow is distributed to each path according to the calculated passenger flow sharing rate. The sum of the passenger flow on all paths is the total passenger flow.

4. The method according to claim 1, characterized in that: The multi-objective non-dominated sorting evolutionary algorithm is used to iteratively solve the multi-objective hybrid programming model to obtain a non-dominated optimal solution set, which includes schedule adjustment schemes corresponding to various objective function values ​​of the multi-objective hybrid programming model, including: The hybrid model is iteratively solved based on an optimized multi-objective non-dominated sorting genetic algorithm to obtain a non-dominated optimal solution set, including the following steps: S41, population initialization; S42, Generate offspring: Generate offspring solutions based on evolutionary crossover, selection, and mutation operators in the multi-objective evolutionary algorithm MOEA; S43, solution selection: non-dominated sorting, and calculate the congestion degree, select the critical layer part of the solution into the solution set to select the offspring; A non-dominated optimal solution set is obtained, which includes the values ​​of the objective function of minimizing the passenger waiting time, the objective function of minimizing the total number of train schedule adjustments, the objective function of minimizing the total adjustment time, and the objective function of minimizing the passenger travel cost, as well as the timetable adjustment plan corresponding to each objective function value.

5. The method according to claim 4, characterized in that The method of obtaining a timetable optimization result that best meets passenger needs in the non-dominated optimal solution set according to the operator's decision preference through the multi-criteria decision method includes: The pseudo-weight method is selected as the decision method of the multi-criteria decision-making method, and the pseudo-weight vector of each non-inferior solution in the non-dominated optimal solution set is calculated and searched, and the maximum value f of each solution in the non-inferior solution set is searched. i max and the minimum value f i min , the weight w of the non-inferior solution i-th objective function is calculated by the following formula i : Where, M is the number of targets; The decision maker inputs preference information and selects the non-inferior solution corresponding to the weight vector closest to the decision maker's preference through the decision function, which is defined as: Among them, d i is the decision function value of the ith solution in the solution set, w m is the i-th pseudo weight vector; is the weight vector preferred by the i-th decision maker; The timetable optimization result that best meets the passenger needs is obtained according to the selected non-inferior solution, and the timetable optimization result includes: the passenger waiting time minimization objective function, the total adjustment amount of the train schedule minimization objective function, the total adjustment time minimization objective function and the passenger travel cost minimization objective function value, the departure time adjustment amount of each train in the multi-modal transportation mode, the multi-modal time-varying passenger flow sharing rate and the optimization model evaluation index value.