A method for setting up a transfer bus station and line based on grid optimization

Abstract

The application discloses a method for setting a transfer bus station and a line based on grid optimization, and comprises the following steps: 1, based on travel big data, respectively calculating travel demand distribution indexes of each unit grid in a region; 2, respectively calculating total travel demand and regional competitive advantage of each evaluation grid by using an evaluation grid optimization method; 3, searching and optimizing the evaluation grid of the site coverage range to respectively calculate and obtain quantitative evaluation values of the evaluation grid and the centroid longitude and latitude of the evaluation grid, thereby obtaining an evaluation grid quantitative evaluation value set in a certain range according to a certain step; 4, determining the optimal unit grid of the site selection of each evaluation grid as the optimal site selection position of the transfer station based on travel cost, and forming a bus driving route scheme; and 5, establishing a multi-objective optimization model to optimize the bus driving route to obtain a final bus driving route scheme, wherein the decision variables of the model optimization include the bus driving route and the number of vehicles equipped on the line. The application realizes scientific and accurate site selection for a city public transport stop.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of urban traffic planning, and particularly relates to a feeder bus station and line layout method based on grid optimization. BACKGROUND

[0002] With the continuous expansion of the city scale, rail transit plays an increasingly important role in public transportation, connecting various groups in the city and effectively alleviating the problem of long-distance travel. However, due to various limitations (cost and construction standards), rail transit is mainly laid along areas with high development intensity, and the accessibility of peripheral functional areas or areas under development needs to be improved. Walking, shared bicycles, taxis and feeder buses are often used to solve such problems, and feeder buses have become an important way to connect trunk traffic due to their high efficiency, reliability and long connection distance. Feeder bus is a kind of auxiliary passenger transport system that serves passengers to and from designated locations, which can effectively improve the radiation range of rail stations, connect travel supply and demand points, increase the attractiveness of rail transportation and reduce travel time.

[0003] Since the current feeder bus system demand data collection method is relatively single, most cases are based on survey data or consider the existing operation, so it cannot accurately capture travel demand, resulting in unscientific feeder bus station or line layout and mismatched supply and demand distribution.

[0004] In the research of feeder bus station site selection and line layout, the existing research often takes the demand point of resident travel as the pickup point, or clusters the demand points as the station site selection scheme. The clustering algorithm often considers the demand of resident travel as the factor, ignores the cost of resident travel, the path of resident travel to the station and the road conditions of station setting, and it is difficult to determine the cluster station position for the scattered demand points. In addition, the position of the cluster station is automatically solved based on the geometric space of the clustering algorithm, and it is difficult to determine the specific point of the station. The line layout problem is generally solved by optimization model, taking the minimum travel cost of passengers and the minimum travel time of passengers as the objective function, but in actual situation, the station layout and line layout often involve multiple conflicting objectives, such as maximizing service coverage or minimizing operation cost, and single objective optimization may not fully consider the comprehensive influence of various factors. SUMMARY

[0005] The present application relates to the technical field of urban traffic planning, and particularly relates to a feeder bus station and line layout method based on grid optimization.

[0006] The technical scheme adopted by the present application is:

[0007] A feeder bus station and line layout method based on grid optimization, comprising the following steps:

[0008] Step 1, obtain the total amount of public transport traffic demand data in the cell based on travel big data; and set the size to divide the corresponding area into a plurality of unit grids, and calculate the travel demand distribution index of each unit grid in the area;

[0009] Further, the travel big data in step 1 is multi-source data, which includes mobile signaling data, bus and rail boarding and alighting card data, shared bicycle travel order data, and high-precision map data.

[0010] Further, step 1 calculates the share rate of public transport and shared bicycle traffic modes; taking a unit square of 150m*200m as the calculation unit, the travel demand distribution index of each unit grid in the area is obtained.

[0011] Further, step 1 specifically includes the following steps:

[0012] Step 1-1, extract the mobile signaling data of all travelers within the service radius of the rail station;

[0013] Step 1-2, pre-process the data, eliminate invalid data, and supplement missing fields for data cleaning;

[0014] Step 1-3, filter the early morning peak period on weekdays, identify the mobile signaling data based on historical data, and identify the early morning peak period based on the cell as the starting point and the rail station as the ending point, to obtain the total amount of travel and travel demand distribution from the cell to the rail station in the area;

[0015] Step 1-4, based on the mobile signaling data, combined with the total amount of travel and travel demand distribution, identify the user's travel mode.

[0016] Further, the specific method for step 1-4 to identify the user's travel mode is:

[0017] (1) For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the mobile signaling data with an average speed less than 10Km / h and a distance from the rail station less than 2Km is identified as walking travel;

[0018] (2) For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the mobile signaling data with an average speed less than 10Km / h and a distance from the rail station less than 2Km is identified as walking travel;

[0019] (3) For communities with shared bicycles available, the total number of shared bicycle connections is obtained by combining shared bicycle order data with the travel demand for rail stations.

[0020] (4) Subtract the pedestrian traffic volume, private electric vehicle traffic volume, public transport connection traffic volume, and shared bicycle connection traffic volume from the total mobile phone signaling data volume, and identify the remaining traffic volume as travel by private car.

[0021] Step 2: Using the evaluation grid optimization method, calculate the total travel demand for each evaluation grid. and regional competitive advantage The evaluation grid is a rectangular grid composed of several unit grids;

[0022] Furthermore, step 2 specifically includes the following steps:

[0023] Step 2-1: Calculate the evaluation grid based on the travel demand distribution index of each unit grid. j Total demand for all transportation The calculation formula is:

[0024]

[0025] Among them, cell i This represents the i-th cell grid within the region. Let i represent the traffic demand of the i-th cell grid, where i = 1, ..., n, and n is the total number of cell grids; the side length of each cell grid is 150-200m.

[0026] Step 2-2: Calculate the competitive advantage of the evaluation grid area. The specific calculation formula is as follows:

[0027]

[0028] in, This represents the average probability of selecting a shared bicycle within the evaluation grid. This represents the average probability of selecting a bus within the evaluation grid; R is the maximum distance from the transportation hub station within the area, and R is a fixed value. This represents the distance between the centroid of the evaluation grid and the transportation hub station;

[0029] Furthermore, the specific steps of step 2-2 are as follows:

[0030] Step 2-2-1: Calculate the average selection probability of choosing shared bicycles in the evaluation grid. The specific calculation formula is as follows:

[0031]

[0032] wherein, is the total demand of all trips in the jth evaluation grid; is the total demand of all trips in the jth evaluation grid, respectively, j is the total amount of trips from the bicycle to the transportation hub in the jth evaluation grid, and the corresponding data can be obtained by counting the number of shared bicycle orders;

[0033] Step 2-2-2, the average selection probability of choosing a bus in the evaluation grid is calculated The specific calculation formula is as follows:

[0034]

[0035] wherein, is the total amount of trips from the regular bus to the transportation hub in the jth evaluation grid, and the corresponding data can be obtained by counting the number of bus card swipes; j

[0036] Step 2-2-3, the distance between the grid centroid of the evaluation grid and the transportation hub site is calculated The specific calculation formula is as follows:

[0037]

[0038] wherein, R c is the equatorial radius, x sta and y sta are the coordinates of the transportation hub site in radian system; and are the longitude and latitude of the grid centroid of the jth evaluation grid grid j in radian system;

[0039] wherein, x sta and y sta are the longitude and latitude of the transportation hub site; and are the longitude and latitude of the grid centroid of the jth evaluation grid grid j ;

[0040] Step 2-2-4, the regional competitive advantage of the evaluation grid is calculated The specific calculation formula is as follows:

[0041]

[0042] wherein, represents the average selection probability of choosing a shared bicycle in the evaluation grid, represents the average selection probability of choosing a bus in the evaluation grid; R is the maximum distance from the transportation hub site within the region, and the value of R is a fixed value; ​

[0043] Step 3, the area of the site coverage range is searched and optimized by using the evaluation grid, and the quantitative evaluation value of each evaluation grid grid j is calculated The centroid longitude and latitude of the evaluation grid And A three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} is constructed Get all grid quantitative evaluation value sets evaluated within a certain range with a certain step size

[0044] Further, the specific steps of step 3 are as follows:

[0045] Step 3-1, the total demand of each evaluation grid and the regional competitive advantage are added to calculate the quantitative evaluation value of each evaluation grid grid j , that is:

[0046]

[0047] Where, θ represents the scaling factor, representing the different inclination of grid quantitative evaluation value to demand and regional competitive advantage, and θ value is obtained according to actual situation;

[0048] Specifically, the entire area is traversed by continuously translating the grid, and the average distance between subway stations is taken as the unit step length L of translation, so as to obtain the quantitative value of each region and sort the quantitative value evaluation array in descending order. The invention needs to cover all areas around the subway station as much as possible, so it is necessary to select a suitable grid length to surround all areas without causing the grid to be too large. Specifically, the distance between stations is generally 400-600m, which is obtained according to actual demand.

[0049] Step 3-2, the centroid longitude and latitude of the evaluation grid And A three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} is constructed Get all grid quantitative evaluation value sets evaluated within a certain range with a certain step size

[0050] Step 4, determine the optimal unit grid of each evaluation grid based on the travel cost to determine the optimal site location of the station, form a bus driving route scheme; specifically including the following steps:

[0051] Step 4-1, the corresponding unit grid through which the road passes is included in the candidate unit grid set: the specific steps are as follows ​

[0052] Step 4-1-1, establishing an evaluation grid grid j and the corresponding relationship of all cell grids {cell1, … cell n} under it;

[0053] Step 4-1-2, selecting all road sets {road1, … road p} of the candidate setting connection stations based on road level and road connectivity factors;

[0054] Step 4-1-3, position association of {road1, … road p} and {cell1, … cell n}; when the corresponding cell grid passed by the road is included in the final set of candidate cell grids, {cell2', … cell m '} is obtained.

[0055] Step 4-2, selecting the standby grid with the minimum total cost in the candidate cell grid set as the optimal site location of the station:

[0056] Step 4-2-1, calculating the total cost of each cell grid of {cell1, … cell n} to each candidate grid;

[0057] Step 4-2-2, selecting the cell grid with the minimum total cost in the candidate set as the optimal cell grid of the site location of the current evaluation grid;

[0058] Specifically, the station sequence of the cell grid with the minimum total cost (i.e. the optimal position station under the grid) is u r (r = 1, …, w).

[0059] Step 5, establishing a multi-objective optimization model to optimize the bus driving route to obtain the final bus driving route scheme, and the decision variables of the model optimization include the bus driving route (i.e. the sequence of stations) and the number of vehicles equipped on the line. The specific steps are as follows:

[0060] Step 5-1, determination of the objective function: calculate the total cost of vehicle cost, the total cost of resident travel and the total demand of resident travel to establish the objective function; construct the constraint function of passenger service constraint, running time constraint, vehicle line constraint and vehicle capacity constraint;

[0061] Further, the specific steps of the multi-objective function are as follows:

[0062] Step 5-1-1, total vehicle driving cost: assuming that the operation departure interval of the vehicle in the peak period is t r , then the total cost of vehicle driving should be the sum of the driving distance of each vehicle on the line, i.e.:

[0063]

[0064] where K is the total time of peak commuting service, is the actual travel distance from the rth station u r to the r+1th station u r+1 ;

[0065] Step 5-1-2, Total travel time cost of residents: The travel time of passengers taking community buses at feeder rail transit stations includes three parts: walking to the bus stop, waiting time, and in-vehicle time, i.e.,

[0066] C2 = T h + T w + T s (12)

[0067] where T h is the total time of residents walking to the bus stop, T w is the waiting time of residents, and T s is the in-vehicle time;

[0068] For station u r , the total walking travel time is equal to Then the total walking travel time of residents for all stations T h is:

[0069]

[0070] Assuming that the vehicle arrival law is uniformly distributed, the waiting time at the station is equal to t r / 2, then the total waiting time of residents for all stations T w is:

[0071]

[0072] Based on the travel data, the starting point of passengers is obtained, i.e., the boarding and alighting quantity of each station sequence, then for u r to u r+1 station sequence, the total in-vehicle time of passengers is:

[0073]

[0074] where t is the travel time between two stations;

[0075] Step 5-1-3, Maximum number of passenger demand satisfaction: The specific expression is as follows:

[0076]

[0077] Step 5-1-4, construct the multi-objective function, the specific expression is:

[0078] minC = C1 + C2 - C3 (18);

[0079] The calculation steps of the constraint function are:

[0080] Step 5-1-21, passenger service constraint: each passenger can only be served by one vehicle

[0081]

[0082] Where, x b is a 0-1 variable, x b is 1 when the passenger is served, otherwise it is 0;

[0083] Step 5-1-22, running time constraint, i.e. the single running time of each community bus is limited within the maximum running time T max The corresponding expression is:

[0084]

[0085] Step 5-1-23, vehicle route constraint, i.e. the route length of each bus single run is controlled within a certain range, the corresponding expression is:

[0086]

[0087] Where, L min represents the minimum route length of single run; L max represents the maximum route length of single run;

[0088] Step 5-1-24, vehicle capacity constraint: i.e. the number of passengers carried by each community bus is not allowed to exceed the capacity C of the vehicle;

[0089]

[0090] Where, represents the number of passengers on the bus from the rth station u r to the r+1th station u r+1 .

[0091] Step 5-2, solution of optimal solution: based on the constructed constraint function, the multi-objective function is solved by applying NSGA-II algorithm.

[0092] Further, the multi-objective solution in step 5-2 specifically includes the following steps:

[0093] Step 5-2-1, randomly generate an initial population P0 and perform non-inferior sorting, perform binary tournament selection, crossover and mutation on the population to obtain a new population Q0, and let t=0;

[0094] Step 5-2-2, form a new population R t t ∪Q t Again, perform non-inferior sorting to obtain F1, F2, …;

[0095] Step 5-2-3, perform crowding degree sorting on all F, select the optimal N individuals to form a population P t+1 ; Wherein, F represents the set composed of all feasible solution populations R t of NSGAII algorithm;

[0096] Step 5-2-4, execute step 5-2-1 on population P t+1 to obtain Q t+1 ;

[0097] Step 5-2-5, if the termination condition is reached, the algorithm ends; otherwise, go to step 5-2-2.

[0098] Specifically, taking the rail station as the center, a certain distance as the radius, the service range of the rail station is determined, the rail station and the surrounding travel demand points are connected through the feeder bus, the demand distribution in the service range is determined, and the layout position of the feeder bus station is determined considering the influence of various factors, and the driving route of the feeder bus is determined.

[0099] The above technical scheme is adopted in the present application, and the following technical advantages are achieved:

[0100] (1) The present application combines mobile signaling data, bus and rail boarding and alighting card data, shared bicycle travel order data and high-precision map data and other multi-source travel big data to obtain the total amount of regional public transportation commuting travel demand, improving the accuracy and completeness of travel distribution.

[0101] (2) The present application considers the coverage range and site spacing of public transportation stations, aggregates the total amount of travel demand in the grid and the quantitative evaluation value of the location competitive advantage, evaluates the site selection location of the public transportation station based on the double-factor evaluation result, and aggregates the grid with high spatial autocorrelation and low demand to determine the optimal site selection location of the public transportation station. The present application solves the problems of the existing clustering algorithm, such as the inability to set the clustering range and the instability of the clustering result, and improves the reliability and scientificity of the station planning.

[0102] ​(3) The application further considers the influence of road conditions and position distribution in the grid, traffic demand and total travel cost of residents, sorts and filters all travel costs, and determines the specific position of the stop station in the grid. The problem that the existing research cannot determine the position of the stop station through quantitative analysis is solved.

[0103] (4) Based on the determination of the site location, the application considers factors such as travel demand of residents, waiting time of residents, driving distance of vehicles and the like, establishes a multi-objective planning model of the optimal route of public transportation travel, and solves through a heuristic algorithm.

[0104] The application can accurately determine the optimal site location of the stop station based on the grid optimization method according to the travel demand distribution situation of the multi-source data, the cost of the residents, the path of the residents to the site and the road conditions of the site setting, and realizes the planning of the public transportation route and the configuration of the vehicle by combining the multi-objective optimization model, so that the site location of the urban public transportation stop station is determined scientifically and accurately. BRIEF DESCRIPTION OF DRAWINGS

[0105] The application will be further described in detail in combination with the drawings and specific embodiments;

[0106] Figure 1 It is a flowchart of the method for the connection of the bus stop station and the route layout based on the grid optimization of the application;

[0107] Figure 2 It is a structural schematic diagram of the research area object of the application;

[0108] Figure 3 It is a grid schematic diagram of the travel demand distribution;

[0109] Figure 4 It is a schematic diagram of the evaluation grid traversal process;

[0110] Figure 5 It is a schematic diagram of the alternative unit grid;

[0111] Figure 6 It is a schematic diagram of the bus route direction;

[0112] Figure 7 It is a schematic diagram of the alternative grid;

[0113] Figure 8 It is a flowchart of the multi-objective solving method;

[0114] Figure 9 It is a schematic diagram of the multi-objective solving process state.

[0115] EMBODIMENT

[0116] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0117] With the increasing intelligence and digitalization of urban development, massive amounts of big data can be more easily obtained, allowing for a deeper understanding of the characteristics of urban residents' travel needs and providing crucial support for urban transportation planning and research. Big data is characterized by its large volume, wide range of sources, and rapid update speed, containing a vast amount of information. Multi-source data, such as public transportation travel data and population spatial distribution data, can accurately reflect the distribution of travel demand, providing support for developing more scientific and reasonable stop location plans and connecting bus route plans, thereby enhancing the attractiveness of urban public transportation.

[0118] like Figures 1 to 9 As shown in the figure, the present invention discloses a method for the layout of connecting bus stops and routes based on grid optimization, which includes the following steps:

[0119] Step 1: Obtain total public transportation demand data for the community based on travel big data; and divide the corresponding area into several unit grids with a set size, and calculate the travel demand distribution index for each unit grid in the area.

[0120] Furthermore, the travel big data in step 1 is multi-source data, which includes mobile phone signaling data, bus and rail card swiping data, shared bicycle travel order data, and high-precision map data.

[0121] Further, in step 1, the modal share of public transportation and shared bicycles is calculated; using a 200m*200m grid as the calculation unit, the travel demand distribution index of each grid unit in the region is obtained.

[0122] Furthermore, step 1 specifically includes the following steps:

[0123] Step 1-1: Extract mobile phone signaling data of all travelers within the service radius of the rail station;

[0124] Steps 1-2 involve preprocessing the data, removing invalid data, and filling in missing fields to clean the data.

[0125] Steps 1-3: Filter mobile phone signaling data from the morning rush hour on weekdays, based on historical data, to identify the total number of trips from residential communities to rail stations and the distribution of travel demand in the region.

[0126] Steps 1-4: Identify users' travel modes based on mobile phone signaling data, combined with total travel volume and travel demand distribution.

[0127] Furthermore, the specific methods for identifying the user's travel mode in steps 1-4 are as follows:

[0128] (1) For communities where there are no shared bicycles available, the average speed from the community to the rail station is statistically analyzed. Mobile phone signaling data with an average speed of less than 10 km / h and a distance of less than 2 km from the rail station are considered as walking trips.

[0129] (2) In cases where there are no regular bus stops or shared bicycle stops near the community, mobile phone signaling data with an average speed between 10-20 km / h and a distance of more than 2 km but less than 5 km from the rail station are identified as private electric vehicle travel; for communities with regular bus stops nearby, the travel demand for rail stations as the destination is statistically analyzed by combining regular bus card swiping data to obtain the total bus connection volume for the corresponding community.

[0130] (3) For communities with shared bicycles available, the total number of shared bicycle connections is obtained by combining shared bicycle order data with the travel demand for rail stations.

[0131] (4) Subtract the pedestrian traffic volume, private electric vehicle traffic volume, public transport connection traffic volume, and shared bicycle connection traffic volume from the total mobile phone signaling data volume, and identify the remaining traffic volume as travel by private car.

[0132] Step 2: Using the evaluation grid optimization method, calculate the total travel demand for each evaluation grid. and regional competitive advantage The evaluation grid is a rectangular grid composed of several unit grids; the length of the evaluation grid is determined by the station spacing, approximately the length of 5 unit grids.

[0133] Furthermore, step 2 specifically includes the following steps:

[0134] Step 2-1: Calculate the evaluation grid based on the travel demand distribution index of each unit grid. j Total demand for all transportation The calculation formula is:

[0135]

[0136] Among them, cell i This represents the i-th cell grid within the region. Let i represent the traffic demand of the i-th cell grid, where i = 1, ..., n, and n is the total number of cell grids; the side length of each cell grid is 150-200m.

[0137] Step 2-2: Calculate the competitive advantage of the evaluation grid area. The specific calculation formula is as follows:

[0138]

[0139] wherein, represents the average selection probability of choosing a shared bicycle in the evaluation grid, represents the average selection probability of choosing a bus in the evaluation grid; R is the maximum distance from the traffic hub site in the regional range, and the value of R is a fixed value; represents the distance between the grid centroid of the evaluation grid and the traffic hub site;

[0140] Further, the specific steps of step 2-2 are as follows:

[0141] Step 2-2-1, the average selection probability of choosing a shared bicycle in the evaluation grid is calculated The specific calculation formula is as follows:

[0142]

[0143] wherein, is the total demand for transportation of the evaluation grid j; is the total amount of bicycle trips to the traffic hub in the evaluation grid grid j , and the corresponding data can be obtained by counting the number of shared bicycle orders;

[0144] Step 2-2-2, the average selection probability of choosing a bus in the evaluation grid is calculated The specific calculation formula is as follows:

[0145]

[0146] wherein, is the total amount of regular bus trips to the traffic hub in the evaluation grid grid j , and the corresponding data can be obtained by counting the number of bus card swipes;

[0147] Step 2-2-3, the distance between the grid centroid of the evaluation grid and the traffic hub site is calculated The specific calculation formula is as follows:

[0148]

[0149] wherein, R c is the equatorial radius, x sta and y sta are the coordinates of the traffic hub site in radian system; and are the longitude and latitude of the grid centroid of the evaluation grid grid j in radian system;

[0150] Step 2-2-4, calculate the regional competitive advantage of the evaluation grid The specific calculation formula is as follows:

[0151]

[0152] wherein, represents the average selection probability of selecting a shared bicycle in the evaluation grid, represents the average selection probability of selecting a bus in the evaluation grid; R is the maximum distance from the transportation hub site in the region, and the value of R is a fixed value;

[0153] Step 3, the region covered by the station is searched and optimized by using the evaluation grid, and the evaluation grid grid j quantitative evaluation value centroid longitude and latitude of the evaluation grid and to construct a three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} get all grid quantitative evaluation value set evaluated in a certain range with a certain step

[0154] Further, the specific steps of step 3 are as follows:

[0155] Step 3-1, add the total traffic demand and regional competitive advantage of each evaluation grid to calculate the quantitative evaluation value of each evaluation grid grid j , that is:

[0156]

[0157] wherein, θ represents a scaling factor, representing the different tendencies of grid quantitative evaluation value to total traffic demand and regional competitive advantage, and the value of θ is obtained according to actual conditions;

[0158] Specifically, the entire region is traversed by continuously translating the grid, and the average distance between subway stations is taken as the unit step length L of translation, thereby obtaining the quantitative value of each region and performing descending order sorting to obtain a quantitative value evaluation array; The invention needs to cover all areas around the subway station as much as possible, so it is necessary to select a suitable grid length to surround all areas without causing the grid to be too large. Specifically, the distance between stations is generally 400-600m, which is obtained according to actual needs.

[0159] Step 3-2, obtain the centroid longitude and latitude of the evaluation grid and ​A three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} is constructed A set of all grid quantitative evaluation values evaluated in a certain range with a certain step size is obtained

[0160] Step 4, determine the optimal unit grid of site location of each evaluation grid based on travel cost as the optimal site location of the transfer station, and form a bus driving route scheme; specifically including the following steps:

[0161] Step 4-1, include the corresponding unit grid passed by the road into the set of candidate unit grids: the specific steps are as follows

[0162] Step 4-1-1, establish the evaluation grid grid j and the corresponding relationship with all the unit grids {cell1,…cell n} under it;

[0163] Step 4-1-2, select all road sets {road1,…road p} for setting the transfer station based on road grade and road connectivity factors;

[0164] Step 4-1-3, positionally associate {road1,…road p} with {cell1,…cell m}; when the corresponding unit grid passed by the road is included in the set of candidate unit grids, {cell1′…cell m ′} is finally obtained.

[0165] Specifically, all unit grids, such as 4*4=16 grids, but after associating with the road, only 5 grids are left, and the remaining 11 grids have no road, so the number of candidate unit grids is 5, that is, the set of {cell1′…cell m ′} in the following

[0166] Step 4-2, select the standby grid with the minimum total cost in the set of candidate unit grids as the optimal site location:

[0167] Step 4-2-1, calculate the total cost of each unit grid of {cell1,…cell n} to each candidate grid;

[0168] Step 4-2-2, select the unit grid with the minimum total cost in the candidate set as the optimal unit grid of site location of the current evaluation grid;

[0169] Specifically, for the unit grid with the minimum total cost (i.e. the optimal position station under the grid), the station sequence is u r(i = 1, …, w).

[0170] Step 5, a multi-objective optimization model is established to optimize the bus driving route to obtain the final bus driving route scheme, and the decision variables of the model optimization include the bus driving route (i.e. the sequence of stations) and the number of vehicles equipped on the line. The specific steps are as follows:

[0171] Step 5-1, determination of objective function: calculate the total cost of vehicle cost, the total cost of resident travel and the total demand of resident travel to establish the objective function; construct the constraint function of passenger service constraint, running time constraint, vehicle line constraint and vehicle capacity constraint;

[0172] Step 5-2, solution of optimal solution: based on the constructed constraint function, the NSGA-II algorithm is applied to solve the multi-objective function.

[0173] The specific principles of the present application are described in detail as follows:

[0174] As shown in Figure 2 The research object of the present application is centered on a rail station, a certain distance is taken as a radius to determine the service range of the rail station, and the rail station is connected with each travel demand point around by a feeder bus; the demand distribution in the service range is determined, and the location of the feeder bus station is determined considering the influence of various factors, and the driving route of the feeder bus is determined. The specific technical points are as follows:

[0175] (1) A method for obtaining community public transportation commuting travel demand based on travel big data

[0176] The existing researches usually estimate the regional public transportation travel demand through sampling artificial investigation and bus historical passenger flow data. The sample size of the investigation result of this kind of method is small, and the distribution of the regional travel demand cannot be completely indicated, and the distribution of the potential travel demand cannot be counted.

[0177] The present application obtains the total amount of regional public transportation commuting travel demand by combining mobile signaling data, bus and rail boarding and alighting card data, shared bicycle travel order data and high-precision map data and other multi-source travel big data.

[0178] ①Extract the mobile signaling data of all travelers within the service radius of the rail station;

[0179] ②Data preprocessing, eliminating invalid data, supplementing missing fields and other data cleaning;

[0180] ③Screening during the morning peak of weekdays, based on historical data to identify the mobile signaling data of the morning peak from the community to the rail station to obtain the total amount of travel and the travel demand distribution from the community to the rail station in the region;

[0181] IV. Identify the mode of transportation for the above mobile signaling data:

[0182] A. For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the average speed less than 10Km / h and the distance to the rail station less than 2Km are considered as walking trips.

[0183] B. For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the average speed less than 10Km / h and the distance to the rail station less than 2Km are considered as walking trips.

[0184] C. For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the average speed less than 10Km / h and the distance to the rail station less than 2Km are considered as walking trips.

[0185] C. The total amount of mobile signaling data is reduced by the amount of walking traffic, private electric vehicle traffic, bus transfer traffic, and shared bicycle transfer traffic, and the remaining traffic is considered as private car travel.

[0186] Based on the above steps, the share rate of bus and shared bicycle traffic is further calculated; taking a unit square 200m*200m as the calculation unit, the travel demand distribution index of each unit grid in the region is obtained.

[0187] (2) Station double-factor evaluation algorithm based on grid optimization:

[0188] 1) Double-factor evaluation algorithm based on traffic demand and location competitive advantage: The purpose of setting the transfer bus is to solve the travel demand problem from the residence to the nearest subway station or transportation hub. Therefore, the setting of the transfer station needs to consider the total traffic demand in the coverage range of the station, and also needs to avoid the situation that the distance to other stations with competitive relationship is too close. Therefore, in the process of station site selection, the traffic demand of the coverage range and the competitive advantage of the station need to be considered.

[0189] ① Evaluation of total traffic demand of grid

[0190] As shown in Figure 3 , the resident travel heat map in the research range can be obtained by step (1). As shown in Figure 3 , assuming that the traffic demand in each unit grid (cell) is known, cell i (i=1,…,n) represents the i-th unit grid in the research range, the length of the unit grid is generally 150-200m, and the traffic demand in the unit grid is At the same time, the coverage range l of the station (calculated according to the service range of the station, the station coverage range is 500-1500m) is considered, and the evaluation grid grid is drawn with the side length l, then for all traffic demands of the jth evaluation grid The calculation formula is:

[0191]

[0192] Wherein, cell i represents the ith unit grid in the area range, is the traffic demand of the ith unit grid, i=1,…,n, n is the total number of unit grids; the side length of each unit grid is 150-200m.

[0193] ②Evaluation of grid location competitive advantage: the location competitive advantage is strong, there is no regular bus to the subway or transportation hub station around, there is no shared bicycle stop, and the farther the distance from the hub, the more necessary it is to build a connecting station, and the necessity of construction is proportional to the location competitive advantage. The regional competitive advantage is defined as As follows:

[0194]

[0195] Wherein, represents the average selection probability of choosing a shared bicycle in the evaluation grid, represents the average selection probability of choosing a bus in the evaluation grid, and the acquisition method is:

[0196]

[0197] Wherein, and are the total amount of bicycles and regular buses to the transportation hub in the grid j grid. This data can be obtained by counting the number of shared bicycle orders and bus card swiping.

[0198] R is the maximum distance from the transportation hub station in the research range, which is a fixed value.

[0199] is the distance between the grid center of grid j and the transportation hub station, which is calculated by the following formula:

[0200]

[0201] Wherein, x sta and y sta are the longitude and latitude of the transportation hub station.

[0202] 2) Grid search optimization method considering site coverage:

[0203] ① Calculation of grid quantitative evaluation value: add grid demand and location competitive advantage to calculate the quantitative evaluation value of each grid That is:

[0204]

[0205] In the above formula, since the travel demand is an actual value, and the grid location competitive advantage is a number in [0, 3], a scaling factor θ is set to represent the different tendencies of grid quantitative evaluation value to demand and location competitive advantage. This value is obtained according to the actual situation.

[0206] In the studied area, the grid is continuously translated to traverse. Considering the distance between sites (usually 300-500m), the distance between sites is taken as the unit step length L of translation, and then the quantitative value of each area is obtained, which is sorted in descending order to obtain the quantitative value evaluation array. At the same time, the centroid latitude and longitude of the rectangle are calculated, and then a three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} is constructed Then get all grid quantitative evaluation value set in a certain range with a certain step

[0207] ② Spatial autocorrelation judgment: if two grids are close in distance and similar in quantitative evaluation value, that is, the spatial autocorrelation is high, and the quantitative evaluation value is less than a certain threshold, then the two grids can be merged, which is equivalent to merging the points with less demand to reduce the cost of site construction. Considering the correlation between grid quantitative evaluation values, Global Moran's I spatial autocorrelation measure index is used to describe the similarity of attribute values of spatial adjacent regions. The calculation method is as follows:

[0208]

[0209] Where y i is the attribute value of element i, which is the quantitative evaluation value in the grid in this paper. The length of the set of all grid quantitative evaluation values; w ij is the spatial weight between elements i and j. In spatial computing analysis, the spatial weight can be represented as an adjacency matrix and a distance matrix. Considering that the scope studied in this invention contains multiple grids, the spatial weight is defined as a distance matrix. I>0, indicating that all adjacent grid attribute values have positive correlation in space, so the grid can be merged. For grids with I≤0 showing negative correlation, a site is set separately.

[0210] 3) Optimal site selection algorithm of transfer station based on travel cost: For each evaluation grid grid, the traffic demand heat of the unit grid cell contained is not evenly distributed, and the position of the transfer station needs to be further determined in the square with side length l, that is, the optimal unit grid in the optimal evaluation grid is searched.

[0211] 3-1, determination of alternative unit grid: as shown in Figure 5 , let grid j be one of the optimal evaluation grids, and all unit grids belonging to it are {cell1,…cell m}. Since the station position must be set on the road, and the road cannot be a dead-end road, based on the road network map data, all possible road sets that can set the transfer station are determined and associated with the grid, so that the subordinate alternative unit grid set {cell1′,…cell′} that meets the station setting condition is screened out from all subordinate unit grids of grid j . The specific steps are as follows:

[0212] 3-1-1, establish the corresponding relationship between grid j and all subordinate unit grids {cell1,…cell n};

[0213] 3-1-2, considering factors such as road grade and road connectivity, select all road sets {road1,…road p} that can set the transfer station;

[0214] 3-1-3, position association between {road1,…road p} and {cell1,…cell n}, if the road passes through the unit grid, the unit grid is included in the alternative unit grid set, and finally {cell1′…cell m ′} is obtained.

[0215] 3-2, determination of optimal site selection of station: after determining the alternative unit grid set {cell1′…cell m ′}, it is necessary to further determine in which cell in the set the specific station site selection is. In fact, the determination of the station is related to the travel cost of the surrounding residents, and the total travel cost in grid is the minimum as the consideration factor of optimal site selection of the station.

[0216] Specifically, as shown in Figure 6 , for one of the alternative unit grids cell n ′, calculate all cells with traffic demand around the alternative unit grid cell nThe actual travel distance is the distance traveled along the actual road, which can be calculated through the relationship between the road and the cell grid.

[0217] exist Figure 6 In the middle, from cell i to cell n The actual distance between them is not a straight-line distance, but a grid distance along the road (e.g., ...). Figure 6 The distance between the two shown is 8). Based on the above method, the distance l between all grids and the candidate grids can be calculated for {cell1′… ... m For each candidate cell grid in the set {cell1, ..., cell}, calculate {cell1, ..., cell2}. n The total cost from each cell grid to the alternative grid is:

[0218]

[0219] For the candidate set, the cell grid with the lowest total cost is the optimal cell grid for site selection. Therefore, the station sequence for the cell grid with the lowest total cost (i.e., the optimal location station within that cell grid) is u. r (r=1,…,w), that is:

[0220]

[0221] 4) Algorithm for Optimizing the Configuration of Shuttle Bus Routes: Given multiple known stops, the algorithm optimizes the configuration of shuttle bus routes by considering the shortest vehicle travel distance, the shortest waiting time for residents, and the maximum passenger capacity of the vehicles. A multi-objective optimization model is established to determine the bus routes. The decision variables determined by the model include the vehicle travel route (i.e., the sequence of stops) and the number of vehicles allocated to each route.

[0222] 4-1) Determining the objective function:

[0223] 4-1-1, Total vehicle operating cost: Let the peak-hour vehicle departure interval be t. r Therefore, the total cost for each vehicle's journey should be the sum of the distances traveled by each vehicle on that route, i.e.:

[0224]

[0225] Where K represents the total time of the morning rush hour commuting service. This represents the actual travel distance from the r-th station to the (r+1)-th station.

[0226] 4-1-2, Total Cost of Residents' Travel Time: The travel time for passengers taking community buses connecting to rail transit stations includes three parts: walking time to the bus stop, waiting time, and time spent on the bus. That is:

[0227] C2 = T g + T w + T s (12)

[0228] where T h is the total walking time of residents to bus stops, for stop u i , the total walking travel time is equal to Then the total walking travel time of residents T h for all stops is:

[0229]

[0230] T w is the waiting time of residents, considering that the location of stops and the number of stops have been determined, the waiting time is related to the frequency of departure, assuming that the vehicle arrival law is uniformly distributed, then the waiting time of the stop is equal to t r / 2, then the total waiting time of residents T w for all stops is:

[0231]

[0232] The in-vehicle time is determined by the passenger's origin-destination sequence, based on the travel data, the passenger's commuting origin can be obtained, that is, the boarding and alighting quantity of each station sequence. Then for u r to u r+1 station sequence, the total in-vehicle time of passengers is:

[0233]

[0234] where, is the travel time between two stations.

[0235] 4-1-3, the maximum number of passenger demand satisfaction, the specific expression is as follows:

[0236]

[0237] From the above, the objective function is:

[0238] min C = C1 + C2 - C3 (18).

[0239] 4-2) Determination of constraint function:

[0240] 4-2-1, passenger service constraint: each passenger can only be served by one vehicle

[0241]

[0242] where x b is a 0-1 variable, if the passenger is served, then xb 1, otherwise 0.

[0243] 4-2-2, running time constraint, i.e. the single running time of each community bus must be limited to the maximum running time T max , the corresponding expression is:

[0244]

[0245] 4-2-3, vehicle route constraint, i.e. the length of the route of each bus single running needs to be controlled within a certain range, the corresponding expression is:

[0246]

[0247] 4-2-4, vehicle capacity constraint: i.e. the number of passengers carried by each community bus is not allowed to exceed the capacity of the vehicle.

[0248]

[0249] represent the total number of all stations, represent the total number of people on the bus, the number of passengers boarding the bus is less than the maximum limit of the vehicle carrying capacity.

[0250] 4-3) Multi-objective solution method: the method of multi-objective solution is relatively mature, and the method adopts NSGA-II multi-objective solution algorithm. The algorithm flow is as follows:

[0251] 4-3-1, randomly generate initial population P0 and perform non-inferior sorting, execute binary tournament selection, crossover and mutation on the population to obtain new population Q0, and let t = 0.

[0252] 4-3-2, form a new population R t = P t ∪Q t Again, perform non-inferior sorting to obtain F1, F2, …;

[0253] 4-3-3, sort all F by crowding degree, select the optimal N individuals to form the population P t+1 ;

[0254] 4-3-4, execute step 4-3-1 on population P t+1 to obtain Q t+1 ;

[0255] 4-3-5, if the termination condition is reached, the algorithm ends; otherwise, go to step 4-3-2;

[0256] The main process is as follows: Figure 9 : Through multi-objective algorithm solving, the optimal solution set satisfying minC can be obtained, and the vehicle driving route (i.e. the sequence of stations) and the number of vehicles equipped with the route are obtained.

[0257] The present application proposes to combine multi-source big data to estimate the demand for feeder bus travel, and can obtain relatively complete travel distribution data, improving the scientificity and reliability of station layout. In the aspect of station layout, the existing research usually selects clustering algorithm to determine the station location. Through the clustering algorithm, only the same type of travel demand is usually classified into a category, and the limitation of the service radius of the bus station is not eliminated, which may lead to the problem that part of the station needs to bear too much demand, or due to the influence of remote areas, the station needs to be specially laid out to meet the needs of a small number of travel demands, resulting in resource waste. Therefore, the present application first proposes a station double-factor evaluation algorithm based on grid optimization. Among them, the double-factor evaluation algorithm based on traffic demand and location competitive advantage is used to evaluate the total travel demand and location competitive advantage of the regional candidate station, and the grid search optimization method considering the coverage range of the station is used to calculate the grid quantitative evaluation value and the judgment of spatial autocorrelation, which is used to evaluate the station. On this basis, the optimal site determination algorithm of the feeder bus station based on travel cost is proposed, so as to determine the candidate grid and the final site of the station, solve the problem caused by the clustering algorithm, and improve the accuracy and reliability of the station layout. On the basis of station site selection, the configuration optimization algorithm of the feeder bus line is determined, and the objective function and solving algorithm are proposed.

[0258] The present application has the following technical advantages by adopting the above technical scheme:

[0259] (1) The present application combines mobile signaling data, bus track boarding and alighting card data, shared bicycle travel order data and high-precision map data and other multi-source travel big data to obtain the total amount of regional public transportation commuting travel demand, improving the accuracy and completeness of travel distribution.

[0260] (2) The present application considers the coverage range and spacing of the public transportation station, collects the total amount of travel demand in the statistical grid and the quantitative evaluation value of the location competitive advantage, evaluates the site selection position of the public transportation station based on the double-factor evaluation result, and collects the grid with high spatial autocorrelation and low demand to determine the optimal site selection position of the public transportation station. The problems of unable to set the clustering range and unstable clustering result in the existing clustering algorithm are solved, and the reliability and scientificity of station planning are improved.

[0261] (3) The present application further considers the influence of road conditions and location distribution, traffic demand and total travel cost of residents in the grid, sorts and filters all travel costs to determine the specific position of the stop station in the grid. The problem that the existing research cannot determine the stop station position through quantitative analysis is solved.

[0262] (4) The application establishes a multi-objective planning model of the optimal route of public transport travel on the basis of site location determination, considering factors such as resident travel demand, resident waiting time, vehicle driving distance, etc., and is solved by a heuristic algorithm.

[0263] The application is based on the statistical travel demand distribution of multi-source data, and considers the cost of resident travel, the path of resident transfer to the site and the road condition of site setting, and can accurately determine the optimal site location of the site based on the grid optimization method, and realizes the planning of public transport route and the configuration of vehicle, and scientifically and accurately determines the site location of urban public transport stop.

[0264] Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

Claims

1. A method for setting a transfer bus station and a line based on grid optimization, characterized in that: It comprises the following steps: Step 1, obtaining the total amount of public transport traffic demand data of the cell based on travel big data; and setting the size to divide the corresponding area into a plurality of unit grids, respectively calculating the travel demand distribution index of each unit grid in the area; travel big data includes mobile signaling data, bus and rail boarding and alighting card data, shared bicycle travel order data and high-precision map data; Step 2, using the evaluation grid optimization method, the total demand for travel of each evaluation grid is calculated respectively and regional competitive advantage The evaluation grid is a rectangular grid composed of a plurality of unit grids; comprising the following steps: Step 2-1, based on the travel demand distribution index of each unit grid, calculate the evaluation grid grid j of all the total demand of traffic travel The calculation formula is: wherein cell i is the i-th cell grid in the region, is the traffic demand of the i-th cell grid, i = 1,..., n, n is the total number of cell grids corresponding to the evaluation grid; the side length of each cell grid is 150-200 m; Step 2-2, calculate the evaluation grid area competitive advantage The specific calculation formula is as follows: wherein, is the average selection probability of choosing a shared bicycle in the grid, is the average selection probability of choosing a bus in the grid; R is the maximum distance from the transportation hub site in the region range, and the value of R is a fixed value; is the distance between the grid centroid of the grid and the transportation hub site. Step 3, the station coverage area uses the evaluation grid to search for optimization, respectively calculates the quantitative evaluation value of the evaluation grid grid j The centroid longitude and latitude of the evaluation grid And To construct a three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} Get all grid quantitative evaluation value sets evaluated within a certain range with a certain step The specific steps are as follows:​ Step 3-1: Calculate the total travel demand for each evaluation grid. and regional competitive advantage The values ​​are added together to obtain each evaluation grid. j Quantitative evaluation value Right now: Wherein, θ is a scaling factor, representing the different inclination of grid quantization evaluation value to total traffic demand and location competitive advantage, θ value is obtained according to actual situation; Step 3-2, get the centroid longitude and latitude of the evaluation grid and Construct a three-dimensional array with {quantitative evaluation value, grid longitude, grid latitude} Get the set of all grid quantitative evaluation values evaluated within a certain range at a certain step size Step 4, determining the optimal unit grid of site location of each evaluation grid based on travel cost as the optimal site location of the transfer station, forming a bus driving route scheme; Step 5, establishing a multi-objective optimization model to optimize the bus driving route to obtain the final bus driving route scheme, the decision variables of model optimization include bus driving route and vehicle number of line arrangement; the specific steps are as follows: Step 5-1, determination of objective function: respectively calculating vehicle cost total fee, total cost of resident travel and total demand of resident travel to establish multi-objective function; constructing constraint functions of passenger service constraint, running time constraint, vehicle line constraint and vehicle capacity constraint; Step 5-2, solution of optimal solution: based on the constructed constraint function, NSGA-II algorithm is applied to solve the multi-objective function.

2. The method according to claim 1, wherein: Step 1 specifically comprises the following steps: Step 1-1, extracting mobile signaling data of all travelers within the service radius of the rail station; Step 1-2, pre-processing the data, eliminating invalid data, and supplementing missing fields for data cleaning; Step 1-3, selecting the morning peak period on weekdays, identifying the mobile signaling data based on historical data, taking the cell as the starting point and the rail station as the terminal, to obtain the total amount of travel and travel demand distribution from the cell to the rail station in the area; Step 1-4, identifying the travel mode of users based on mobile signaling data combined with travel total amount and travel demand distribution.

3. The method according to claim 2, wherein: The specific method of step 1-4 for identifying the travel mode of users is: (1) For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the mobile signaling data with average speed less than 10Km / h and distance less than 2Km from the rail station is identified as walking travel; (2) For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the mobile signaling data with average speed less than 10Km / h and distance less than 2Km from the rail station is identified as walking travel; (3) For the cell near the shared bicycle supply, the average speed from the cell to the rail station is calculated, and the mobile signaling data with average speed less than 10Km / h and distance less than 2Km from the rail station is identified as walking travel; (4) The total amount of mobile signaling data is reduced by the amount of walking traffic, private electric vehicle traffic, bus transfer traffic and shared bicycle transfer traffic, and the remaining traffic is identified as private car travel.

4. The method according to claim 1, wherein: The specific steps of step 2-2 are as follows: Step 2-2-1, calculate the average selection probability of selecting a shared bicycle in the evaluation grid The specific calculation formula is as follows: wherein, is the total demand of all trips in the evaluation grid grid j ; is the total amount of trips from bicycles to the transport hub in the evaluation grid grid j , the corresponding data can be obtained by counting the number of shared bicycle orders; Step 2-2-2, calculate the average selection probability of the selected bus in the evaluation grid The specific calculation formula is as follows: in, To evaluate the grid j The total number of trips from regular buses to transportation hubs can be obtained from the statistics of bus card swipes. Step 2-2-3, calculate the distance between the grid centroid of the evaluation grid and the transportation hub site The specific calculation formula is as follows: wherein R c is the equatorial radius, x sta and y sta are the coordinates in the radian system of the transport hub site; and are the coordinates in the radian system of the center of the grid grid j of evaluation; The calculation formula of converting the angle system to the radian system is as follows: rad = z·π / 180°, wherein z is the latitude and longitude in the angle system, and rad is the latitude and longitude in the radian system after conversion. Step 4 specifically includes the following steps: Step 2-2-4, calculate the evaluation grid area competitive advantage The specific calculation formula is as follows: wherein, represents the average selection probability of choosing a shared bicycle in the evaluation grid, represents the average selection probability of choosing a bus in the evaluation grid; R is the maximum distance from the transport hub site within the area range, and the value of R is a fixed value.

5. The method according to claim 1, wherein: Step 4-1, the corresponding unit grid through which the road passes is included in the set of candidate unit grids: the specific steps are as follows Step 4-2, the standby grid with the minimum total cost in the set of candidate unit grids is selected as the optimal site location of the station: Step 4-1-1, Establish evaluation grid grid j corresponding relationship with all the subordinate unit grids {cell1,... cell n} Step 4-1-2, based on road level, road connectivity factor, select all road set {road1, …road p} of candidate setting connection station. Step 4-1-3, position association between {road1,...,road p} and {cell1,...,cell n}; the corresponding cell grid crossed by the road is included into the candidate cell grid set, and finally {cell1',...,cell m '} is obtained. Step 4-2-2, the unit grid with the minimum total cost in the candidate set is selected as the optimal unit grid for site selection of the current evaluation grid. Step 4-2-1, compute the total cost of each cell grid of {cell1,... cell n} to each candidate grid; The specific steps of establishing the multi-objective function are as follows:

6. The method according to claim 1, wherein: Step 5-1-2, total cost of resident travel time: the travel time of passengers taking the community bus of the feeder rail transit station includes three parts: walking to the bus station, waiting time and in-vehicle time, that is: Step 5-1-1, Total cost of vehicle travel: Let the operating headway of the vehicle during peak hours be t r , then the total cost of vehicle travel should be the sum of the distance traveled by each vehicle on the line, i.e., where K is the total time of the morning rush hour service, is the rth station u r is the actual travel distance to the r+1th station u r+1 is the actual travel distance to the r+1th station u Step 5-1-3, maximum number of passenger demand satisfaction: the specific expression is as follows: C2 = T h + T w + T s (12) where T h is the total time for a resident to walk to a bus stop, T w is the wait time for a resident, and T s is the in-vehicle time. For a site u r , the total walking travel time is equal to Then the total walking travel time T h for all sites is: Assuming that the vehicle arrival law is uniform, the waiting time at the station is equal to t r / 2, the total waiting time T of all residents for all stations is: w ​ Based on the trip data, the passenger's commuting origin is obtained, i.e. the boarding and alighting quantity of each station sequence, then for u r to u r+1 station sequence, the total time of passengers in the car is: wherein, is the travel time between two stations; Step 5-1-4, construct a multi-objective function, and the specific expression is: minC = C1 + C2 - C3 (18); The calculation steps of the constraint function are as follows: Step 5-1-21, passenger service constraint: each passenger can only be served by one vehicle Step 5-1-23, vehicle route constraint, that is, the length of the line of each bus in single operation is controlled within a certain range, and the corresponding expression is: where x b is a 0-1 variable, x b is 1 when the passenger is served, and 0 otherwise. Step 5 - 1-22, running time constraint, i.e. the single running time of each community bus is limited to the maximum running time T max In the inner loop, the corresponding expression is: Step 5-1-24, vehicle capacity constraint: that is, the number of passengers carried by each community bus is not allowed to exceed the capacity C of the vehicle; wherein L min represents the lowest line length for a single run; L max represents the highest line length for a single run; The multi-objective solving in step 5-2 specifically includes the following steps: wherein, denotes the rth station u r to the r+1th station u r+1 the number of passengers on the vehicle at that time.

7. The method according to claim 1, wherein: Step 5-2-1, randomly generate an initial population P0 and perform non-inferior sorting, execute binary tournament selection, crossover and mutation on the population to obtain a new population Q0, and let t = 0; Step 5-2-5, if the termination condition is reached, the algorithm ends; otherwise, go to step 5-2-2. Step 5 - 2 - 2, form new group R t = P t ∪ Q t Again, non-dominant sorting is performed to obtain F1, F2,...; Step 5-2-3, crowding degree sorting is performed on all F, and the optimal N individuals are selected from F to form a population P t+1 ; wherein F represents a set composed of all feasible solution populations R in the NSGAII algorithm t . Step 5-2-4, for population P t+1 Perform step 5-2-1 operation, get Q t+1 ; ​

Images