Bus station location selection and route design method and device based on passenger demand

Abstract

The present invention belongs to the field of transportation, and specifically relates to a method and device for bus station site selection and route design based on passenger demand. The method steps are as follows: S1: Establish a directed graph representing traffic conditions based on the geographical characteristics and road distribution of the region. S2: Determine the alternative routes, starting and ending arcs and virtual arcs of the bus route, and establish an operation map. S3: Determine the alternative boarding and alighting points for passengers, passenger boarding and alighting arcs, operation arcs, transfer arcs and virtual arcs, and establish a demand map. S4: Take the arcs in the operation map and the demand map as decision variables; the minimum comprehensive cost of travel cost and route operation cost is the optimization goal; set the constraint condition of the rationality constraint of the route, and establish a line network optimization model. S5: Use a variable neighborhood search algorithm to iteratively optimize the line network optimization model to obtain the optimized site selection and line layout. The present invention solves the problem that the existing bus line site layout is unreasonable and cannot meet the travel needs of passengers.

Description

Technical Field

[0001] The invention belongs to the field of transportation, and in particular relates to a method and device for bus station site selection and line design based on passenger demand. Background Art

[0002] Considering that public transportation is greener and more environmentally friendly than other modes of transportation and can also reduce residents' travel costs, how to improve the convenience of residents in the region to use public transportation is one of the main issues of sustainable transportation.

[0003] The existing public transportation line design and bus station site selection process rely on old standards, and the station settings in the old standards are usually aimed at achieving the maximum resident coverage rate. This results in that in the actual bus operation process, although passengers are within the coverage range calculated by straight-line distance in the national standard, due to the complex road conditions in local areas such as towns and urban villages, the distance passengers walk to the bus stop is far longer than the coverage distance specified in the national standard.

[0004] In addition, there are several defects in the existing public transportation route design: 1. Little attention is paid to the traffic convenience in remote areas; 2. Modeling and optimization are not carried out according to the actual road network, and the optimization results produced do not conform to the actual road conditions. 3. The consideration is not comprehensive enough, and multiple factors such as passenger travel needs, passenger transfer plans, actual road conditions, passenger travel routes, bus stop locations, bus routes, and the number of bus routes are not considered at the same time. In summary, the existing public transportation routes and stations are unreasonable. On the one hand, it will lead to the inability to fully meet the travel needs of passengers. On the other hand, it will also cause the bus company's capacity configuration to be mismatched with the travel needs of passengers, reducing the operating income of the bus company. Summary of the invention

[0005] In order to solve the problem that unreasonable public transportation routes and station settings bring inconvenience to users and lead to reduced operating income of bus companies, the present invention provides a bus station site selection and route design method and device based on passenger demand.

[0006] The technical solution provided by the present invention is:

[0007] A bus station location and route design method based on passenger demand is used to redesign all existing bus routes in a specified area, thereby achieving optimal cost while meeting the travel needs of most passengers. The method includes the following steps:

[0008] S1: Establish a directed graph representing the current regional traffic conditions based on the geographical characteristics and road distribution of the specified area G : .in, Represents the set of all candidate sites that meet the site setting conditions;A It represents the set of travel paths between any two candidate sites that can be reached in the same way.

[0009] S2: In a directed graph G Determine any bus route l The line alternative stations, line starting arcs, line ending arcs and line virtual arcs are then used to establish the corresponding operation map G l .

[0010] S3: In a directed graph G Determine any passenger p The alternative boarding points, alternative alighting points, passenger boarding arcs, passenger travel arcs, passenger transfer arcs and passenger virtual arcs in the travel demand are used to establish the corresponding demand map G p .

[0011] S4: All arcs in the operation maps of all bus routes are used as decision variables for bus route selection, and all arcs in all demand maps are used as decision variables for passenger route selection; the unit cost of each type of arc contained in the operation map and the demand map is preset, and then the minimum comprehensive cost including the travel costs of all passengers and the operation costs of all bus routes is taken as the optimization goal; a line network optimization model is established based on the constraint conditions of line setting rationality constraints.

[0012] S5: Adopting a variable neighborhood search algorithm to iteratively optimize the line network optimization model, and adding a transfer selection operation between the perturbation operation and the repair operation of the variable neighborhood search algorithm to accelerate iteration; finally, the lines in each operation map that do not contain line virtual arcs in the optimization results are used as the final bus lines to be opened, and the alternative stations passed by the bus lines contained in each operation map are used as the actual set stations.

[0013] As a further improvement of the present invention, in step S2, any bus route l Running map G l The construction method is as follows:

[0014] (1) Remove the candidate sites corresponding to the starting and ending points of the original route in the directed graph G, and use the remaining sites as the route candidate sites. (2) Take each path in G starting from the original route starting point as the route starting point arc ; Each pass path that reaches the end of the original line is used as the line end arc ; The remaining travel path is used as the line operation arc (3) Define the jump path from the line starting point directly to the line end point without passing through each line alternative station as the line virtual arc .

[0015] As a further improvement of the present invention, in step S3, any passenger p Demand Map G p The construction method is as follows:

[0016] (1) The corresponding passengers in G p Each candidate station within a specified radius around the boarding and alighting locations of the historical ride record is used as the current user's candidate boarding point and alternative drop-off points (2) Change G from Each departure path is used as the passenger boarding arc ;arrive Each path of the passenger drop-off arc ; The remaining travel path is the passenger running arc, still recorded as (3)Definition of passenger p The jump path from the line alternative station i of one bus line to the line alternative station j of another bus line is the passenger transfer arc (4) Define the jump path from the alternative boarding point to the alternative alighting point without passing through the alternative stations of each line as the passenger virtual arc .

[0017] As a further improvement of the present invention, in step S4, the expression of the optimization target is as follows:

[0018]

[0019]

[0020]

[0021] In the above formula, c p ( i , j ) means that the passenger is in any The arc ( i , j )’s unit operating cost; represents the passenger path selection decision variable; P represents the set of all bus users; L represents the set of all bus routes; A l Indicates running map G l The set of all arcs contained in ; A p Demand map G p The set of all arcs contained in ; cl ( i , j ) means any bus route belonging to A l The arc ( i , j )’s unit operating cost; represents the bus route selection decision variable.

[0022] As a further improvement of the present invention, in the line network optimization model, the line setting rationality constraints include bus line flow balance constraints, passenger path flow balance constraints, coupling constraints, line length constraints, transfer number constraints, path integrity constraints and sub-loop elimination constraints.

[0023] As a further improvement of the present invention, the bus line flow balance constraint means that when the bus line is opened, each operation map must contain a line starting point arc and a line end point arc, and any line alternative station in the operation map must have a line operation arc that arrives at the station and departs from the station at the same time.

[0024] The passenger path flow balance constraint means that if the passenger takes public transportation, each demand map must contain a passenger boarding arc and a passenger getting off arc, and any alternative station on the demand map must have a route operation arc that arrives at the station and departs from the station.

[0025] The coupling constraint means that each passenger operation arc in the demand map must have a corresponding line operation arc in the operation map; and the number of passenger operation arcs allocated to each line operation arc is not greater than the preset line capacity C.

[0026] The route length constraint means that the total distance from the route start point to the route end point in any running map is not greater than the preset length limit. L max .

[0027] The transfer number constraint means that the number of transfer stations included in each demand map does not exceed the preset transfer number upper limit M.

[0028] The line integrity constraint means that the previous pass path and the next pass path of the passenger transfer arc contained in any demand map are not allowed to be a passenger getting off arc and a passenger getting on arc at the same time.

[0029] The sub-loop elimination constraint means that no loop is allowed between the remaining stations except the passenger starting point and the passenger end point in each demand route.

[0030] As a further improvement of the present invention, in step S5, the iterative optimization process performed by the variable neighborhood search algorithm is as follows:

[0031] S51: Encode the feasible solution consisting of the operation map of each bus line and initialize the objective function.

[0032] S52: On the premise of satisfying the constraint conditions, the sites included in the demand maps of all users are inserted into the feasible solution in turn to obtain an initial solution.

[0033] S53: using a preset perturbation operator to perturb the initial solution obtained in the previous step.

[0034] S54: Using a preset transfer selection operator, a specified number of stations are selected from each operation map in the previous step as transfer stations.

[0035] S55: Re-obtain a new intermediate solution through a neighborhood local search, update the value of the objective function, and make the following decisions:

[0036] If the objective function value of the solution of the current round is greater than the objective function value of the solution of the previous round, the solution of this round will be used as the initial solution of the next round; otherwise, the solution of the previous round will be used as the initial solution of the next round;

[0037] S56: Repeat steps S53-S55 until the maximum number of iterations is reached; retain the optimal solution that can minimize the objective function value;

[0038] S57: Decode the optimal solution to obtain several optimized bus routes and a station sequence in each bus route.

[0039] As a further improvement of the present invention, in step S53, the disturbance operator includes: a transfer point removal operator, an alighting and alighting cluster removal operator, and a history-based removal operator.

[0040] In step S54, the transfer selection operator includes: a random selection operator, a nearest neighbor operator, a centralized transfer operator and a historical performance operator.

[0041] In step S55, the operators used in the neighborhood local search include: an optimal insertion operator that simultaneously considers transfers, an optimal insertion operator that gives priority to transfers, and an insertion operator with the largest gap.

[0042] As a further improvement of the present invention, in step S54, a scoring mechanism is introduced in each round of iteration, and then after each round of iteration, the performance of each operator used in the transfer point selection operation is evaluated, and then the selection probability of each transfer selection operator in the next round is adjusted according to the performance evaluation value.

[0043] The present invention also includes a bus station site selection and route design device based on passenger demand, which includes a memory, a processor, and a computer program stored in the memory and running in the processor. When the processor executes the computer program, the bus station site selection and route design method based on passenger demand as described above is adopted to achieve redesign of all existing bus routes in a designated area, thereby achieving optimal cost while meeting the travel needs of most passengers.

[0044] The technical solution provided by the present invention has the following beneficial effects:

[0045] The present invention abstracts the layout of bus routes in a specified area and the travel needs of passengers into graph structure data, and then defines route virtual arcs, passenger transfer arcs, and passenger virtual arcs to more accurately model the optimization scenario, thereby obtaining an operation map corresponding to each bus route and a demand map corresponding to each customer's travel record.

[0046] Based on the operation map and the demand map, the present invention creates a network optimization model that can integrate and jointly optimize the passenger's riding route and the bus operation route; in constructing the network and the model, multiple factors such as the actual road conditions in the region, passenger travel needs, passenger transfer plans, passenger travel routes, bus stop locations, and the number of bus routes are considered at the same time, making it more realistic.

[0047] The present invention adopts a variable neighborhood search algorithm to solve the constructed wire network model. The algorithm adopts the idea of ​​systematically changing the neighborhood of the current solution during the search process. At the same time, the present invention adds a transfer selection operation to the perturbation operation of the traditional algorithm, so that the algorithm can further reduce the local search space. In particular, unlike the traditional variable neighborhood search algorithm, the present invention also designs a scoring mechanism in the algorithm process, and then uses the adaptive idea to flexibly select a transfer selection operator with better effect before each local search, so as to reduce the local search space as much as possible, which has an accelerating effect on the algorithm solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flowchart of the steps of the bus stop site selection and route design method based on passenger demand provided in Example 1 of the present invention.

[0049] Figure 2 A typical example of a demand map constructed with passenger transfer arcs.

[0050] Figure 3 A typical example of a run map where eliminating constraints for missing subcircuits results in the optimization containing incorrect subcircuits.

[0051] Figure 4 This is a flow chart of the variable neighborhood search algorithm used in Example 1 of the present invention.

[0052] Figure 5 This is a schematic diagram of a feasible solution before filling in passenger travel demands in Example 1 of the present invention.

[0053] Figure 6 This is a schematic diagram of the initialization solution after filling in the passenger travel demand in Example 1 of the present invention.

[0054] Figure 7 This is a schematic diagram of inserting passenger travel demands without transfer in Embodiment 1 of the present invention.

[0055] Figure 8 This is a schematic diagram of inserting passenger travel demands using an insertion method with transfer in Example 1 of the present invention.

[0056] Fig. 9 This is a schematic diagram of an application scenario of the optimal insertion operator that gives priority to transfers designed in Example 1 of the present invention.

[0057] Fig.10 This is a flow chart of a variable neighborhood search algorithm including a scoring mechanism used in Example 1 of the present invention. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0059] Example 1

[0060] This embodiment provides a method for bus stop site selection and route design based on passenger demand. The method is used to reselect the site of bus stops and redesign bus routes according to the actual road conditions in the area and the spatial distribution of residents' travel needs. This can shorten the travel time of residents, improve their travel convenience, and save the travel costs of residents taking public transportation. On the other hand, it can reduce the operating costs of bus companies and improve their operating efficiency and income.

[0061] Specifically, the solution provided in this embodiment abstracts the optimization design problem of the bus network in the region into a PDPT (pickup and delivery problem with transfers) problem, and establishes a corresponding linear programming mathematical model, namely, a network optimization model. Then, the VNS variable neighborhood search algorithm is used to solve the discrete optimization problem, and the optimal solution of the network optimization model including multiple new lines and new stations is obtained. In particular, in order to make the obtained optimization solution more practical, the present invention also constructs a transfer mechanism in the established network optimization model, and introduces the customer demand on the resident side and the operating cost on the enterprise side into the network optimization model for comprehensive consideration; this makes the optimization result of the solution of this embodiment more practical.

[0062] In detail, Figure 1 As shown, the bus station site selection and route design method based on passenger demand provided in this embodiment includes the following steps:

[0063] S1: Establish a directed graph representing the current regional traffic conditions based on the geographical features and road distribution of the specified area.

[0064] Taking the township area as an example, in this embodiment, it is necessary to first obtain the road network distribution of the roads that have been opened to traffic in the township area and the spatial distribution of the residential areas in the township area in the geographic image, and then combine the relevant standards (such as "Public Gas and Tram Network Evaluation Indicators" JT / T1457-2023) to mark the spatial locations of all candidate sites that meet the station setting conditions in the geographic image. Next, with each candidate site as a node and the roads that can be connected between sites as edges, the full-factor information supporting the operation of the regional public transportation network is abstracted into a directed graph. G : .

[0065] in, Represents the set of all candidate sites that meet the site setting conditions; A It represents the set of travel paths between any two candidate sites that can be reached in the same way.

[0066] It should be noted that, in this embodiment, the directed graph G The candidate stations in include stations that already exist in the existing bus routes in the area, as well as new stations that do not exist in the existing bus routes but meet the station setting conditions according to relevant standards. G The travel path between any two candidate sites is bidirectional; that is, for any two candidate sites i and j , i ∈ , j ∈ , a directed graphG Including from i arrive j The passage path ( i , j ), also including j arrive i The passage path ( j , i ).

[0067] The directed graph created in this step G It is a blank map that represents the current state of public transportation in the area. During the design phase, any bus line can select alternative sites to set up, and then set the route of the bus line based on the communication paths between the sites.

[0068] S2: In a directed graph G Determine any bus route l The line alternative stations, line starting arcs, line ending arcs and line virtual arcs are then used to establish the corresponding operation map G l .

[0069] Considering that the designed bus routes should be in the directed graph G Therefore, in this embodiment, the route design schemes of all bus lines are also abstracted into a graph, which is recorded as the operation map. G l Before optimization, any bus route should be able to run between all stations in the directed graph. Therefore, the initial operation graph should contain the directed graph G In addition, since each bus line usually has a fixed starting point, the alternative stations corresponding to the first and last stops should be removed from the operation diagram of each bus line.

[0070] Specifically, for any bus route l , l ∈ L , where L represents the set of all existing bus routes before optimization; its operation map G l The construction method is as follows:

[0071] (1) The original bus routes in the directed graph G are l The starting point of the line ori ( l ) and the end point of the line dep ( l ) are removed, and the remaining sites are used as line alternative sites.

[0072] (2) Take each path in G starting from the original route starting point as the route starting point arc , which satisfies the following formula:

[0073]

[0074] in, Indicates bus routes l The first stops that can be reached after departing from the starting point of the bus line are collected.

[0075] Each pass path that reaches the end point of the original line is used as the line end arc , which satisfies the following formula:

[0076]

[0077] in, Indicates bus routes l Meet at the last stop before reaching the end of the bus route.

[0078] The remaining travel path is used as the line operation arc , which satisfies the following formula:

[0079]

[0080] in, Indicates the bus route except the starting point ori ( l ) and the end point of the line dep ( l ) is the set of alternative sites for all lines except the first one.

[0081] (3) Define the jump path from the line starting point directly to the line end point without passing through the alternative stations of each line as the line virtual arc .

[0082] The virtual arc of the route is a custom arc in this embodiment. The bus route with the virtual arc of the route can reach the end point of the route "out of thin air" from the starting point of the route, which means that the bus route does not actually run on the road and does not pass through the directed graph. G The concept of virtual arcs is introduced in this embodiment to indicate the deleted bus routes in the optimization process. When all the candidate sites in a bus route operation diagram are removed in the optimization process, a virtual arc is used to connect the starting point and the end point of the bus route, indicating that the bus route has been deleted.

[0083] S3: In a directed graph G Determine any passenger p The alternative boarding points, alternative alighting points, passenger boarding arcs, passenger travel arcs, passenger transfer arcs and passenger virtual arcs in the travel demand are used to establish the corresponding demand map G p .

[0084] In this embodiment, considering that the travel needs of all passengers when taking public transportation can only follow the travel paths in the directed graph G, the travel needs of all passengers can also be abstracted into corresponding demand maps based on the directed graph G. G p Specifically, each passenger's behavior of swiping a card to board a bus can be abstracted into a demand map. p , p ∈ P , P represents the set of all users. Its demand map G p The construction method is as follows:

[0085] (1) The corresponding passengers in G p The boarding location of historical ride records ori ( p ) and drop-off location dep ( p ) are used as alternative boarding points and alternative alighting points for the current user.

[0086] Unlike the fixed starting and ending points of bus routes, the boarding and alighting points of each passenger's ride are not actually fixed. For example, for a passenger, assuming that there are multiple stations around him, and only the bus line that passes through one of the stations can reach the destination he needs to go, then the passenger will inevitably use this station as the boarding point. However, assuming that the different bus lines that can be taken from the various stations around him can all reach the corresponding destinations, it is actually uncertain which station the passenger chooses as the boarding point. Based on this situation, this embodiment uses multiple stations within a specified radius around the actual boarding point of each user's ride as alternative boarding points for the passenger, and accordingly, uses multiple stations within a specified radius around the actual alighting point of each user's ride (the maximum walking distance that the user can accept when riding) as alternative alighting points for the passenger. Among them, N ori as well as N dep They are the set of alternative boarding sites and the set of alternative alighting sites for all passengers.

[0087] (2) Change G from Each departure path is used as the passenger boarding arc ; It satisfies the following formula:

[0088]

[0089] in, Indicates passengers p On bus routesl A collection of alternative boarding stops on the .

[0090] arrive Each path of the passenger drop-off arc , which satisfies:

[0091]

[0092] in, Indicates passengers p On bus routes l A collection of alternative drop-off stops on the .

[0093] The remaining travel path is taken as the passenger operation arc. p On the same bus route l The travel path between the alternative line stations is called the passenger operation arc. Because it is the same as the line operation arc, in order to reduce the number of parameters, it is also used express.

[0094] (3) Definition of passenger p The operation between alternative route stations belonging to different bus lines is called a passenger transfer arc, in which passengers p From one of the bus lines l The alternative stop i of the route goes to another bus route The jump path of the line candidate site j is recorded as . It satisfies the following formula:

[0095]

[0096] in, and Respectively represent bus routes l , The set of candidate sites in , and .

[0097] In the demand map of this embodiment, if Figure 2 As shown in FIG, the transfer behavior can be characterized by adding a skip connection arc between the stations in the middle of the operation maps of two bus lines. The skip connection arc indicates that the passenger jumps from the operation map of one bus line to the operation map of another bus line.

[0098] In practice, there are two types of transfer behaviors for passengers during the ride: same-station transfer and different-station transfer. Same-station transfer refers to the situation where the physical stations are at the same location but belong to different bus lines. For passengers, the transfer fee (time cost) of same-station transfer is lower. Different-station transfer refers to the situation where the physical stations are at different locations and belong to different bus lines. The transfer fee of different-station transfer is higher. In this embodiment, not any two line stations can transfer. Only when two physical stations are connected and the distance between each station obtained by the Floyd-Warshall algorithm is within the specified distance range (corresponding to the maximum transfer walking distance acceptable to passengers), can passengers transfer between the line stations represented by these two physical stations.

[0099] (4) Define the jump path from the alternative boarding point to the alternative alighting point without passing through the alternative stations of each line as the passenger virtual arc .

[0100] Like the running virtual arc, the passenger virtual arc of this embodiment is also a newly defined arc, which indicates that the user ultimately chooses not to take public transportation when traveling. For example, in the designed bus route, if the travel demand of a passenger whose departure point or destination is in a remote area is too far away from the station of the optimized bus route, or the optimized bus route cannot be directly reached, resulting in too high time cost, the user will no longer choose to take public transportation. In this regard, this embodiment directly jumps the boarding point and the alighting point in the user's travel demand in the demand map, indicating that the current bus network cannot meet the user's travel demand.

[0101] S4: In order to solve the bus station site selection and route design problem in this embodiment, it is necessary to model the problem as a line network optimization model. Among them, based on the operation maps of each bus line and the demand maps of all users established above, it can be seen that the bus station site selection and route design problem that needs to be solved in this embodiment is essentially to delete the alternative sites included in each operation map, and then set the travel path between the sites. Therefore, when modeling the bus line network optimization problem in this embodiment, all arcs in the operation maps of all bus lines are used as bus path selection decision variables, and all arcs in all demand maps are used as passenger path selection decision variables.

[0102] In this embodiment, the optimized bus routes should be able to cover the travel needs of all users in the demand map as much as possible, and minimize the cost of each user taking the bus, which can include time cost and economic cost, etc. At the same time, the operating cost of the bus company should be minimized. On this basis, in the line network optimization model established in this embodiment, the unit cost of each type of arc contained in the operation map and the demand map is preset, and the optimization goal is to minimize the comprehensive cost including the travel cost of all passengers and the operating cost of all bus routes. Specifically, the expression of the optimization goal is as follows:

[0103] (1)

[0104] In the above formula, c p ( i , j ) means that the passenger is in any The arc ( i , j )’s unit operating cost; represents the passenger path selection decision variable; if the passenger uses arc but is 1, otherwise is 0; P represents the set of all bus users; L represents the set of all bus routes. A l Indicates running map G l The set of all arcs contained in , which satisfies: . A p Demand map G p The set of all arcs contained in , which satisfies: . c l ( i , j ) means any bus route belonging to A l The arc ( i , j )’s unit operating cost; represents the bus route selection decision variable. If the bus uses arc but is 1, otherwise is 0.

[0105] In the above objective function corresponding to the optimization target, the first part It refers to the total travel cost of all passengers on the passenger operation arc, which corresponds to the cost spent by passengers after getting on the bus in their actual riding behavior. It refers to the total travel cost of all passengers on the passenger boarding arc, passenger transfer arc, passenger alighting arc and passenger virtual arc, which corresponds to the cost of preparing to board, giving up, boarding, alighting, transferring, etc. in the actual riding behavior of passengers when they do not get on the bus. It represents the total operating cost of the bus line on the line starting arc, line running arc, line end arc and line virtual arc.

[0106] Among the various types of arcs included in the operation map and demand map, the passenger virtual arc refers to the bus line that fails to meet the travel needs of customers, which is actually not in line with the original intention of public transportation settings. Therefore, a larger unit cost should be set for the passenger virtual arc. The line virtual arc refers to a deleted bus line, which can reduce the operating costs of the bus company. Therefore, a lower unit cost should be set for the line virtual arc, or even set to 0. Passenger transfer arcs can be further divided into different types of transfer arcs such as same-station transfers and different-station transfers. The unit cost of each type of transfer arc should be set differently. The unit costs of the remaining types of arcs are usually set accordingly according to the distance between stations.

[0107] Finally, in the line network optimization model, the constraints of line setting rationality constraints are used. In this embodiment, the line setting rationality constraints include bus line flow balance constraints, passenger path flow balance constraints, coupling constraints, line length constraints, transfer number constraints, path integrity constraints, and sub-loop elimination constraints.

[0108] (I) Bus line flow balance constraint means that when the bus line is opened, each operation map must contain a line starting arc and a line ending arc, and any line alternative station in the operation map must have a line operation arc that arrives at the station and departs from the station at the same time.

[0109] The expression of bus line flow balance constraint is as follows:

[0110] (2)

[0111] (3)

[0112] (4)

[0113] Among them, formula (2) means: there must be a bus line starting from i The arc with the starting point as i , j ). Formula (3) shows that there must be a bus route ending atj is the arc reaching the point ( i , j ). Formula (4) shows that: for bus routes l Line stations on j , based on the line station j The arc starting from The number of arcs ending at the line station ( i , j The established bus line flow balance constraint ensures that each bus line l There is a connected route from the starting point of the bus route to the end point of the bus route.

[0114] (II) Passenger path flow balance constraint means that if the passenger takes public transportation, each demand map must contain a passenger boarding arc and a passenger alighting arc, and any alternative station on the demand map must have a route operation arc that arrives at the station and departs from the station.

[0115] The expression of passenger path flow balance constraint is as follows:

[0116] (5)

[0117] (6)

[0118] (7)

[0119] Among them, formula (5) indicates that there must be a path from the passenger’s starting point i The arc with the starting point as i , j ). Formula (6) indicates that there must be a route with the passenger’s destination j is the arc reaching the point ( i , j ). Formula (7) shows that for a physical site j , the arc starting from the physical site and the number of arcs ending at the physical site ( i , j The established passenger path flow balance constraint ensures that each passenger p From the passenger boarding point to the passenger alighting point is a complete passenger travel path.

[0120] (III) Coupling constraint means that each passenger operation arc in the demand map must have a corresponding line operation arc in the operation map; and the number of passenger operation arcs allocated to each line operation arc shall not exceed the preset line capacity C.

[0121] The expression of the coupling constraint is as follows:

[0122] (8)

[0123] Formula (8) means: For each passenger operation arc ( i , j ), can only be assigned to the opened line operation arc ( i , j ), the total number of passenger running arcs allocated on each opened line running arc shall not exceed the line capacity C of the line running arc.

[0124] (IV) Route length constraint means that the total distance from the starting point to the end point of any running map is not greater than the preset length limit. L max . Its expression is:

[0125] (9)

[0126] (V) Transfer frequency constraint means that the number of transfer stations included in each demand map does not exceed the preset transfer frequency limit. M . Its expression is:

[0127] (10)

[0128] (VI) Route integrity constraint means that the previous path and the next path of the passenger transfer arc contained in any demand map are not allowed to be the passenger getting off arc and the passenger getting on arc at the same time. Its expression is:

[0129] (11)

[0130] (12)

[0131] According to the actual situation of passenger travel, if a passenger uses a transfer arc during the travel process, there will be no passengers getting on and off the bus at the same time before and after the transfer arc, which avoids the unrealistic situation that the passenger transfers as soon as he walks to the boarding station or arrives at the alighting station as soon as he transfers. Formula (11) shows that for passenger p, if the transfer arc is used ( i , j ) walks from station i to station j for transfer, then the arc with station i as the arrival point must be the passenger running arc. Formula (12) shows that for passenger p, if the transfer arc ( i , j ) Walk from station i to station j for transfer, then the arc starting from station j must be the passenger running arc.

[0132] (VII) Sub-loop elimination constraint means that no loop is allowed between the stations other than the passenger starting point and the passenger end point in each demand route. Its expression is:

[0133]

[0134] The constraints defined above require that the stations in the complete line of the operation map should be connected and accessible to the starting and ending points of the line. Figure 3 As shown, the optimized line graph may still contain errors including small loops, which are not realistic, but meet the above constraints. Therefore, this embodiment eliminates this situation by setting sub-loop elimination constraints. The sub-loop elimination constraint adopts Miller-Tucker-Zemlin (MTZ), which uses the access order of each site to limit the generation of loops, where: is an integer greater than or equal to 0, indicating the order in which site i is visited. For example, if site 2 is the third visit in the route plan, then =3. N is the number of all sites in the run map.

[0135] S5: Use the variable neighborhood search algorithm to iteratively optimize the line network optimization model, and add a transfer selection operation between the disturbance operation and the repair operation of the variable neighborhood search algorithm to accelerate the iteration. Figure 4 As shown in Figure 2, the iterative optimization process performed by the variable neighborhood search algorithm is as follows:

[0136] S51: Encode the feasible solution consisting of the operation map of each bus line and initialize the objective function.

[0137] In this embodiment, a solution to the problem is encoded by multiple vectors having the same number of elements as the set of bus routes L. Each vector in the solution represents a bus route, and the length of the vector representing each bus route is not fixed and needs to be determined according to the travel demand of passengers allocated to each subsequent bus route.

[0138] This embodiment stipulates that each bus line starts from the bus line departure station and ends at the bus line terminal station. Therefore, in the solution encoding, the beginning and end of each bus line must be the bus line departure station and the bus line terminal station. Figure 5 As shown, O represents the starting point of the bus line, and D represents the end point of the bus line.

[0139] When passenger p takes bus route l When traveling, the bus routes lIt must pass through one of the passenger's boarding stations and one of the passenger's alighting stations. Under this premise, if Figure 6 As shown, there are two situations when passengers travel:

[0140] Case 1: When there is no transfer for passengers, only one bus line is used for travel. For example, when a passenger A can choose to board "1, 2" within a certain walking distance from his departure point, and can choose to get off "9, 10" within a certain walking distance from his destination. Then, if route 1 is used to meet the travel needs of the passenger, the path of route 1 can be: "O→1→9→D", "O→1→10→D", "O→2→9→D", "O→2→10→D". If the objective function is calculated, the path plan "O→2→9→D" is the path with the smallest objective function, that is, route 1 starts from point O, picks up passengers at platform "2", goes to platform "9" to send the passenger, and finally route 1 arrives at point D to terminate the trip, then we choose this path plan to meet the travel needs of passenger A.

[0141] Case 2: When passengers have to transfer during their trip, that is, they use more than one bus line to travel. For example: if passenger C can choose to board at "2, 3" within a certain walking distance from his starting point, and can choose to get off at "8, 9" within a certain walking distance from his destination. The first transfer point selected is: "7", and the second transfer point selected is: "6". If the passenger uses bus line 2 for the first half of the trip and chooses bus line 3 for the second half of the trip, then after the objective function calculation, bus line 2 starts from point O, picks up passengers at platform "3", goes to platform "7" for passengers to transfer, and finally arrives at point D to terminate the trip. Bus line 3 starts from point O, picks up passenger C who transferred from bus line 2 at platform "6", goes to platform "8" to drop off passengers, and finally arrives at point D to terminate the trip.

[0142] S52: On the premise of satisfying the constraint conditions, the sites included in the demand maps of all users are inserted into the feasible solution in turn to obtain an initial solution.

[0143] In this embodiment, the construction of the initial solution adopts the greedy idea, and the best option in the current state is taken in each step, that is, it does not consider the overall optimal solution, but pursues the local optimal solution in a certain sense. Before generating the initial solution, two facts must be understood first. Each time a passenger travel demand is selected and a route is assigned to it:

[0144] 1. For each passenger, there is at least one alternative boarding point and one alternative alighting point, so only the best boarding and alighting point that meets the passenger's travel needs is selected. For each passenger, there is a corresponding set of alternative boarding points and a set of alternative alighting points. In the final route plan, the passenger can only choose the best boarding and alighting point to travel.

[0145] 2. If the passenger chooses to transfer to a certain station in the first leg of the journey, and for each station, there is a corresponding set of transfer stations, then the passenger must transfer to at least one station in the second leg of the journey. In the final route plan, the passenger can only choose a pair of optimal transfer stations to travel.

[0146] On this basis, insert each passenger's travel demand into the bus route in turn. When inserting, consider both the insertion without transfer and the insertion with transfer, and then choose the insertion method that minimizes the increase in the objective function. The following are the steps for the two insertion methods:

[0147] (1) Insertion without transfer:

[0148] First, iterate over the passengers p A set of candidate boarding stations and a set of candidate alighting stations, from which one boarding station and one alighting station are selected, and then Figure 7 As shown, the boarding and alighting stations are inserted into each encoded bus route in sequence. O , D For each position between the two, the alighting station is required to be inserted after the boarding station. Finally, the increase in the objective function after the insertion is determined.

[0149] (2) Insertion with transfer:

[0150] First, iterate over the passengers p A set of candidate boarding points and a set of candidate alighting points, from which one boarding station and one alighting station are selected, and then Figure 8 As shown, select a station as the transfer station. Select two bus lines accordingly, and try to insert the boarding station and the transfer station into the first bus line in turn. O , D For each position between , determine the increase in the objective function after insertion. Select a transfer boarding station from the transfer boarding station set corresponding to the transfer boarding station, and try to insert the transfer boarding station and the alternative boarding station into the second bus line in turn. O , D For each position between the two, the alighting station is required to be inserted after the transfer boarding station. Finally, the increase in the objective function after the insertion is determined.

[0151] When inserting passenger travel demand into a bus route, it is necessary to consider that the constraints cannot be violated during the insertion. Constraint 1: Maximum length constraint of the bus route. When inserting a passenger travel demand into a bus route, if the length of the bus route exceeds the maximum route length limit after the insertion, the insertion method is not available; Constraint 2: Maximum capacity constraint of the bus route. When the passenger travel demand of a certain bus route has reached the maximum capacity of the bus route, no more passenger demand can be inserted into this bus route.

[0152] After trying to insert each passenger's travel demand, if an insertion plan is generated, the insertion plan that minimizes the increase in the objective function is selected from all feasible insertion plans and saved as the initial solution for the next insertion of passenger travel demand, and the current objective function is updated. If the passenger's travel demand violates the constraints after trying all the insertion plans, the passenger can only travel by virtual arc, that is, by other travel modes other than public transportation.

[0153] The above steps are then repeated for all passengers’ travel demands until all passengers’ travel demands are attempted to be inserted, and finally the travel route selections for all passengers and the station arrangement plans for each bus route are generated. The vectors containing all passengers’ travel demands and bus route plans are taken as the initial solutions of the algorithm and recorded as X init .

[0154] S53: using a preset perturbation operator to perturb the initial solution obtained in the previous step.

[0155] The first step of the variable neighborhood search algorithm to iterate is to perturb the initial solution. The purpose of the perturbation is to prevent the current solution from falling into the local optimum by using a set of perturbation operators. It uses a perturbation operator to perturb the current solution to an intermediate solution. The perturbation process of this embodiment is as follows:

[0156] S531: First, let the current solution X = X init and the optimal solution X best = X init ;

[0157] S532: Define a set N k Used to store a set of perturbation operators used for perturbation, k is the index using a certain perturbation operator, k max Represents the number of designed perturbation operators in the set.

[0158] S533: Selectk The perturbation operator affects the current solution X Perturb to get the intermediate solution :

[0159] .

[0160] Among them, Shaking X , k ) indicates the solution X Adopt the k A perturbation operator is used to perform the perturbation.

[0161] S54: Using a preset transfer selection operator, a specified number of stations are selected from each operation map in the previous step as transfer stations.

[0162] The core process of solving the optimization problem by the variable neighborhood search algorithm is to repeat "perturbation" and "local search". In particular, for the specific problem of this embodiment, the technicians have added a special transfer selection operation between the two. Specifically, the transfer point selection operation in this embodiment is to select the most likely transfer point set for local search before each local search based on the usage of various transfer points, thereby reducing the local search space and accelerating the algorithm solution process.

[0163] The operation process of transfer point selection is as follows:

[0164] S541: Define a set H i Used to store a set of transfer selection operators, i is the index of the transfer selection operator used, i max Represents the number of designed operators in the set.

[0165] S541: Use the roulette strategy to select a transfer selection operator and obtain a set of candidate transfer sites T:

[0166] .

[0167] in, Indicates all alternative route sites Adopt the H i A transfer selection operator.

[0168] S55: Obtain a new solution through a neighborhood local search. The core of the local search is to conduct a detailed search in the neighborhood of the current solution to find the local optimal solution, ensuring that the solution after each disturbance is fully optimized. The operation process of the local search is as follows:

[0169] S551: Define a set M l Used to store a set of operators for local search. l is the index using a certain operator, l max Represents the number of designed operators in the set.

[0170] S552: Select a local search operator to search for the intermediate solution Perform operations to get new solutions :

[0171] .

[0172] in, Represents the intermediate solution Adopt the l A local search operator is used to perform local search.

[0173] S553: ​​Based on the new intermediate solution, update the value of the objective function and make the following decisions:

[0174] If the objective function value of the solution of the current round is greater than the objective function value of the solution of the previous round, the solution of this round will be used as the initial solution of the next round; otherwise, the solution of the previous round will be used as the initial solution of the next round.

[0175] S56: Repeat steps S53-S55 until the maximum number of iterations is reached; retain the optimal solution that can minimize the objective function value;

[0176] S57: Decode the optimal solution to obtain several optimized bus routes and the station sequence in each bus route. Finally, the routes in each operation map that do not contain route virtual arcs in the optimization results are used as the final bus routes to be opened (if there are route virtual arcs, it means that the bus route is cancelled), and the alternative stations passed by the bus routes contained in each operation map are used as the actual set stations.

[0177] In the solution of this embodiment, the technicians have designed a variety of operators for the disturbance operation, transfer selection operation and local search operation in the neighborhood local search algorithm. The details of each operator used are introduced below:

[0178] 1. Perturbation Operator

[0179] In this embodiment, the disturbance operators that can be selected for the disturbance operation include: transfer point removal operator, boarding and alighting cluster removal operator, and history-based removal operator. Each operator is as follows:

[0180] The transfer point removal operator is used to reallocate the transfer demand of passengers using one transfer point to another transfer point. The operation removes the transfer demand of passengers using a transfer point so that the passenger has a chance to try using another transfer point. The execution strategy of this operator is as follows:

[0181] (1) Randomly select a transfer point to be used.

[0182] (2) Make a judgment: If the number of passengers transferring at the transfer point is less than or equal to the number of passengers to be removed R , then all passengers who use this point for transfer will be removed from the current solution, and the transfer point will also be deleted from the bus route plan. Otherwise, a random number of size R A subset of passengers is selected and all passengers in the subset of passengers are removed from the solution.

[0183] (3) If a passenger subset of size is selected randomly to be removed, then the removed passenger R The passengers are determined according to the following strategy:

[0184] First, calculate all the passengers who transfer at this transfer point (the distance between the boarding point and the boarding point of the passengers who did not take public transportation recently + the distance between the alighting point and the alighting point of the passengers who did not take public transportation recently) / 2. Second, arrange all passengers from small to large according to the results calculated in step 4. Finally, select passengers in order from small to large and delete them from the current solution.

[0185] The boarding and alighting cluster removal operator is used to simultaneously remove a group of passengers who may use the same transfer point. The scenario it targets is: if there are two or more passengers whose boarding or alighting stations are close to each other, then the close boarding or alighting stations can form a cluster, and passengers in the same cluster may use the same bus line to travel, or transfer at the same transfer point. The execution strategy of this operator is as follows:

[0186] (1) Arrange the boarding and alighting points of all passengers within the specified radius d Clustering is done by randomly selecting a root node for each cluster.

[0187] (2) Arrange all passengers’ boarding and alighting stops according to their distance to the root node from near to far to obtain a list.

[0188] (3) Iterative deletion from the list in the previous step j The positions are: of passengers, among whom y is a random number between [0, 1], p is a fixed parameter, which is taken as 9 in this embodiment.P is the set of all passengers in the list of the previous step, U is the set of passengers who do not travel on bus routes.

[0189] The history removal operator considers the usage of each transfer point in the optimal solution of each iteration, finds the transfer point that satisfies the least passenger demand, and then deletes the passenger demand of the transfer point so that the passenger has a chance to try to insert using other solutions. The execution strategy of this operator is as follows:

[0190] (1) Count the number of passenger transfer demands satisfied by each transfer station in the current solution, and arrange the transfer points from small to large according to the number of passenger transfer demands satisfied by each transfer point.

[0191] (2) Delete the transfer point that satisfies the least number of passengers’ transfer demands, and delete the passengers who transfer at this transfer point from the current solution.

[0192] 2. Transfer Selection Operator

[0193] In this embodiment, the available transfer selection operators include: random selection operator, nearest neighbor operator, centralized transfer operator and historical performance operator. Each operator is as follows:

[0194] The random selection operator is used to randomly select a m A subset of potential transfer points is used, and these transfer points are tried in each passenger demand insertion. The execution strategy of this operator is as follows:

[0195] (1) Randomly select from all the alternative transfer points m transfer points are saved as a subset.

[0196] (2) In the subsequent local search, transfer points are only selected from this subset.

[0197] The nearest neighbor operator is used to select the nearest m The execution strategy of this operator is as follows:

[0198] (1) Calculate the sum of the distances from each station to all unassigned passenger boarding stations and the distances between all unassigned passenger alighting stations.

[0199] (2) Arrange these sites in ascending order according to the sum of their distances.

[0200] (3) Select the one with the smallest sum of distances m stations as transfer points.

[0201] The concentrated transfer operator is used to find the most transfers that can meet the most transfer demands after the disturbance operation. m The execution strategy of this operator is as follows:

[0202] (1) In the intermediate solution after the perturbation operator operation, the number of passenger transfer demands satisfied by each transfer point is counted.

[0203] (2) Arrange the transfer points in descending order according to the number of transfer needs they can satisfy.

[0204] (3) Select the bus that meets the largest number of passengers’ transfer needs m stations as transfer points.

[0205] The historical performance operator is defined to use the travel cost of each transfer point according to a certain passenger demand, and provide transfer points to be considered for the next selection. The execution strategy of this operator is as follows:

[0206] (1) For each solution generated in each iteration, count the number of times each passenger uses a certain transfer point and record the optimal travel cost of the passenger when using the transfer point. i Passing the transfer point t travel, then the optimal travel cost for the passenger is h i，t If the transfer point t Never been a passenger i Use, then .

[0207] (2) For each passenger, a roulette wheel is used to select the transfer point to be considered based on the passenger's optimal travel cost through each transfer point.

[0208] (3) All passenger demands to be assigned are processed in the same manner as in the above two steps; all transfer points selected by the roulette wheel are used as candidate transfer points to be considered in the local search.

[0209] 3. Local Search Operator

[0210] The local search operators used in the neighborhood local search include: the optimal insertion operator that simultaneously considers transfers, the optimal insertion operator that gives priority to transfers, and the insertion operator with the largest gap.

[0211] The optimal insertion operator that considers transfers at the same time means: for each passenger, the cost of insertion without transfer is calculated, and then the insertion with transfer is considered. For transfer points, the insertion costs from the boarding station to the transfer drop-off point and from the transfer boarding point to the drop-off station are calculated separately, and then the insertion method with the lowest total cost is selected. The execution strategy of this operator is as follows:

[0212] (1) Calculate the objective function value of the intermediate solution obtained in the previous step.

[0213] (2) Insert the passenger travel demand into the optimal position of the intermediate solution without transfer, and calculate its objective function value.

[0214] (3) Consider inserting into the middle solution with transfer For the objective function of insertion with transfer, the distance from the passenger boarding station to the transfer drop-off point and the transfer drop-off point to the passenger drop-off station are calculated accordingly, and then the insertion position that minimizes the total objective function is selected.

[0215] (4) Compare the difference between the objective functions of the two insertion methods and the objective function of the first step, and select the insertion scheme that minimizes the increase in the objective function.

[0216] The optimal insertion operator with transfer priority is a strategy that prioritizes insertion with transfers. Under this strategy, the system will force the use of transfer points even if the cost of direct insertion is lower, because this may lead to a better solution overall. Although the cost of using the insertion with transfer for a single passenger demand is slightly higher, the overall cost is reduced.

[0217] For example: Fig. 9 As shown, for passenger demand 1 and passenger demand 2, when passenger demand 1 directly starts from its starting point p 1 to the end d 1, the cost is 9. When passenger demand 1 moves from its starting point p 1Passing the transfer point T To the end d 1, the cost is 10 (5+5). The same is true for passenger demand 2. In this scenario, although the cost of a single passenger demanding a transfer is slightly higher (from 9 to 10), if both passengers choose to transfer, the total cost for the two passengers as a whole is reduced from 18 (9+9) to 15 (5+5+5). In other words, although transfer increases the cost of a single passenger, it can reduce the total cost of all passengers.

[0218] The execution strategy of the optimal insertion operator that prioritizes transfers is as follows:

[0219] (1) For passengers who do not take bus routes, the first option is to use an insertion method that involves transfers.

[0220] (2) If no feasible insertion solution with transfer is found, consider insertion without transfer and insert the passenger demand.

[0221] (3) When there is no feasible insertion solution for the remaining passenger demand, it is assigned to the virtual arc.

[0222] (4) Iteratively delete the passenger demand that is inserted with transfers, and try to see if the travel cost is lower by inserting without transfers. If so, update the insertion plan for the passenger demand.

[0223] The insertion operator with the largest gap means that for each unassigned passenger demand, the difference between the best insertion cost using the insertion method with transfer and the best insertion cost without transfer is calculated. The passenger demand with the largest difference is inserted to its best position first. The execution strategy of this operator is as follows:

[0224] (1) Calculate the optimal insertion cost of each unassigned passenger demand using the insertion method with transfer and the optimal insertion cost without transfer.

[0225] (2) Calculate the difference between the costs of using these two insertion methods for each passenger demand.

[0226] (3) Calculate the cost gap for each passenger and arrange them in order from largest to smallest.

[0227] (4) Insert passenger demands into their optimal positions according to the gap from largest to smallest.

[0228] For each of the above perturbation operators, transfer selection operators and local search operators, such as Fig.10 As shown, in a further optimized solution of this embodiment, a scoring mechanism is introduced in each round of iteration, and then the performance of each operator used in the transfer selection operation is evaluated after each round of iteration, and then the selection probability of each operator in the next round is adjusted according to the performance evaluation value.

[0229] Example 2

[0230] Based on the scheme of Example 1, this embodiment further provides a bus station site selection and route design device based on passenger demand, which includes a memory, a processor, and a computer program stored in the memory and running in the processor. When the processor executes the computer program, the bus station site selection and route design method based on passenger demand as in Example 1 is adopted to achieve redesign of all existing bus routes in a designated area, thereby achieving optimal cost while meeting the travel needs of most passengers.

[0231] The bus stop site selection and route design device based on passenger demand provided in this embodiment is essentially a computer device used to implement the solutions in embodiments 1 and 2.

[0232] In this embodiment, the computer device may also be a smart terminal, tablet computer, laptop computer, desktop computer, rack server, blade server, tower server or cabinet server (including an independent server or a server cluster composed of multiple servers) that can execute programs.

[0233] The computer device indicated in this embodiment includes at least but is not limited to: a memory and a processor that can be connected to each other through a system bus. Among them, the memory (i.e., a readable storage medium) includes a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory can be an internal storage unit of a computer device, such as a hard disk or a memory of the computer device. In other embodiments, the memory can also be an external storage device of a computer device, such as a plug-in hard disk equipped on the computer device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory can also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory is generally used to store an operating system and various application software installed on the computer device. In addition, the memory can also be used to temporarily store various types of data that have been output or are to be output.

[0234] The processor may be a central processing unit (CPU), a graphics processing unit (GPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor is generally used to control the overall operation of a computer device. In this embodiment, the processor is used to run program codes stored in a memory or process data.

[0235] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A bus station site selection and route design method based on passenger demand, which is used to redesign the existing bus routes in a specified area, thereby achieving optimal cost while meeting the travel needs of most passengers; characterized by: It includes the following steps: S1: Establish a directed graph representing the current regional traffic conditions based on the geographical characteristics and road distribution of the specified area G : ； in, Represents the set of all candidate sites that meet the site setting conditions; A represents the set of travel paths between any two candidate sites that can be equally reached; S2: In a directed graph G Determine any bus route l The line alternative stations, line starting arcs, line ending arcs and line virtual arcs are then used to establish the corresponding operation map G l ; S3: In a directed graph G Determine any passenger p The alternative boarding points, alternative alighting points, passenger boarding arcs, passenger running arcs, passenger transfer arcs and passenger virtual arcs in the travel demand are used to establish the corresponding demand map G p ; S4: taking all arcs in the operation maps of all bus routes as bus route selection decision variables, and taking all arcs in all demand maps as passenger route selection decision variables; presetting the unit cost of each type of arc contained in the operation map and the demand map, and then taking the minimum comprehensive cost including the travel cost of all passengers and the operation cost of all bus routes as the optimization goal; establishing a line network optimization model based on the constraint condition of line setting rationality constraint; S5: Adopting a variable neighborhood search algorithm to iteratively optimize the line network optimization model, and adding a transfer selection operation between the disturbance operation and the repair operation of the variable neighborhood search algorithm to accelerate iteration; finally, the lines in each operation map that do not contain line virtual arcs in the optimization results are used as the final bus lines to be opened, and the alternative stations passed by the bus lines contained in each operation map are used as the actual set stations.

2. The bus stop location selection and route design method based on passenger demand as claimed in claim 1, characterized in that: In step S2, any bus route l Running map G l The construction method is as follows: (1) Remove the candidate sites corresponding to the starting point and the end point of the original route in the directed graph, and use the remaining sites as the route candidate sites ; (2) Take each pass path in G starting from the original route starting point as the route starting point arc ; Each pass path that reaches the end point of the original line is used as the line end arc ; The remaining travel path is used as the line operation arc ; (3) Define the jump path from the line starting point directly to the line end point without passing through each line alternative station as the line virtual arc .

3. The bus stop site selection and route design method based on passenger demand as claimed in claim 2, characterized in that: In step S3, any passenger p Demand Map G p The construction method is as follows: (1) The corresponding passengers in G p Each candidate station within a specified radius around the boarding and alighting locations of the historical ride record is used as the current user's candidate boarding point and alternative drop-off points ; (2) Change G from Each departure path is used as the passenger boarding arc ;arrive Each path of the passenger drop-off arc ; The remaining travel path is the passenger running arc, still recorded as ; (3) Definition of passenger p The jump path from the line alternative station i of one bus line to the line alternative station j of another bus line is the passenger transfer arc ; (4) Define the jump path from the alternative boarding point to the alternative alighting point without passing through the alternative stations of each line as the passenger virtual arc .

4. The bus stop location selection and route design method based on passenger demand as claimed in claim 3, characterized in that: In step S4, the expression of the optimization objective is as follows: ， ， ， In the above formula, c p ( i , j ) means that the passenger is in any The arc ( i , j )’s unit operating cost; represents the passenger path selection decision variable; P represents the set of all bus users; L represents the set of all bus routes; A l Indicates running map G l The set of all arcs contained in ; A p Demand map G p The set of all arcs contained in ; c l ( i , j ) means any bus route belonging to A l The arc ( i , j )’s unit operating cost; represents the bus route selection decision variable.

5. The method for bus station location selection and route design based on passenger demand as claimed in claim 1, characterized in that: In the line network optimization model, the line setting rationality constraints include bus line flow balance constraints, passenger path flow balance constraints, coupling constraints, line length constraints, transfer number constraints, path integrity constraints and sub-loop elimination constraints.

6. The method for bus station location selection and route design based on passenger demand as claimed in claim 5, characterized in that: The bus line flow balance constraint means that when the bus line is opened, each operation map must contain a line starting arc and a line ending arc, and any line alternative station in the operation map must have a line operation arc that arrives at the station and departs from the station at the same time; The passenger path flow balance constraint means that if the passenger takes public transportation, each demand map must contain a passenger boarding arc and a passenger getting off arc, and any line alternative station in the demand map must have a line running arc that arrives at the station and departs from the station at the same time; The coupling constraint means that each passenger operation arc in the demand map must have a corresponding line operation arc in the operation map; And the number of passenger running arcs allocated to each line running arc is not greater than the preset line capacity C; The route length constraint means that the total distance from the route start point to the route end point of any running map is not greater than the preset length limit L max ; The transfer number constraint means that the number of transfer stations included in each demand map does not exceed the preset transfer number upper limit M; The line integrity constraint means that the previous pass path and the next pass path of the passenger transfer arc included in any demand map are not allowed to be a passenger getting off arc and a passenger getting on arc at the same time; The sub-loop elimination constraint means that no loop is allowed to be formed between the remaining stations except the passenger starting point and the passenger end point in each demand route.

7. The method for bus station location selection and route design based on passenger demand as claimed in claim 1, characterized in that: In step S5, the iterative optimization process performed by the variable neighborhood search algorithm is as follows: S51: Encode the feasible solution consisting of the operation map of each bus line and initialize the objective function: S52: On the premise of satisfying the constraint conditions, inserting the sites included in the demand maps of all users into the feasible solution in turn to obtain an initial solution; S53: using a preset perturbation operator to perturb the initial solution obtained in the previous step; S54: using a preset transfer selection operator to select a specified number of stations from each operation map in the previous step as transfer stations; S55: Re-obtain a new intermediate solution through a neighborhood local search, update the value of the objective function, and make the following decisions: If the objective function value of the solution of the current round is greater than the objective function value of the solution of the previous round, the solution of this round will be used as the initial solution of the next round; otherwise, the solution of the previous round will be used as the initial solution of the next round; S56: Repeat steps S53-S55 until the maximum number of iterations is reached; retain the optimal solution that can minimize the objective function value; S57: Decode the optimal solution to obtain a number of optimized bus routes and a station sequence in each bus route.

8. The method for bus station location selection and route design based on passenger demand as claimed in claim 7, characterized in that: In step S53, the disturbance operators include: a transfer point removal operator, an alighting and alighting cluster removal operator, and a history-based removal operator; In step S54, the transfer selection operator includes: a random selection operator, a nearest neighbor operator, a centralized transfer operator and a historical performance operator; In step S55, the operators used in the neighborhood local search include: an optimal insertion operator that simultaneously considers transfers, an optimal insertion operator that gives priority to transfers, and an insertion operator of the maximum gap.

9. The method for bus station location selection and route design based on passenger demand as claimed in claim 8, characterized in that: In step S54, a scoring mechanism is introduced in each round of iteration, and then the performance of each operator used in the transfer selection operation is evaluated after each round of iteration is completed, and then the selection probability of each transfer selection operator in the next round is adjusted according to the performance evaluation value.

10. A bus stop location selection and route design device based on passenger demand, comprising a memory, a processor, and a computer program stored in the memory and running in the processor, characterized in that: When the processor executes the computer program, the bus stop site selection and route design method based on passenger demand as described in any one of claims 1 to 9 is adopted to redesign the existing bus routes in the specified area, thereby achieving the optimal cost while meeting the travel needs of most passengers.

Images