A travel route planning method for electric vehicles considering the energy supply capacity of scenic spots

Abstract

This invention discloses a method for planning electric vehicle travel routes that considers the energy supply capacity of a scenic area. The method comprises: obtaining information on all scenic spots and charging stations in the scenic area; establishing a multi-objective function and corresponding constraints, including total route time, number of charging times, attribute diversity, and number of scenic spots; generating a set of feasible solutions for routes that meet the constraints based on the information of all scenic spots and all charging stations, with each path in the set of feasible solutions including all required attractions and some non-required attractions; using the set of feasible solutions as the initial population, a genetic algorithm is used to solve the multi-objective problem based on the multi-objective function and the constraints, obtaining a Pareto front solution, which is then input into a simulated annealing algorithm to obtain the optimal route planning result. By comprehensively considering constraints such as charging, route, and scenic area attribute diversity, the method aims to optimize electric vehicle travel routes within the scenic area and improve travel efficiency.

Description

Technical Field

[0001] The present invention relates to the field of route planning, and in particular to a travel route planning method for electric vehicles taking into account the energy supply capacity of a scenic area. Background Art

[0002] With the increasing popularity of electric vehicles, their route planning has gradually become an important research area, especially for holiday travel planning. Compared with traditional vehicles, electric vehicles face many challenges when traveling. Especially when traveling long distances, due to their own battery life, travel route planning needs to consider the location of refueling, which brings certain challenges. Tourism travel often occurs during holidays, and traffic conditions can also affect the travel range of electric vehicles, which in turn affects travel route planning. Therefore, compared with traditional route planning problems, electric vehicle travel needs to consider more practical factors.

[0003] Most of the existing electric vehicle travel route planning algorithms are based on the shortest path algorithms of graph theory, such as the Dijkstra algorithm and the A* algorithm. Combined with the actual energy consumption model, they solve the path selection problem by calculating the path with the minimum energy consumption or the shortest path. For the multi-objective path planning problem, the optimization algorithm based on the NSGA genetic algorithm is used to realize the multi-objective path planning selection by calculating the Pareto front solution.

[0004] Currently, most electric vehicle route planning algorithms focus primarily on conventional route planning. However, travel requires a wide range of practical considerations, making them difficult to address the unique needs of electric vehicle scenic area tourism. For example, current electric vehicle route planning approaches lack comprehensive consideration of factors such as the variety of attractions to be visited, the number of charging times, maximizing attraction coverage, and the number of must-visit attractions. They also neglect actual holiday traffic conditions, leaving the practical challenges of electric vehicle travel unresolved.

[0005] Therefore, existing technologies have great limitations and cannot meet the path planning needs of electric vehicles under the complex requirements of travel. Summary of the Invention

[0006] The technical problem to be solved by the present invention is as follows: In response to the above-mentioned problems of the prior art, a travel route planning method for electric vehicles is provided that takes into account the energy supply capacity of scenic spots. It comprehensively considers constraints such as charging, routes, and diversity of scenic spot attributes, aiming to optimize the travel routes of electric vehicles within scenic spots and improve travel efficiency.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0008] A method for planning electric vehicle travel routes taking into account the energy supply capacity of a scenic area comprises the following steps:

[0009] Obtain information about all scenic spots in the scenic area and all charging piles, including must-see and non-must-see attractions;

[0010] A multi-objective function and corresponding constraints are established, including the total route time, number of charging times, attribute diversity, and number of attractions. Based on the information of all attractions and all charging piles, a set of feasible solutions for the path that meets the constraints is generated. Each path in the set of feasible solutions includes all must-see attractions and some non-must-see attractions.

[0011] The set of feasible solutions of the path is used as the initial population, and the genetic algorithm is used to solve the multi-objective problem according to the multi-objective function and constraints to obtain the Pareto front solution of the path. The Pareto front solution of the path is input into the simulated annealing algorithm to obtain the optimal result as the final path planning result.

[0012] Furthermore, the multi-objective function includes:

[0013] Minimize the total path time F1, the formula is as follows:

[0014] min F1= ∑T(i)+∑D(i,i+1) / V*L(i,i+1)

[0015] Where i is the node passed by the path, T(i) is the stay time at node i, D(i,i+1) is the distance from node i to node i+1, V is the driving speed of the electric vehicle, L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L, and D(i,i+1) / L(i,i+1) is the equivalent distance from node i to node i+1;

[0016] Minimize the number of charging times F2, the formula is as follows:

[0017] min F2= C

[0018] Where C is the number of charging times of the path;

[0019] Maximize attribute diversity F3, the formula is as follows:

[0020] max F3= -∑plnp

[0021] Where p=S j / ΣS j , j is the attribute of the scenic spot, S j is the number of scenic spots with attribute j in the path, ΣS j is the total number of attractions in the route;

[0022] Maximize the number of scenic spots F4, the formula is as follows:

[0023] max F4= S

[0024] Where S is the number of scenic spots on the path.

[0025] Furthermore, the constraints include:

[0026] Time constraint: The total travel time of the route cannot exceed the maximum allowed time, which is calculated as follows:

[0027] ∑T(i)+∑D(i,i+1) / V*L(i,i+1)≤T MAX

[0028] Where i is the node passed by the path, T(i) is the stay time at node i, D(i,i+1) is the distance from node i to node i+1, V is the driving speed of the electric vehicle, L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L, D(i,i+1) / L(i,i+1) is the equivalent distance from node i to node i+1, T MAX is the maximum allowed time;

[0029] Charge times constraint: The charge times are less than or equal to the set maximum charge times. The formula is as follows:

[0030] C≤C MAX

[0031] C is the number of charging times of the current path, C MAX Maximum number of charging times;

[0032] Battery power constraint: The battery power in the route must not exceed the remaining battery power in each section. The formula is as follows:

[0033] Y(i)-E(i,i+1)≥0

[0034] Where Y(i) is the remaining battery charge of the electric vehicle at node i, and E(i,i+1) is the energy consumption of the electric vehicle from node i to node i+1. The formula is as follows:

[0035] E(i,i+1)=α·D(i,i+1) / L(i,i+1)

[0036] Among them, α is the unit energy consumption of electric vehicles, D(i,i+1) is the distance from node i to node i+1, and L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L.

[0037] Constraints on must-visit scenic spots: The must-visit scenic spots are fully covered in the path. The formula is as follows:

[0038] R(b)=1,b∈B

[0039] Where b is a scenic spot in the set B of must-see scenic spots, and R(b) is a 0-1 variable. When R(b)=1, it means that the current path includes scenic spot b, otherwise it does not.

[0040] Furthermore, the information of the scenic spot includes attribute information and location information of the scenic spot. When generating a set of feasible solutions of the path that meets the constraint conditions based on the information of all scenic spots and the information of all charging piles, it specifically includes:

[0041] Get the starting point of the electric car and all the must-see attractions and randomly arrange them to get the initial node of the path. The starting point of the electric car is used as the starting point and end point of the path;

[0042] Count the attribute distribution of all attractions in the path, select non-essential attractions with the same attributes as the one with the smallest percentage as new nodes and insert them between the start and end points of the path. Then count the attribute distribution of all attractions in the path again until the specified number of non-essential attractions is reached.

[0043] Perform path repair, then use a multi-objective function to calculate the total path time, number of charging times, attribute diversity, and number of attractions, and determine whether the corresponding constraints are met. If not, re-execute the steps of obtaining the starting point of the electric vehicle and all must-visit attractions and randomly arrange them. If the corresponding constraints are met, add the path to the set of feasible solutions.

[0044] Furthermore, the information about the scenic spot also includes the tour time, and the information about the charging pile includes the location information and charging time of the charging pile. When performing path repair, the following steps are included:

[0045] Check whether the starting point and end point of the current path are both the starting point of the electric car. If not, modify the starting point and end point to the starting point of the electric car.

[0046] Check whether the current path includes all the must-see attractions. If there are any missing must-see attractions, insert the missing must-see attractions before the end point;

[0047] Initialize the battery to be fully charged, calculate the energy consumption between two nodes section by section from the starting point of the current path, and update the remaining power to each node. When the remaining power is insufficient, find the charging station closest to the corresponding node and add it as a new node to the current path, and update the charging times. If the charging times exceed the corresponding constraint, randomly delete a node where a non-essential tourist attraction is located from the current path, and then re-execute the steps of initializing the battery to be fully charged, calculating the energy consumption between two nodes section by section from the starting point of the current path, and updating the remaining power to each node, until the nodes of the current path are traversed and the charging times meet the corresponding constraint.

[0048] The total path time is obtained by adding up the driving time of each section of the current path, the visiting time of each scenic spot, and the charging time of each charging pile. If the total path time exceeds the corresponding constraint, the non-essential scenic spot with the most repeated attributes in the current path is deleted, and then the steps of adding up the driving time of each section of the current path, the visiting time of each scenic spot, and the charging time of each charging pile are repeated again until the total path time of the current path meets the corresponding constraint, thus obtaining the final overall path of the current path.

[0049] Furthermore, the charging piles include scenic spot charging piles located within the scenic spot and independent charging piles located outside the scenic spot. When the remaining power is insufficient, the charging pile closest to the corresponding node position is found and added to the current path, and the charging times are updated, specifically including:

[0050] If the remaining power of the current node is less than the energy consumption between the current node and the next node, then find the scenic spot charging pile or independent charging pile with the nearest equivalent distance to the current node. If the remaining power of the current node can reach the scenic spot charging pile or independent charging pile, insert the scenic spot charging pile or independent charging pile as a new node between the current node and the next node, and update the power to full charge, and increase the number of charging times by one. If the remaining power of the current node cannot reach the scenic spot charging pile or independent charging pile, then find the scenic spot charging pile or independent charging pile with the nearest equivalent distance to the previous node as a new node and insert it between the previous node and the current node, and update the power to full charge, and increase the number of charging times by one. The equivalent distance is the actual distance between the two nodes divided by the congestion coefficient between the two nodes.

[0051] Furthermore, when using genetic algorithms to solve multi-objective problems based on multi-objective functions and constraints, it specifically includes:

[0052] Take each path in the initial population as a parent;

[0053] Randomly select two parent paths, perform crossover and mutation operations, and obtain the child path;

[0054] Use a multi-objective function to calculate the total path time, number of charging times, attribute diversity, and number of scenic spots of the child path, and determine whether the corresponding constraints are met. If not, the path is repaired until the total path time, number of charging times, attribute diversity, and number of scenic spots of the child path meet the corresponding constraints.

[0055] Merge the parent path and the child path to obtain a new population. Use a multi-objective function to calculate the total path time, number of charging times, attribute diversity, and number of scenic spots for each path in the new population. Based on the calculation results, perform non-dominated sorting on the paths in the new population, count the number of dominated paths for each path, and classify them into different frontier levels.

[0056] If the maximum number of iterations has not been reached, the individual congestion of all paths in each frontier level is calculated. Specifically, the individual congestion is the sum of the normalized congestion of each objective function in the multi-objective function. In descending order of frontier level, the paths in each frontier level are added to the population of the next iteration in descending order of congestion. The paths in the population of the next iteration are used as new parents, and two parent paths are randomly selected to perform crossover and mutation operations.

[0057] If the maximum number of iterations is reached, the path in the highest frontier level is output as the Pareto frontier solution.

[0058] Furthermore, the crossover operation specifically refers to exchanging the road sections between the nodes where the common must-see attractions are located in the two parent paths to obtain child paths, and eliminating the nodes where the duplicate attractions are located in each child path. Specifically, for the child paths with duplicate attractions, the node where the corresponding attraction appears for the first time is retained, and other nodes of the corresponding attractions are eliminated; the mutation operation includes: statistically analyzing the attribute distribution of all attractions in the path, selecting the nodes where the non-must-see attractions have the same attributes as the attributes with the smallest proportion and inserting them between the starting point and the end point of the path, or randomly deleting the nodes where the non-must-see attractions are located in the path, or selecting the nodes where the attractions corresponding to the attraction charging piles are located to replace the nodes where the independent charging piles are located.

[0059] Furthermore, when the Pareto front solution of the path is input into the simulated annealing algorithm to obtain the optimal result, it specifically includes:

[0060] Convert the multi-objective function into a single objective function and calculate the single objective function value of each path in the Pareto front solution. The formula of the single objective function is as follows:

[0061]

[0062] Where F1, F2, F3, and F4 are the total path time, number of charging times, attribute diversity value, and number of scenic spots of a path in the Pareto front solution output by the genetic algorithm, respectively. W 、C W 、D W 、S W are the maximum total path time, maximum number of charging times, minimum attribute diversity value, and minimum number of scenic spots of the Pareto front solution output by the genetic algorithm, respectively. B 、C B 、D B 、S B are the minimum total path time, minimum number of charging times, maximum attribute diversity value, and maximum number of scenic spots of the Pareto front solution output by the genetic algorithm, respectively. T 、W C 、W D、W S They are the weight proportions of total route time, number of charging times, attribute diversity value, and number of scenic spots;

[0063] Perform neighborhood operations on each path in the Pareto front solution to obtain the corresponding new path. Use a multi-objective function to calculate whether the new path meets all constraints, including total path time, number of charging times, attribute diversity, and number of attractions. Then determine whether the corresponding constraints are met. If not, repair the path until the total path time, number of charging times, attribute diversity, and number of attractions of the new path meet the corresponding constraints.

[0064] Calculate the single objective function value of each new path, and calculate the difference between each new path and the single objective function value of the corresponding original path in the Pareto front solution. If the single objective function difference Δf between the original path and the new path is less than 0, the new path is directly accepted, otherwise The acceptance probability of the new path is accepted, where T is the temperature. If the new path is accepted, the corresponding original path in the Pareto front solution is updated to the new path. If the new path is not accepted, the original path in the Pareto front solution is retained.

[0065] Lower the temperature T and perform the steps of performing neighborhood operations on each path in the Pareto front solution of the path until the temperature T is lower than the termination temperature T F , output the path with the largest single objective function value in the Pareto front solution as the final path.

[0066] Furthermore, the neighborhood operation includes: exchanging the positions of the nodes where two must-see attractions are located in the path, or selecting a node where a non-must-see attraction is located that is not in the path to replace the node where a non-must-see attraction is located in the path, or selecting the node where the attraction corresponding to the attraction charging pile is located to replace the node where the independent charging pile is located, or statistically analyzing the attribute distribution of all attractions in the path, selecting the node where the non-must-see attraction has the same attribute as the attribute with the smallest proportion and inserting it between the start and end points of the path, or deleting the node where the non-must-see attraction is located in the path, or reversing the order between any two nodes in the path.

[0067] Compared with the prior art, the advantages of the present invention are:

[0068] The present invention establishes a multi-objective function including the total path time, the number of charging times, the diversity of attributes and the number of scenic spots and the corresponding constraints, and generates a set of feasible solutions for the path that meets the constraints. It can obtain a path that meets multiple constraints such as power constraints and time constraints, and realize the application of energy charging facilities for scenic spots.

[0069] The present invention adopts a path repair mechanism to ensure that the paths that meet the constraint conditions include all must-see attractions and some non-must-see attractions, so that the paths cover some additional attractions beyond the must-see attractions, thereby meeting the travel needs of travelers.

[0070] The present invention uses the genetic-simulated annealing algorithm to effectively realize multi-objective tourism route planning, wherein the genetic algorithm encodes the path and uses crossover, mutation and other methods to find the optimal solution to the multi-objective problem, and then cooperates with the simulated annealing algorithm to break out of the limitations of local optimality and obtain multi-objective optimization paths such as time, scenic area attributes, and power consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Schematic diagram of a flow chart of an embodiment of the present invention.

[0072] Figure 2 Schematic diagram of the final path planning result of an embodiment of the present invention. DETAILED DESCRIPTION

[0073] The present invention will be further described below in conjunction with the accompanying drawings and specific preferred embodiments, but the scope of protection of the present invention is not limited thereby.

[0074] The current electric vehicle tourism route planning method has a mismatch between the algorithm and actual needs. For example, the current planning method hardly considers the scenic area attribute factors that affect travelers' travel, nor does it take into account the actual situation of charging facilities in the scenic area to flexibly select energy replenishment resources.

[0075] In order to fill the gap in the current multi-constraint and multi-objective planning methods for electric vehicle holiday travel route planning, this embodiment proposes an electric vehicle travel route planning method that considers the energy supply capacity of scenic spots. This is a multi-objective planning method that combines travel demand and energy supply resources. It can consider conditions such as scenic spot charging supply, scenic spot attributes, time, distance, etc., and provide a multi-objective optimized path to traverse the must-visit scenic spots.

[0076] like Figure 1 As shown, the method specifically comprises the following steps:

[0077] S100) Acquiring information about all scenic spots and all charging piles in the scenic area. In this embodiment, the scenic spots include must-see scenic spots and non-must-see scenic spots. The charging piles include scenic spot charging piles located within the scenic spot and independent charging piles located outside the scenic spot. The relevant information about the scenic spot charging piles can be obtained by using the relevant attribute parameters of the charging facilities in the corresponding scenic spot.

[0078] S101) Establishing a multi-objective function including total route time, number of charging times, attribute diversity, and number of scenic spots and corresponding constraints;

[0079] S102) generating a set of feasible solutions for paths that meet the constraints based on the information of all scenic spots and all charging piles, wherein each path in the set of feasible solutions includes all must-see scenic spots and some non-must-see scenic spots;

[0080] S103) Using the set of feasible solutions of the path as the initial population, using a genetic algorithm to solve the multi-objective problem according to the multi-objective function and constraints, and obtaining a Pareto front solution of the path;

[0081] S104) Inputting the Pareto front solution of the path into the simulated annealing algorithm to obtain the optimal result as the final path planning result.

[0082] Through the above steps, the final path planning result is as follows Figure 2 As shown in the figure, a path that satisfies multiple constraints such as power constraints and time constraints can be obtained, and the application of energy-recharging facilities at scenic spots can be realized. At the same time, the path covers some additional attractions besides the must-see attractions to meet the travel needs of travelers. Finally, the genetic-simulated annealing algorithm can effectively realize multi-objective tourism path planning.

[0083] The following describes the relevant steps in detail.

[0084] In step S100, the scenic spot information includes attribute information, visit time, and location information, and the charging pile information includes the charging time and location information of the charging pile. Based on the location information of the scenic spot and the charging pile, a topological map can be established. The nodes in the map include all scenic spots and independent charging piles. The scenic spot charging piles belong to the same node as the scenic spot. For each node in the topological map, the actual distance between them can be calculated to obtain the distance matrix D. At the same time, by actually observing the travel time between holidays and between each node and comparing it with the corresponding benchmark data, the corresponding congestion coefficient can be obtained, and the congestion coefficient matrix L can be established.

[0085] Step S101 is intended to set constraints on travel time, power consumption, number of charging times, and attribute duplication, and to establish a multi-objective function that considers factors such as equivalent distance and scenic area attributes. Specifically, the multi-objective function of this embodiment includes:

[0086] Minimize the total path time F1, the formula is as follows:

[0087] min F1= ∑T(i)+∑D(i,i+1) / V*L(i,i+1)(1)

[0088] Where i is the node passed by the path, T(i) is the dwell time at node i, which specifically refers to the time spent at the scenic spot node when not charging, or the time spent at the scenic spot charging station or independent charging station node when charging, D(i,i+1) is the distance from node i to node i+1 in the distance matrix D, V is the driving speed of the electric vehicle, L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L, and D(i,i+1) / L(i,i+1) is the equivalent distance from node i to node i+1;

[0089] Minimize the number of charging times F2, the formula is as follows:

[0090] min F2 = C(2)

[0091] Where C is the number of charging times of the path;

[0092] Maximize attribute diversity F3, the formula is as follows:

[0093] max F3 = -∑plnp (3)

[0094] Where p=S j / ΣS j , j is the attribute of the scenic spot, S j is the number of scenic spots with attribute j in the path, ΣS j is the total number of attractions in the route;

[0095] Maximize the number of scenic spots F4, the formula is as follows:

[0096] max F4= S(4)

[0097] Where S is the number of scenic spots on the path.

[0098] Correspondingly, the constraints include:

[0099] Time constraint: The total travel time of the route cannot exceed the maximum allowed time, which is calculated as follows:

[0100] ∑T(i)+∑D(i,i+1) / V*L(i,i+1)≤T MAX (5)

[0101] Where i is the node passed by the path, T(i) is the stay time at node i, D(i,i+1) is the distance from node i to node i+1 in the distance matrix D, V is the driving speed of the electric vehicle, L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L, D(i,i+1) / L(i,i+1) is the equivalent distance from node i to node i+1, T MAX is the maximum allowed time;

[0102] Charge times constraint: The charge times are less than or equal to the set maximum charge times. The formula is as follows:

[0103] C≤C MAX (6)

[0104] C is the number of charging times of the current path, C MAX Maximum number of charging times;

[0105] Battery power constraint: The battery power in the route must not exceed the remaining battery power in each section. The formula is as follows:

[0106] Y(i)-E(i,i+1)≥0(7)

[0107] Where Y(i) is the remaining battery charge of the electric vehicle at node i, and E(i,i+1) is the energy consumption of the electric vehicle from node i to node i+1. The formula is as follows:

[0108] E(i,i+1)=α·D(i,i+1) / L(i,i+1)(8)

[0109] Where α is the unit energy consumption of electric vehicles, D(i,i+1) is the distance from node i to node i+1 in the distance matrix D, and L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L.

[0110] Constraints on must-visit scenic spots: The must-visit scenic spots are fully covered in the path. The formula is as follows:

[0111] R(b)=1, b∈B(9)

[0112] Where b is a scenic spot in the set B of must-see scenic spots, and R(b) is a 0-1 variable. When R(b)=1, it means that the current path includes scenic spot b, otherwise it does not.

[0113] Step S102 of this embodiment aims to generate feasible solutions that meet the requirements of time, power, number of charging times, and attribute repetition by repairing the constraint mechanism, and to create an initial population with the feasible solutions as a set. Specifically, when generating a set of feasible solutions for paths that meet the constraints based on the information of all scenic spots and all charging piles, the following steps are specifically included:

[0114] S201) Obtain the starting point of the electric vehicle and all must-see attractions and randomly arrange them to obtain an initial node set R = [O, B1, B2, ... Bn, O] of the path, where O represents the starting point of the electric vehicle, the set R serves as the starting point and end point of the path, and B1, B2, ... Bn represent all must-see attractions after random arrangement;

[0115] S202) randomly inserting non-essential attractions, specifically, counting the attribute distribution of attractions corresponding to all nodes in the node set R of the path, selecting non-essential attractions with the same attribute as the attribute with the smallest proportion as new nodes added to the node set R and inserting them between the starting point and the end point of the path, and then counting the attribute distribution of all attractions in the path again until a specified number of non-essential attractions are inserted;

[0116] S203) performs path repair, and then uses the formula of the multi-objective function, i.e., formula (1) to formula (4), to calculate the total time, number of charging times, attribute diversity and number of scenic spots of the path, and uses formula (5) to formula (9) to judge whether these calculation results meet the corresponding constraints. If the corresponding constraints are not met, return to step S201 and re-execute the steps of obtaining the starting point of the electric vehicle and all the must-visit scenic spots and performing random arrangement to regenerate the path. If the corresponding constraints are met, the path is added to the set of feasible solutions.

[0117] S204) Repeat steps S201 to S203 until the number of paths in the set of feasible solutions reaches a specified population size.

[0118] In step S203, when performing path repair, specifically adjusting the non-essential attractions to achieve a path that meets the constraint conditions, the following steps are included:

[0119] S301) Check whether the starting point and the end point in the node set R of the current path are both the starting point O of the electric vehicle. If not, modify the starting point and the end point to be the starting point O of the electric vehicle;

[0120] S302) Check whether the node set R of the current path includes nodes corresponding to all must-see attractions. If there are any missing must-see attractions, insert the nodes corresponding to the missing must-see attractions before the end point of the node set R;

[0121] S303) Initialize the battery to be fully charged. Starting from the starting point of the node set R of the current path, calculate the energy consumption between two nodes according to formula (8) and update the remaining power to each node. The remaining power of the current node is equal to the remaining power of the previous node minus the energy consumption between the previous node and the current node. When the remaining power is insufficient, find the charging pile closest to the corresponding node as a new node to add to the current path and update the charging times. Specifically, if the remaining power of the current node is less than the energy consumption between the current node and the next node, then according to the calculation formula of the equivalent distance D(i,i+1) / L(i,i+1) Find the nearest scenic spot charging pile or independent charging pile with an equivalent distance to the current node. If the remaining power of the current node can reach the scenic spot charging pile or independent charging pile, insert the scenic spot charging pile or independent charging pile as a new node between the current node and the next node, update the power to full, and increase the number of charging times by one. If the remaining power of the current node cannot reach the scenic spot charging pile or independent charging pile, find the nearest scenic spot charging pile or independent charging pile with an equivalent distance to the previous node, insert it as a new node between the previous node and the current node, update the power to full, and increase the number of charging times by one.

[0122] Since there are rechargeable scenic spots, in order to distinguish whether the nodes in the node set R are used for charging or for sightseeing, this embodiment can add a charging flag to the nodes in the node set R. The charging flag is 0 by default. In step S303, the charging flag of the newly added node in the node set R can be set to 1;

[0123] S304) Use formula (2) to calculate the number of charging times. If the number of charging times exceeds the corresponding constraint in formula (6), randomly delete a node where a non-must-visit scenic spot is located from the node set R of the current path, then jump to step S303 and re-execute the steps of initializing the power to full power, calculating the energy consumption between two nodes section by section from the starting point of the current path, and updating the remaining power to each node, until the nodes of the current path are traversed and the number of charging times meets the corresponding constraint.

[0124] In order to avoid the existence of redundant nodes, when the number of charging times exceeds the corresponding constraint condition in formula (6), the nodes with the charging flag set to 1 in the node set R can be removed first, and then a step of randomly deleting a node where a non-must-visit attraction is located from the node set R of the current path is executed;

[0125] It should be noted that if a path that meets the constraints cannot be obtained after repeating steps S303 and S304 for a maximum number of times, the current path is discarded and the process returns to step S201 to re-execute the steps of obtaining the starting point of the electric vehicle and all the must-see attractions and randomly arrange them to regenerate a path.

[0126] S305) According to formula (1), the travel time of each section in the current path node set R (i.e., the travel time between each two adjacent nodes), the sightseeing time of each scenic spot (nodes with the charging flag set to 0), and the charging time of each charging pile (nodes with the charging flag set to 1) are accumulated to obtain the total path time. If the total path time exceeds the constraint condition corresponding to formula (5), the non-essential scenic spot with the most repeated attributes in the current path node set R is deleted, and then the travel time of each section in the current path node set R, the sightseeing time of each scenic spot, and the charging time of each charging pile are accumulated again according to formula (1) to obtain the total path time, until the total path time of the current path meets the corresponding constraint condition, and the final overall path of the current path is obtained.

[0127] It should be noted that if step S305 is repeated until there are no non-must-see attractions and still no path that meets the constraints can be obtained, the current path is discarded and the process returns to step S201 to re-execute the steps of obtaining the starting point of the electric vehicle and all the must-see attractions and randomly arranging them to regenerate the path.

[0128] Step S103 of this embodiment is to iterate the initial population using a genetic algorithm to obtain a Pareto frontier solution, i.e., a solution with the shortest travel time, the least number of charging times, the lowest attribute duplication, the largest number of scenic spots, etc., and includes the following steps:

[0129] S401) Each path in the initial population is used as a parent. In this embodiment, the paths in the population are encoded into a chromosome format, where each chromosome represents a path, and the points in the chromosome are arranged in the order in which the paths pass through.

[0130] S402) Randomly select two parent paths, perform crossover and mutation operations, and obtain a child path;

[0131] In this embodiment, the crossover operation specifically refers to exchanging the road segments between the nodes of the common must-visit scenic spots in the two parent paths to obtain two child paths, and removing the nodes of the duplicate scenic spots in each child path. Specifically, it is checked whether there are two or more nodes in each child path corresponding to the same scenic spot. For the child paths with duplicate scenic spots, the node of the first appearance of the corresponding scenic spot is retained, and the other nodes of the corresponding scenic spots are removed.

[0132] In this embodiment, the mutation operation includes: statistically analyzing the attribute distribution of all scenic spots in the path, selecting the node where the non-essential scenic spot has the same attribute as the attribute with the smallest proportion and inserting it between the starting point and the end point of the path, or randomly deleting the node where the non-essential scenic spot is located in the path, or selecting the node where the scenic spot charging pile is located to replace the node where the independent charging pile is located;

[0133] S403) Use the multi-objective functions of formula (1) to formula (4) to calculate the total path time, number of charging times, attribute diversity and number of scenic spots of the child path, and determine whether the corresponding constraints in formula (5) to formula (9) are met. If the corresponding constraints are not met, execute steps S301 to S305 to repair the path until the total path time, number of charging times, attribute diversity and number of scenic spots of the child path meet the corresponding constraints;

[0134] S404) The parent path and the child path are merged to obtain a new population. The total path time, charging times, attribute diversity, and number of scenic spots of each path in the new population are calculated using the multi-objective functions of formula (1) to formula (4). The paths in the new population are non-dominated according to the calculation results. Specifically, for the maximization objective function, the function value of path A is not less than that of path B. For the minimization objective function value, path A is not greater than that of path B. Moreover, if path A is greater than path B on at least one maximization objective function or less than path B on at least one minimum function, then path A dominates path B, and the number of B dominated increases by 1.

[0135] After obtaining the dominated number of each path, the dominated number of each path is counted and divided into different frontier levels, so that each path is divided into different frontier levels according to the dominated number;

[0136] S405) Check whether the number of loops has reached the maximum number of iterations. If not, calculate the individual congestion of all paths in each frontier level. The individual congestion is specifically the sum of the normalized congestion of each objective function in the multi-objective function. The calculation process is as follows:

[0137] First, use formula (1) to formula (4) to calculate the values ​​of the four objective functions of each path, sort them in descending order according to the different objective functions, and obtain the congestion degree of each path in this objective function according to the congestion degree formula. The congestion degree of the boundary solution (maximum and minimum solutions) is infinite, and the formula is as follows:

[0138] X i =P i-1 -P i+1 (10)

[0139] Among them, P i For the i-th value in descending order of a target function, X i P i The congestion degree of the corresponding path in this objective function;

[0140] Then, the normalized congestion is calculated as follows:

[0141] X i ’ = Xi / X B -X W (11)

[0142] X B 、X W For the maximum and minimum values ​​in descending order, X i ’ P i The corresponding path in this objective function is the normalized congestion;

[0143] Finally, the normalized crowding degree of all objective functions of the individual is accumulated to obtain the individual crowding degree;

[0144] After obtaining the individual congestion of all paths, the paths in each frontier level are added to the next iteration of the population in descending order of congestion, in descending order of the frontier level. Specifically, if the number of individuals in the current dominant level is less than the population requirement after all individuals in the current dominant level are added to the population, the individuals of the next dominant level are selected in descending order until the population requirement is reached.

[0145] The path in the next iteration population is used as the new parent and the process jumps to step S402 to randomly select two parent paths, perform crossover and mutation operations, and thus start the next cycle.

[0146] If the maximum number of iterations is reached, the path in the highest frontier level in step S404 is output as the Pareto frontier solution.

[0147] Step S104 of this embodiment aims to optimize the Pareto front solution using a simulated annealing algorithm to obtain a multi-objective weighted optimal path, and specifically includes the following steps:

[0148] S501) Convert the multi-objective function into a single objective function, and calculate the single objective function value of each path in the Pareto front solution. The formula of the single objective function is as follows:

[0149] (12)

[0150] Where F1, F2, F3, and F4 are the total path time, number of charging times, attribute diversity value, and number of scenic spots of a path in the Pareto front solution output by the genetic algorithm, respectively. W 、C W 、D W 、S W are the maximum total path time, maximum number of charging times, minimum attribute diversity value, and minimum number of scenic spots of the Pareto front solution output by the genetic algorithm, respectively. B 、C B 、D B 、S Bare the minimum total path time, minimum number of charging times, maximum attribute diversity value, and maximum number of scenic spots of the Pareto front solution output by the genetic algorithm, respectively. T 、W C 、W D 、W S They are the weight proportions of total route time, number of charging times, attribute diversity value, and number of scenic spots;

[0151] S502) performing a neighborhood operation on each path in the Pareto front solution of the path to obtain a corresponding new path, wherein the neighborhood operation specifically includes: exchanging the positions of two nodes at which must-see attractions are located in the path, or selecting a node at which a non-must-see attraction is located that is not in the path to replace a node at which a non-must-see attraction is located in the path, or selecting a node at which a scenic spot corresponding to a scenic spot charging pile is located to replace a node at which an independent charging pile is located, or statistically analyzing the attribute distribution of all scenic spots in the path, selecting a node at which a non-must-see attraction has the same attribute as the attribute with the smallest proportion and inserting it between the start and end points of the path, or deleting a node at which a non-must-see attraction is located in the path, or reversing the order of any two nodes in the path;

[0152] Use the multi-objective function of formula (1) to formula (4) to calculate whether the new path meets all constraints, the total path time, the number of charging times, the attribute diversity and the number of scenic spots, and determine whether the corresponding constraints of formula (5) to formula (9) are met. If the corresponding constraints are not met, execute steps S301 to S305 to repair the path until the total path time, the number of charging times, the attribute diversity and the number of scenic spots of the new path meet the corresponding constraints.

[0153] S503) Use formula (12) to calculate the single objective function value of each new path, and calculate the difference between each new path and the single objective function value of the corresponding original path in the Pareto front solution. The formula is as follows:

[0154] Δf=F C -F N (13)

[0155] Among them, F C is the single objective function value of the original path, F N is the single objective function value of the new path;

[0156] S504) If the difference Δf between the original path and the new path is less than 0, the new path is directly accepted, otherwise The acceptance probability of the new path is accepted, where T is the temperature. If the new path is accepted, the corresponding original path in the Pareto front solution is updated to the new path. If the new path is not accepted, the original path in the Pareto front solution is retained.

[0157] S505) Lower the temperature T and jump to step S502 to perform the neighborhood operation on each path in the Pareto front solution of the path until the temperature T is lower than the termination temperature T F , output the path with the largest single objective function value in the Pareto front solution as the final path, and the formula for cooling the temperature T is as follows:

[0158] T N+1 =T N -(T S -T F ) / I (14)

[0159] Where, T N+1 With T N Represents the temperature at the N+1th and Nth iterations, T S With T F Represents the initial temperature and the end temperature, and I is the cooling coefficient.

[0160] The effect of the method of this embodiment is described below through specific experiments.

[0161] Select some scenic spots as the optional attractions for this trip. The traveler drives an electric car from node A (starting point) and wants to visit attractions B, D, and F. There are other attractions C, E, and G along the way, and a pure charging station H for energy supply. Finally, the trip is completed and returned to node A. The electric car battery capacity is 60kwh, which is fully charged at the starting point. The energy consumption of the electric car is assumed to be 0.1kwh / km, the speed is 60km / h, and the distance matrix between each point is D.

[0162] D=[A,B, C, D, E, F, G, H],

[0163] [A,0,120, 150, 200, 300, 350, 280, 100],

[0164] [B,120, 0,100, 180, 250, 400, 320, 150],

[0165] [C, 150, 100, 0,80,280, 320, 350, 180],

[0166] [D,200, 180, 80,0,220, 300, 380, 220],

[0167] [E,300, 250, 280, 220, 0,200, 150, 280],

[0168] [F,350, 400, 320, 300, 200, 0,180, 320],

[0169] [G,280, 320, 350, 380, 150, 180, 0,250],

[0170] [H,100, 150, 180, 220, 280, 320, 250, 0]

[0171] The congestion coefficient matrix is ​​L,

[0172] L= [A , B ,C , D , E ,F,G , H],

[0173] [A, 1.0, 0.8, 1.0, 0.9, 1.0, 1.0, 0.7, 1.0],

[0174] [B ,0.8, 1.0, 0.7, 1.0, 0.9, 1.0, 1.0, 0.9],

[0175] [C ,1.0, 0.7, 1.0, 0.8, 1.0, 0.9, 1.0, 1.0],

[0176] [D ,0.9, 1.0, 0.8, 1.0, 0.7, 1.0, 0.8, 1.0],

[0177] [E ,1.0, 0.9, 1.0, 0.7, 1.0, 0.8, 0.9, 1.0],

[0178] [F ,1.0, 1.0, 0.9, 1.0, 0.8, 1.0, 0.7, 0.9],

[0179] [G ,0.7, 1.0, 1.0, 0.8, 0.9, 0.7, 1.0, 0.8],

[0180] [H ,1.0, 0.9, 1.0, 1.0, 1.0, 0.9, 0.8, 1.0]

[0181] Total time weight W T =0.3, charging times weight W C =0.3, attribute diversity value weight W D =0.2, weight of number of attractions W S =0.2, the parameters of each point are shown in Table 1. The travel time is required to be no more than 50 hours, the maximum number of charging times is 3 times, and the number of attractions with the same scenic spot attributes is at most 3.

[0182] Table 1 Specific parameters of each node

[0183]

[0184] Input the initial parameter population size 50, the number of iterations 50 times, the initial temperature of the simulated annealing algorithm is 1000, the termination temperature is 0, and the cooling coefficient is 5.

[0185] According to step S102, a path that meets the constraints is generated: A -> F -> E -> B -> H -> D -> A. The number of charging times is 2. According to formula (1), the total time of this path is calculated to be 40.56 hours. According to formula (3), the attribute diversity is 1.0549. The number of attractions in the path is calculated to be 4, which meets all the constraints and is added to the initial population. Repeat multiple times until the initial population reaches 50 paths.

[0186] According to step S103, the generated initial population is iterated using a genetic algorithm. After 50 iterations, two Pareto front solutions are obtained. They are the shortest time solution A -> F -> E -> H -> D -> B -> A, with the shortest time being 39.02 hours, and the highest diversity and most attractions solution A -> F -> E -> C -> D -> H -> B -> A, with the maximum attribute diversity of 1.0790 and the maximum number of attractions being 5.

[0187] According to step S104, the leading solution obtained by the genetic algorithm is used as the initial path of the simulated annealing algorithm, and the weighted score F is calculated. The shortest time solution A -> F -> E -> H -> D -> B -> A has a weighted score of 0.9246, and the solution with the highest diversity and the most scenic spots A -> F -> E -> C -> D -> H -> B -> A has a weighted score of 0.9574. Neighborhood operations are then performed to continue optimization until the iteration stopping condition is met.

[0188] The final output solution is the path: A -> F -> E -> H -> B -> D -> C -> A, with a total time of 42.54 hours, two charging times, 1.0790 attribute diversity, 5 attractions, and a weighted score of 1. While satisfying the constraints in advance, the weighted score should be maximized to better meet the travel needs of tourists. Therefore, the path A -> F -> E -> H -> B -> D -> C -> A is selected.

[0189] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for planning electric vehicle travel routes considering the energy supply capacity of scenic spots, characterized by , including the following steps: Obtain information about all scenic spots in the scenic area and all charging piles, including must-see and non-must-see attractions; A multi-objective function including total path time, number of charging times, attribute diversity, and number of scenic spots and corresponding constraints are established. The multi-objective function includes: Minimize the total path time F1, the formula is as follows: min F1 = ∑T(i)+∑D(i,i+1) / V*L(i,i+1) Where i is the node passed by the path, T(i) is the stay time at node i, D(i,i+1) is the distance from node i to node i+1, V is the driving speed of the electric vehicle, L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L, and D(i,i+1) / L(i,i+1) is the equivalent distance from node i to node i+1; Minimize the number of charging times F2, the formula is as follows: min F2 = C Where C is the number of charging times of the path; Maximize attribute diversity F3, the formula is as follows: max F3 = -∑plnp Where p=S j / ΣS j , j is the attribute of the scenic spot, S j is the number of scenic spots with attribute j in the path, ΣS j is the total number of attractions in the route; Maximize the number of scenic spots F4, the formula is as follows: max F4= S Among them, S is the number of scenic spots on the path; The constraints include: Time constraint: The total travel time of the route cannot exceed the maximum allowed time, which is calculated as follows: ∑T(i)+∑D(i,i+1) / V*L(i,i+1)≤T MAX Where i is the node passed by the path, T(i) is the stay time at node i, D(i,i+1) is the distance from node i to node i+1, V is the driving speed of the electric vehicle, L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L, D(i,i+1) / L(i,i+1) is the equivalent distance from node i to node i+1, T MAX is the maximum allowed time; Charge times constraint: The charge times are less than or equal to the set maximum charge times. The formula is as follows: C≤C MAX C is the number of charging times of the current path, C MAX Maximum number of charging times; Battery power constraint: The battery power in the route must not exceed the remaining battery power in each section. The formula is as follows: Y(i)-E(i,i+1)≥0 Where Y(i) is the remaining battery charge of the electric vehicle at node i, and E(i,i+1) is the energy consumption of the electric vehicle from node i to node i+1. The formula is as follows: E(i,i+1)=α·D(i,i+1) / L(i,i+1) Among them, α is the unit energy consumption of electric vehicles, D(i,i+1) is the distance from node i to node i+1, and L(i,i+1) is the congestion coefficient from node i to node i+1 in the congestion coefficient matrix L. Constraints on must-visit scenic spots: The must-visit scenic spots are fully covered in the path. The formula is as follows: R(b)=1,b∈B Where b is a scenic spot in the set B of must-see scenic spots, R(b) is a 0-1 variable, when R(b)=1, it means that the current path includes scenic spot b, otherwise it does not include it; A set of feasible solutions for paths that satisfy the constraints is generated based on the information of all scenic spots and all charging piles. Each path in the set of feasible solutions includes all required scenic spots and some non-required scenic spots. The information of the scenic spots includes attribute information and location information of the scenic spots. When a set of feasible solutions for paths that satisfy the constraints is generated based on the information of all scenic spots and all charging piles, the following is specifically included: Get the starting point of the electric car and all the must-see attractions and randomly arrange them to get the initial node of the path. The starting point of the electric car is used as the starting point and end point of the path; Count the attribute distribution of all attractions in the path, select non-essential attractions with the same attributes as the one with the smallest percentage as new nodes and insert them between the start and end points of the path. Then count the attribute distribution of all attractions in the path again until the specified number of non-essential attractions is reached. Perform path repair, then use a multi-objective function to calculate the total path time, number of charging times, attribute diversity, and number of attractions, and determine whether the corresponding constraints are met. If not, re-execute the steps of obtaining the starting point of the electric vehicle and all the must-visit attractions and randomly arrange them. If the corresponding constraints are met, add the path to the set of feasible solutions; The set of feasible solutions of the path is used as the initial population, and the genetic algorithm is used to solve the multi-objective problem according to the multi-objective function and constraints to obtain the Pareto front solution of the path. The Pareto front solution of the path is input into the simulated annealing algorithm to obtain the optimal result as the final path planning result.

2. The electric vehicle travel route planning method considering the energy supply capacity of scenic spots according to claim 1 is characterized in that The information of the scenic spot also includes the tour time, and the information of the charging pile includes the location information and charging time of the charging pile. When performing path repair, the following steps are included: Check whether the starting point and end point of the current path are both the starting point of the electric car. If not, modify the starting point and end point to the starting point of the electric car. Check whether the current path includes all the must-see attractions. If there are any missing must-see attractions, insert the missing must-see attractions before the end point; Initialize the battery to be fully charged, calculate the energy consumption between two nodes section by section from the starting point of the current path, and update the remaining power to each node. When the remaining power is insufficient, find the charging station closest to the corresponding node and add it as a new node to the current path, and update the charging times. If the charging times exceed the corresponding constraint, randomly delete a node where a non-essential tourist attraction is located from the current path, and then re-execute the steps of initializing the battery to be fully charged, calculating the energy consumption between two nodes section by section from the starting point of the current path, and updating the remaining power to each node, until the nodes of the current path are traversed and the charging times meet the corresponding constraint. The total path time is obtained by adding up the driving time of each section of the current path, the visiting time of each scenic spot, and the charging time of each charging pile. If the total path time exceeds the corresponding constraint, the non-essential scenic spot with the most repeated attributes in the current path is deleted, and then the steps of adding up the driving time of each section of the current path, the visiting time of each scenic spot, and the charging time of each charging pile are repeated again until the total path time of the current path meets the corresponding constraint, thus obtaining the final overall path of the current path.

3. The electric vehicle travel route planning method considering the energy supply capacity of scenic spots according to claim 2 is characterized in that: The charging piles include scenic spot charging piles located within the scenic spot and independent charging piles located outside the scenic spot. When the remaining power is insufficient, the charging pile closest to the corresponding node is found and added to the current path, and the charging times are updated, specifically including: If the remaining power of the current node is less than the energy consumption between the current node and the next node, then find the scenic spot charging pile or independent charging pile with the nearest equivalent distance to the current node. If the remaining power of the current node can reach the scenic spot charging pile or independent charging pile, insert the scenic spot charging pile or independent charging pile as a new node between the current node and the next node, and update the power to full charge, and increase the number of charging times by one. If the remaining power of the current node cannot reach the scenic spot charging pile or independent charging pile, then find the scenic spot charging pile or independent charging pile with the nearest equivalent distance to the previous node as a new node and insert it between the previous node and the current node, and update the power to full charge, and increase the number of charging times by one. The equivalent distance is the actual distance between the two nodes divided by the congestion coefficient between the two nodes.

4. The electric vehicle travel route planning method considering the energy supply capacity of scenic spots according to claim 2 or 3 is characterized in that , when using genetic algorithms to solve multi-objective problems based on multi-objective functions and constraints, specifically including: Take each path in the initial population as a parent; Randomly select two parent paths, perform crossover and mutation operations, and obtain the child path; Use a multi-objective function to calculate the total path time, number of charging times, attribute diversity, and number of scenic spots of the child path, and determine whether the corresponding constraints are met. If not, the path is repaired until the total path time, number of charging times, attribute diversity, and number of scenic spots of the child path meet the corresponding constraints. Merge the parent path and the child path to obtain a new population. Use a multi-objective function to calculate the total path time, number of charging times, attribute diversity, and number of scenic spots for each path in the new population. Based on the calculation results, perform non-dominated sorting on the paths in the new population, count the number of dominated paths for each path, and classify them into different frontier levels. If the maximum number of iterations has not been reached, the individual congestion of all paths in each frontier level is calculated. Specifically, the individual congestion is the sum of the normalized congestion of each objective function in the multi-objective function. In descending order of frontier level, the paths in each frontier level are added to the population of the next iteration in descending order of congestion. The paths in the population of the next iteration are used as new parents, and two parent paths are randomly selected to perform crossover and mutation operations. If the maximum number of iterations is reached, the path in the highest frontier level is output as the Pareto frontier solution.

5. The electric vehicle travel route planning method considering the energy supply capacity of scenic spots according to claim 4 is characterized in that: The crossover operation specifically refers to exchanging the road sections between the nodes where the common must-see attractions are located in the two parent paths to obtain child paths, and eliminating the nodes where the duplicate attractions are located in each child path. Specifically, for the child paths with duplicate attractions, the node where the corresponding attraction appears for the first time is retained, and other nodes of the corresponding attractions are eliminated; the mutation operation includes: counting the attribute distribution of all attractions in the path, selecting the nodes where the non-must-see attractions have the same attributes as the attributes with the smallest proportion and inserting them between the starting point and the end point of the path, or randomly deleting the nodes where the non-must-see attractions are located in the path, or selecting the nodes where the attractions corresponding to the attraction charging piles are located to replace the nodes where the independent charging piles are located.

6. The electric vehicle travel route planning method considering the energy supply capacity of scenic spots according to claim 2 or 3 is characterized in that , when the Pareto front solution of the path is input into the simulated annealing algorithm to obtain the optimal result, it specifically includes: Convert the multi-objective function into a single objective function and calculate the single objective function value of each path in the Pareto front solution. The formula of the single objective function is as follows: Where F1, F2, F3, and F4 are the total path time, number of charging times, attribute diversity value, and number of scenic spots of a path in the Pareto front solution output by the genetic algorithm, respectively. W 、C W 、D W 、S W are the maximum total path time, maximum number of charging times, minimum attribute diversity value, and minimum number of scenic spots of the Pareto front solution output by the genetic algorithm, respectively. B 、C B 、D B 、S B are the minimum total path time, minimum number of charging times, maximum attribute diversity value, and maximum number of scenic spots of the Pareto front solution output by the genetic algorithm, respectively. T 、W C 、W D 、W S They are the weight proportions of total route time, number of charging times, attribute diversity value, and number of scenic spots; Perform neighborhood operations on each path in the Pareto front solution to obtain the corresponding new path. Use a multi-objective function to calculate whether the new path meets all constraints, including total path time, number of charging times, attribute diversity, and number of attractions. Then determine whether the corresponding constraints are met. If not, repair the path until the total path time, number of charging times, attribute diversity, and number of attractions of the new path meet the corresponding constraints. Calculate the single objective function value of each new path, and calculate the difference between each new path and the single objective function value of the corresponding original path in the Pareto front solution. If the single objective function difference Δf between the original path and the new path is less than 0, the new path is directly accepted, otherwise The acceptance probability of the new path is accepted, where T is the temperature. If the new path is accepted, the corresponding original path in the Pareto front solution is updated to the new path. If the new path is not accepted, the original path in the Pareto front solution is retained. Lower the temperature T and perform the steps of performing neighborhood operations on each path in the Pareto front solution of the path until the temperature T is lower than the termination temperature T F , output the path with the largest single objective function value in the Pareto front solution as the final path.

7. The electric vehicle travel route planning method considering the energy supply capacity of scenic spots according to claim 6 is characterized in that: The neighborhood operation includes: exchanging the positions of two nodes where must-see attractions are located in the path, or selecting a node where a non-must-see attraction is located that is not in the path to replace a node where a non-must-see attraction is located in the path, or selecting a node where a scenic spot corresponding to a scenic spot charging pile is located to replace a node where an independent charging pile is located, or statistically analyzing the attribute distribution of all scenic spots in the path, selecting a node where a non-must-see attraction has the same attribute as the attribute with the smallest proportion and inserting it between the start and end points of the path, or deleting a node where a non-must-see attraction is located in the path, or reversing the order between any two nodes in the path.

Images