A route planning method with uncertain vehicle delivery tasks

Abstract

The application discloses a route planning method with uncertainty vehicle distribution tasks, which is applied to vehicle distribution tasks composed of a single warehouse, K vehicles and L distribution task nodes, and comprises the following steps: 1, determining the current scene number, and initializing corresponding parameters; 2, randomly generating an initial population and performing individual evaluation under the current switched scene; 3, performing cross and mutation operations on the population to generate temporary offspring; 4, comparing the population after the cross and the offspring according to an environment selection strategy to obtain a next generation population; 5, evaluating the current population according to a robust evaluation value, and archiving the more robust offspring; 6, if the current environment switching condition is reached, performing step 7; otherwise, returning to step 3; 7, if the termination condition is reached, outputting a final population; otherwise, performing scene switching. The application can balance the optimality and robustness of the route planning scheme, so that a route planning scheme suitable for various scenes can be obtained.

Description

Technical Field

[0001] This invention belongs to the field of vehicle routing problems, specifically a route planning method based on game theory applicable to vehicle delivery tasks with uncertain travel times. Background Technology

[0002] In urban logistics, distribution and transportation are crucial links in the logistics system. Choosing reasonable distribution routes can reduce transportation distance and time, improve the satisfaction of distribution stations, and has significant economic and environmental benefits.

[0003] Route planning methods for addressing the uncertainty of travel time caused by urban traffic can be mainly divided into two categories: exact algorithms and metaheuristic algorithms. Exact algorithms include branch and bound, integer linear programming, and dynamic programming. These exact methods are only suitable for small-scale route planning problems because applying them to large-scale problems is very time-consuming, and they cannot find the optimal solution within a reasonable timeframe. Metaheuristic algorithms include ant colony optimization (ACO), particle swarm optimization (PSO), and genetic algorithms. Ant colony optimization is inspired by the food-searching behavior of ants in nature, using a positive feedback mechanism to continuously search the target space and gradually approach the optimal solution. However, ACO has a slow convergence speed, requiring a long search time to obtain the optimal solution, and is prone to getting trapped in local optima when pheromones are scarce in the population. PSO originates from the study of the predatory behavior of flocks of birds and schools of fish, utilizing the information-sharing function of individuals within the group to obtain the optimal solution. However, PSO is prone to premature convergence and getting trapped in local optima. Genetic algorithms are search heuristic algorithms that mimic the natural selection process. They utilize natural evolutionary techniques such as heredity, mutation, selection, and crossover to generate solutions to the problem to be optimized. Traditional genetic algorithms have good global optimization capabilities, but they also have drawbacks such as low search efficiency leading to slow convergence speed.

[0004] The existing route planning problem with uncertain travel times in urban traffic assumes that, under undisturbed conditions, the travel time from each delivery station to the warehouse, and between delivery stations, is known before route planning. However, in practical applications, the travel time between stations and between stations and warehouses is affected by the road transport environment, making it impossible to accurately predict the exact travel time of each delivery station. When travel times are uncertain, the solutions obtained by previously proposed methods for solving route planning problems with uncertain travel times are rarely cost-effective and resilient to adverse traffic conditions. Furthermore, they may not meet the requirements of all stations in certain emergency scenarios, and it remains difficult to find a high-performance planning scheme that satisfies the requirements of all stations. Summary of the Invention

[0005] The present invention addresses the shortcomings of the existing technology by proposing a route planning method for vehicle delivery tasks with uncertainty. This method aims to balance the optimality and robustness of the route planning scheme, thereby obtaining a route planning scheme that is adaptable to various scenarios and performs well in multiple scenarios.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0007] The present invention provides a route planning method for vehicle delivery tasks with uncertainty, characterized by its applicability to a system consisting of a warehouse (Depot), K delivery vehicles, and L customer stations V = {Cus1, Cus2, ..., Cus...}. i ...,Cus L In the urban transportation environment constituted by}, Cus i Let i represent the i-th delivery station, i∈[1,L]; let the i-th delivery station Cus i and the j-th delivery station Cus j Let the route between them be (i, j), and let the distance between the routes be c. ij ;

[0008] Define a scenario where traffic changes cause changes in travel time as Scene = {sc1, sc2, ..., sc...} p , ...,sc P}, sc p Let p represent the p-th scene, p∈[1,P]; P represents the total number of scenes, and sc0 represents the original scene that has not been disturbed;

[0009] Define arr i For the transport vehicle to reach the i-th delivery station Cus i Arrival time, lb i For the i-th delivery station Cus i Service time window limit, ub i For the i-th delivery station Cus i Service window deadline;

[0010] Let the maximum load capacity of each transport vehicle be Q;

[0011] The route planning method is carried out according to the following steps:

[0012] Step 1: Define the number of times the current scene switches, evngen, and initialize evngen = 0; define the current iteration number as gen;

[0013] Step 2: If evngen = 0, set the current scene for the evnth transition to scene = sc0; where sc0 represents the initial scene. Otherwise, start from Scene = {sc1, sc2, ..., sc...} p , ...,sc P Select a scene from the list as the current scene for the evngen-th transition;

[0014] Initialize gen = 1;

[0015] Step 3: Randomly generate the genth generation population in the current scene. And on All individuals are evaluated;

[0016] Step 3.1: Define the gen-th generation archive population in the current scene as... and initialize Empty;

[0017] Let the gen generation population It contains NP individuals, where the num-th individual is denoted as And the num-th individual K represents the current scene. num Route planning schemes for transport vehicles to visit L delivery stations; and K num ≤K;

[0018] Step 3.2: Initialize num = 1;

[0019] Step 3.3: Randomly generate the gen-th generation population in the current scene. The num-th individual

[0020] Step 3.4: Use equation (1) to process the num-th individual. The evaluation is performed to obtain the num-th travel distance evaluation value in the gen-th generation.

[0021]

[0022] In equation (1), This represents the i-th delivery station Cusi and the j-th delivery station Cus in the current scene. j The distance between them;

[0023] Step 3.5: Use equation (2) to process the num-th individual. An evaluation was conducted to obtain the vehicle count evaluation value.

[0024]

[0025] Step 3.6: After assigning num+1 to num, check if num>NP is true. If true, it means that the population of generation gen has been obtained. Otherwise, return to step 1.2 and execute sequentially;

[0026] Step 4: Based on the two evaluation values, evaluate the population of generation gen. Perform crossover and mutation operations to generate the subpopulation of generation gen.

[0027] Step 4.1, based on A binary tournament selection method is used to generate a mating pool of size NP for the genth generation.

[0028] Step 4.2, from Select the uth individual and the vth individual Two offspring individuals are used, and after repairing them using the crossover operator, the repaired u-th offspring individual is obtained. and the vth offspring individual

[0029] Step 4.3, for and Perform mutation operations;

[0030] Generate a random value p2. If p2 > B, then directly... and As Two offspring individuals; otherwise, use the mutation operator to respectively target the two offspring individuals. and The process was performed, resulting in two mutated offspring individuals that were then placed into... In the equation, B is the mutation probability, and p2∈[0,1];

[0031] Step 4.4: Following the process in Steps 4.2-4.3, obtain NP offspring individuals and form the offspring population of the gen generation.

[0032] Step 5 and Merging yields the merged population of generation gen. And based on environmental selection strategies, the populations are merged from generation gen. Select NP individuals and form the gen+1 generation population.

[0033] Step Six, from and archive population Obtain the save population of generation+1

[0034] Step 6.1: Determine the archived population Is it empty? If it is empty, then... All individuals are stored in the archive population. Otherwise, proceed to step 6.2;

[0035] Step 6.2, and Merging yields the merged population of generation gen. Based on equation (3) After robustness evaluation of all individuals, the evaluation value Pot for all individuals is obtained;

[0036]

[0037] In equation (3), express The robustness evaluation value of the t-th individual; where t≤2NP;

[0038] Step 6.3, according to The evaluation value corresponding to each individual in the data is... The individuals in the population are sorted in ascending order, and the first NP individuals are selected to form the archived population of generation (gen+1). gen+1 ;

[0039] Step 7: After assigning gen+1 to gen, check if gen has reached the maximum number of iterations. max If yes, proceed to step eight; otherwise, return to step two and execute sequentially.

[0040] Step 8: After assigning evngen+1 to evngen, check if evngen has reached the maximum number of scene switching attempts. max If so, then stop iterating and obtain the final archive population under the current scene. Find the optimal route planning scheme for each of the NP individuals, and then... The NP route planning schemes of the population are sorted according to non-dominated individuals to determine the non-dominated individuals, and the route planning schemes corresponding to all non-dominated individuals are selected as the final scheme output; otherwise, proceed to step nine.

[0041] Step 9: Based on the archived population The scenario where the individual's performance confirms the switch;

[0042] Step 9.1: Define the maximum number of individuals that violate the constraint. max and initialize violentmax =0;

[0043] Step 9.2: Initialize p = 1;

[0044] Step 9.3: Use equation (4) to determine the p-th scene sc p Below, archived population The number of infeasible individuals

[0045]

[0046] In equation (4), Represents the p-th scene sc p Down The number of infeasible individuals in the array, Γ represents a Boolean function; if arr i >ub i Then Γ(arr) i -ub i ) = 1; otherwise, Γ(arr) = 1. i -ub i ) = 0; express The t-th individual in;

[0047] Step 9.4, Judgment Is it greater than violent? max If so, Assigned to violent max , sc p Use it as the current scene; otherwise, proceed to step 9.5.

[0048] Step 9.5: After assigning p+1 to p, check if p≥P is true. If yes, proceed to step 9.6; otherwise, return to step 9.3 and execute sequentially.

[0049] Step 9.6: Initialize gen = 1.

[0050] Step 9.7: Based on the evaluation value corresponding to each individual, from Arc gen+1 Select the top NP / 2 individuals and place them into the middle.

[0051] After initializing num = NP / 2 + 1 in step 9.8, proceed to step 3.3 for sequential execution.

[0052] The method for route planning of vehicle delivery tasks with uncertainty, as described in this invention, is also characterized in that step 3.3 includes:

[0053] Step 3.3.1: Initialize the set of unvisited delivery stations I = {Cus1, Cus2, ..., Cus...} i ...,Cus L};

[0054] Step 3.3.2: Define the current vehicle's sequence number as k, and initialize k = 1. Define the path scheme for the k-th vehicle as R. k and initialize

[0055] Step 3.3.3: Initialize R k ={Depot};

[0056] Step 3.3.4: Initialize i = 1;

[0057] Step 3.3.5: Select the i-th delivery station Cus in I. i Determine the path scheme R to the k-th vehicle. k Insert Cus i Check if the maximum load capacity Q of the vehicle is exceeded; if not, proceed to step 3.2.6; if the maximum load capacity Q is exceeded, then... Assign to After assigning k+1 to k, return to step 3.2.3 and execute sequentially;

[0058] Step 3.3.6: Determine arrival at Cus i Arrival time arr i Has the deadline been exceeded? i If the deadline has not passed, then Cus will be... i Remove path scheme R from set I and add it to set R. k At the end of the list, specify the next delivery station visited by the k-th vehicle, and update set I accordingly; otherwise, do not add Cus. i ;

[0059] Step 3.3.7: Determine if the updated set I is empty. If it is empty, output the num-th individual. Otherwise, return to step 3.3.4 to select from the updated set I.

[0060] Step 4.2 includes:

[0061] Step 4.2.1, in Select two individuals and As the parent generation, among them, Represents the u-th individual The kth u Route planning scheme for delivery vehicles to visit distribution stations. Represents the vth individual The kth v Route planning scheme for delivery vehicles to visit distribution stations;

[0062] Step 4.2.2: Generate a random value p1. If p1 > A, then... Assigned to the u-th child individual Will Assigning value to the v-th child individual Then proceed to step 4.3; otherwise, proceed to step 4.2.3; where A is the crossover probability, p1∈[0,1];

[0063] Step 4.2.3, from Randomly select the kth... u Route planning scheme And split into two sub-path segments and from Randomly select the kth... v Route planning scheme And split into two sub-path segments and

[0064] Step 4.2.4, and Replacement body after splicing In Thus, the u-th temporary individual is obtained. Will and Replace after splicing In Thus, the vth temporary individual is obtained.

[0065] Step 4.2.5, Delete and The system identifies repeatedly visited delivery stations and adds any missing delivery stations to the feasible locations of each individual, thus obtaining the repaired u-th child individual. and the vth offspring individual

[0066] The present invention provides an electronic device, including a memory and a processor, wherein the memory is used to store a program that supports the processor in executing the route planning method, and the processor is configured to execute the program stored in the memory.

[0067] The present invention discloses a computer-readable storage medium on which a computer program is stored, wherein the computer program is executed by a processor to perform the steps of the route planning method.

[0068] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0069] 1. This invention proposes a framework to improve the performance of evolutionary algorithms in searching for robust solutions. Specifically, the uncertainty of urban traffic environments affects whether vehicle delivery tasks can be completed on time. Existing evolutionary algorithms for solving such problems, by randomly matching elite solutions, struggle to generate offspring with better robustness. Unlike existing algorithms, this framework continuously improves delivery solutions under different urban traffic scenarios, rather than directly eliminating them. This proactively enhances the robustness of delivery solutions, thereby quickly finding robust delivery methods.

[0070] 2. This invention employs a game theory strategy to balance the optimality and robustness of evolutionary algorithms in solving this type of problem, maintaining population diversity and allowing the algorithm to search for route planning schemes with better evaluation values. Specifically, in game theory, there are two players, a leader and a follower, who attempt to determine a strategy from the set of strategies to maximize their respective payoffs. The robustness of the delivery scheme population for uncertain vehicle delivery tasks is set as the leader's, and the optimal performance is set as the follower's. The leader focuses on the feasibility of the delivery scheme across multiple scenarios. The follower focuses on the optimal evaluation value of the population. Ultimately, the leader and follower reach a Stackelberg equilibrium, and the delivery scheme population achieves good robustness and optimal performance.

[0071] 3. This invention proposes a robust performance evaluation value for measuring planning schemes in logistics transportation. Specifically, for a delivery scheme, delivery stations have service time requirements. This strategy compares the parent and offspring generations from two perspectives: robustness evaluation value and objective function convergence. It prioritizes solutions with high robustness and good convergence, thus achieving a balance between robustness and convergence. This ensures that the final route planning scheme for uncertain vehicle delivery tasks in urban traffic environments can guarantee both a short total transportation distance and the ability to complete uncertain vehicle delivery tasks in various scenarios. Attached Figure Description

[0072] Figure 1 This is a flowchart of the algorithm of the present invention;

[0073] Figure 2a This is an example diagram illustrating the initialization of the locations of various sites and warehouses in this invention;

[0074] Figure 2b This is an example diagram of the initial delivery scheme of the present invention;

[0075] Figure 3 This is an example diagram of the repair operator of the present invention;

[0076] Figure 4This is an example diagram of the crossover operator of the present invention;

[0077] Figure 5a This is an example diagram of the redistribution operator in the mutation operator of the present invention;

[0078] Figure 5b This is an example diagram of the exchange operator in the mutation operator of the present invention. Detailed Implementation

[0079] This embodiment presents a game theory-based route planning method for vehicle delivery tasks with uncertain travel times. It aims to solve route planning problems in uncertain transportation environments, such as road construction and severe weather causing changes in travel time, and under conditions where travel time is uncertain and delivery stations have strict time requirements. By applying game theory strategies, while maintaining population diversity, it balances the robustness and optimality of the population, improving the robustness of the optimal solution and thus obtaining a robust optimal solution that performs well in various scenarios. Specifically, this route planning method is applied to a system consisting of a warehouse (Depot), K delivery vehicles, and L customer stations V = {Cus1, Cus2, ..., Cus...}. i ...,Cus L In the urban transportation environment constituted by}, Cus i Let i represent the i-th delivery station, i∈[1,L]; let the i-th delivery station Cus i and the j-th delivery station Cus j Let the route between them be (i, j), and let the distance between the routes be c. ij ;

[0080] Define a scenario where traffic changes cause changes in travel time as Scene = {sc1, sc2, ..., sc...} p , ...,sc P}, sc p Let p represent the p-th scene, p∈[1,P]; P represents the total number of scenes, and sc0 represents the original scene that has not been disturbed;

[0081] Define arr i For the transport vehicle to reach the i-th delivery station Cus i Arrival time, lb i For the i-th delivery station Cus i Service time window limit, ub i For the i-th delivery station Cus i Service window deadline;

[0082] Let the maximum load capacity of each transport vehicle be Q;

[0083] like Figure 1 As shown, this route planning method proceeds according to the following steps:

[0084] Step 1: Define the number of times the current scene switches, evngen, and initialize evngen = 0; define the current iteration number as gen;

[0085] Step 2: If evngen = 0, set the current scene for the evnth transition to scene = sc0; where sc0 represents the initial scene. Otherwise, start from Scene = {sc1, sc2, ..., sc...} p , ...,sc P Select a scene from the list as the current scene for the evngen-th transition;

[0086] Initialize gen = 1;

[0087] Step 3: Randomly generate the genth generation population in the current scene. And on All individuals are evaluated;

[0088] Step 3.1: Define the gen-th generation archive population in the current scene as... and initialize Empty;

[0089] Let the gen generation population It contains NP individuals, where the num-th individual is denoted as And the num-th individual K represents the current scene. num Route planning schemes for transport vehicles to visit L delivery stations; and K num ≤K;

[0090] Step 3.2: Initialize num = 1;

[0091] Step 3.3: Randomly generate the gen-th generation population in the current scene. The num-th individual

[0092] Step 3.3.1: Initialize the set of unvisited delivery stations I = {Cus1, Cus2, ..., Cus...} i ...,Cus L};

[0093] Step 3.3.2: Define the current vehicle's sequence number as k, and initialize k = 1. Define the path scheme for the k-th vehicle as R. k and initialize

[0094] Step 3.3.3: Initialize R k ={Depot};

[0095] Step 3.3.4: Initialize i = 1;

[0096] Step 3.3.5: Select the i-th delivery station Cus in I. i Determine the path scheme R to the k-th vehicle. k Insert Cus i Check if the maximum load capacity Q of the vehicle is exceeded; if not, proceed to step 3.2.6; if the maximum load capacity Q is exceeded, then... Assign to After assigning k+1 to k, return to step 3.2.3 and execute sequentially;

[0097] Step 3.3.6: Determine arrival at Cus i Arrival time arr i Has the deadline been exceeded? i If the deadline has not passed, then Cus will be... i Remove path scheme R from set I and add it to set R. k At the end of the list, specify the next delivery station visited by the k-th vehicle, and update set I accordingly; otherwise, do not add Cus. i ;

[0098] Step 3.3.7: Determine if the updated set I is empty. If it is empty, output the num-th individual. Otherwise, return to step 3.3.4 to select from the updated set I;

[0099] like Figure 2a As shown, the access sequence is I = (6, 3, 4, 1, 2, 5, 8, 7). Figure 2b Subpaths 6, 3, 4 and 1, 2, 5, 8 are generated sequentially. Inserting 7 does not satisfy the capacity constraint and time window constraint of the second subpath. Determine whether subpath 1 can be inserted. If it can be inserted, then insert it. If it cannot be inserted, create a new subpath and place it in the subpath.

[0100] Step 3.4: Use equation (1) to process the num-th individual. The evaluation is performed to obtain the num-th travel distance evaluation value in the gen-th generation.

[0101]

[0102] In equation (1), This represents the i-th delivery station Cusi and the j-th delivery station Cus in the current scene. j The distance between them;

[0103] Step 3.5: Use equation (2) to process the num-th individual. An evaluation was conducted to obtain the vehicle count evaluation value.

[0104]

[0105] Step 3.6: After assigning num+1 to num, check if num>NP is true. If true, it means that the population of generation gen has been obtained. Otherwise, return to step 1.2 and execute sequentially;

[0106] Step 4: Based on the two evaluation values, evaluate the population of generation gen. Perform crossover and mutation operations to generate the subpopulation of generation gen.

[0107] Step 4.1, based on A binary tournament selection method is used to generate a mating pool of size NP for the genth generation.

[0108] Step 4.2: Generate two offspring individuals using the crossover operator:

[0109] Step 4.2.1, in Select two individuals and As the parent generation, among them, Represents the u-th individual The kth u Route planning scheme for delivery vehicles to visit distribution stations. Represents the vth individual The kth v Route planning scheme for delivery vehicles to visit distribution stations;

[0110] Step 4.2.2: Generate a random value p1. If p1 > A, then... Assigned to the u-th child individual Will Assigning value to the v-th child individual Then proceed to step 4.3; otherwise, proceed to step 4.2.3; where A is the crossover probability, p1∈[0,1];

[0111] Step 4.2.3, from Randomly select the kth... u Route planning scheme And split into two sub-path segments and from Randomly select the kth... vRoute planning scheme And split into two sub-path segments and

[0112] Step 4.2.4, and Replacement body after splicing In Thus, the u-th temporary individual is obtained. Will and Replace after splicing In Thus, the vth temporary individual is obtained.

[0113] like Figure 4 As shown, two child paths from two parent paths are randomly selected: R1 = (3, 4, 7) for parent 1 and R2 = (8, 6, 3) for parent 2. R1 is split into (3, 4) and 7, and R2 is split into 8 and (6, 3). Then, a new route is generated by combining the first sequence from the first route (3, 4) and the second sequence from the second route (6, 3), and then inserted into parent 1 to replace the selected route (3, 4, 7). Finally, customers in parent path 1 that have already appeared in the new path (such as customer 6) are deleted, while missing customers (such as customer 7) are inserted into feasible positions.

[0114] Step 4.2.5, Delete and The system identifies repeatedly visited delivery stations and adds any missing delivery stations to the feasible locations of each individual, thus obtaining the repaired u-th child individual. and the vth offspring individual

[0115] like Figure 3 As shown, the repair strategy examines delivery stations within an individual that do not meet the time window requirement, such as stations 3 and 5. It explores all feasible insertion points and inserts stations 3 and 5 into the shortest route. If no insertion point meets the time window requirement, the station is inserted into a new route, such as delivery station 3.

[0116] Step 4.3, for and Perform mutation operations;

[0117] Generate a random value p2. If p2 > B, then directly... and As Two offspring individuals; otherwise, use the mutation operator to respectively target the two offspring individuals. and The process was performed, resulting in two mutated offspring individuals that were then placed into... In the equation, B is the mutation probability, and p2∈[0,1];

[0118] Step 4.4: Following the process in Steps 4.2-4.3, obtain NP offspring individuals and form the offspring population of the gen generation.

[0119] like Figure 5a As shown, a sub-path R = (8, 6, 3) is selected, and a reassignment trial is performed on all delivery stations of R. The result shows that only station 3 can be inserted into the sub-path (4, 7), so this operation is performed. Figure 5b The swap operation first selects two sequences from the two selected routes, such as (1, 2) and (3) shown in the figure, and then tries to insert these sequences into other routes in order to reduce the total travel distance while satisfying all constraints.

[0120] Step 5 and Merging yields the merged population of generation gen. And based on environmental selection strategies, the populations are merged from generation gen. Select NP individuals and form the gen+1 generation population.

[0121] Step Six, from and archive population Obtain the save population of generation+1

[0122] Step 6.1: Determine the archived population Is it empty? If it is empty, then... All individuals are stored in the archive population. Otherwise, proceed to step 6.2;

[0123] Step 6.2, and Merging yields the merged population of generation gen. Based on equation (3) After robustness evaluation of all individuals, the evaluation value Pot for all individuals is obtained;

[0124]

[0125] In equation (3), express The robustness evaluation value of the t-th individual; where t≤2NP;

[0126] Step 6.3, according to The evaluation value corresponding to each individual in the data is... The individuals in the population are sorted in ascending order, and the first NP individuals are selected to form the archived population of generation (gen+1). gen+1 ;

[0127] Step 7: After assigning gen+1 to gen, check if gen has reached the maximum number of iterations. max If yes, proceed to step eight; otherwise, return to step two and execute sequentially.

[0128] Step 8: After assigning evngen+1 to evngen, check if evngen has reached the maximum number of scene switching attempts. max If so, then stop iterating and obtain the final archive population under the current scene. Find the optimal route planning scheme for each of the NP individuals, and then... The NP route planning schemes of the population are sorted according to non-dominated individuals to determine the non-dominated individuals, and the route planning schemes corresponding to all non-dominated individuals are selected as the final scheme output; otherwise, proceed to step nine.

[0129] Step 9: Based on the archived population The scenario where the individual's performance confirms the switch;

[0130] Step 9.1: Define the maximum number of individuals that violate the constraint. max and initialize violent max =0;

[0131] Step 9.2: Initialize p = 1;

[0132] Step 9.3: Use equation (4) to determine the p-th scene sc p Below, archived population The number of infeasible individuals

[0133]

[0134] In equation (4), Represents the p-th scene sc p Down The number of infeasible individuals in the array, Γ represents a Boolean function; if arr i >ub i Then Γ(arr) i -ub i ) = 1; otherwise, Γ(arr) = 1. i -ub i ) = 0; express The t-th individual in;

[0135] Step 9.4, Judgment Is it greater than violent? max If so, Assigned to violent max , sc p Use it as the current scene; otherwise, proceed to step 9.5.

[0136] Step 9.5: After assigning p+1 to p, check if p≥P is true. If yes, proceed to step 9.6; otherwise, return to step 9.3 and execute sequentially.

[0137] Step 9.6: Initialize gen = 1.

[0138] Step 9.7: Based on the evaluation value corresponding to each individual, from Arc gen+1 Select the top NP / 2 individuals and place them into the middle.

[0139] After initializing num = NP / 2 + 1 in step 9.8, proceed to step 3.3 for sequential execution.

[0140] In this embodiment, an electronic device includes a memory and a processor. The memory stores a program that supports the processor in executing the above-described method, and the processor is configured to execute the program stored in the memory.

[0141] In this embodiment, a computer-readable storage medium stores a computer program, which is executed by a processor to perform the steps of the above method.

Claims

1. A route planning method for vehicle delivery tasks with uncertainty, characterized in that: Applied to a warehouse , transport vehicles and Customer sites In the urban transportation environment that it constitutes, Indicates the first One delivery station Order No. delivery stations and the delivery stations The route between them is denoted as ,route The distance is denoted as ; Define the scenario where traffic changes lead to changes in travel time as , Indicates the first One scenario, ; Indicates the total number of scenes. This represents the original, undisturbed scene. definition For transport vehicles to the first Geda delivery station Arrival time, For the first delivery stations The upper limit of the service time window, For the first delivery stations Service window deadline; Let the maximum loading capacity of each transport vehicle be ; The route planning method is carried out according to the following steps: Step 1: Define the number of times the current scene will switch. and initialize Define the current iteration number as ; Step 2, if , order the The current scene after the switch ;in, Indicates the initial scene; otherwise, from Choose one scene as the first The current scene after the switch ; initialization ; Step 3, in the current scenario Randomly generated in the middle Generation population and to All individuals are evaluated; Step 3.1: Define the current scene The first in The archived population is and initialize Empty; Order No. Generation population Include Individuals, of which the first Each individual is recorded as And the first individual Represents the current scene middle A transport vehicle visited Route planning schemes for each delivery station; and ; Step 3.2, Initialization ; Step 3.3: Randomly generate the current scene. The Middle Generation population The first in individual ; Step 3.4: Use equation (1) to process the first... individual An evaluation was conducted, and the result was obtained. The generation Individual driving distance evaluation value : (1) In equation (1), Indicates the current scene The Middle delivery stations and the j-th delivery station The distance between them; Step 3.5, using equation (2) to process the first... individual An evaluation was conducted to obtain the vehicle count evaluation value. : (2) Step 3.6, Assign to Then, make a judgment Whether it is true or false; if it is true, then it means that the first... Generation population Otherwise, return to step 1.2 and execute sequentially. Step 4: Based on the two evaluation values, evaluate the... Generation population Perform crossover and mutation operations to generate the first... Subpopulation of generations : Step 4.1, based on A binary tournament selection method is applied to generate a pool of size . The Mating pool ; Step 4.2, from Select the uth individual and the vth individual Two offspring individuals are used, and after repairing them using the crossover operator, the repaired u-th offspring individual is obtained. and the vth offspring individual ; Step 4.3, for and Perform mutation operations; Generate a random value ,like Then directly and As Two offspring individuals; otherwise, use the mutation operator to respectively target the two offspring individuals. and The process was performed, resulting in two mutated offspring individuals that were then placed into... In the equation, B represents the mutation probability. ; Step 4.4: Following the process of steps 4.2-4.3, obtain... Each offspring individual constitutes the first Offspring population ; Step 5 and Merge to obtain the first Merging populations And based on the environmental selection strategy, from the first Merging populations Select Individuals and form the first Generation population ; Step Six, from and archive population Get the Archived population : Step 6.1: Determine the archived population Is it empty? If it is empty, then... All individuals are stored in the archive population. Otherwise, proceed to step 6.2; Step 6.2, and Merge to obtain the first Merging populations According to equation (3), the combination After robustness evaluation of all individuals, the evaluation values ​​for all individuals are obtained. ; (3) In equation (3), express The robustness evaluation value of the t-th individual; where, ; Step 6.3, according to The evaluation value corresponding to each individual in the data is... The individuals in the list are sorted in ascending order to select the top ones. Individuals and form the first Archived population ; Step 7: Assign to Then, make a judgment Has the maximum number of iterations been reached? If yes, proceed to step eight; otherwise, return to step two and execute sequentially. Step 8: Assign to Then, make a judgment Has the maximum number of scene switching attempts been reached? If so, stop iterating and obtain the current scene. The final save population of The optimal route planning scheme corresponding to each individual, and for population The non-dominated individuals are determined by sorting the route planning schemes according to their non-dominated order. The route planning schemes corresponding to all non-dominated individuals are then selected as the final output scheme. Otherwise, proceed to step nine. Step 9: Based on the archived population The scenario where the individual's performance confirms the switch; Step 9.1: Define the maximum number of individuals that violate the constraints. and initialize =0; Step 9.2, Initialize p ; Step 9.3: Use equation (4) to determine the p-th scenario. Below, archived population The number of infeasible individuals ; (4) In equation (4), Represents the p-th scene Down The number of infeasible individuals in the area Represent a Boolean function; if ,but =1; otherwise, =0; express The t-th individual in; Step 9.4, Judgment Is it greater than If so, Assign to ,Will As the current scenario Otherwise, proceed to step 9.5; Step 9.5, Assign to Then, make a judgment If the condition is met, proceed to step 9.6; otherwise, return to step 9.3 and execute sequentially. Step 9.6, Initialization , ; Step 9.7: Based on the evaluation value corresponding to each individual, from... Select the top NP / 2 individuals and place them into the middle. ; After initializing num = NP / 2 + 1 in step 9.8, proceed to step 3.3 for sequential execution.

2. The route planning method for vehicle delivery tasks with uncertainties according to claim 1, characterized in that, Step 3.3 includes: Step 3.3.1: Initialize the set of unvisited delivery stations. ; Step 3.3.2: Define the current vehicle's sequence number as k and initialize... , define the first The vehicle's route plan is as follows and initialize ; Step 3.3.3, Initialization ; Step 3.3.4, Initialization ; Step 3.3.5, in Select the i-th delivery station , judge towards the first Vehicle routing Insert Does it exceed the vehicle's maximum load capacity? If the load is not exceeded, proceed to step 3.2.6; if the maximum load Q is exceeded, then... Assign to ,Will Assign to Then, return to step 3.2.3 and execute sequentially; Step 3.3.6: Determine arrival Arrival time Has the deadline passed? If the deadline has not passed, then From the set Delete and add path schemes At the end of as the first The vehicle visits the next delivery station, and the collection is updated accordingly. Otherwise, do not add. ; Step 3.3.7: Determine the set to be updated. Is it empty? If it is empty, output the first... individual Otherwise, return to step 3.3.4 from the updated set. Choose from the options provided.

3. The route planning method for vehicle delivery tasks with uncertainties according to claim 2, characterized in that, Step 4.2 includes: Step 4.2.1, in Select two individuals and As the parent generation, among them, Represents the u-th individual The Middle Route planning scheme for delivery vehicles to visit distribution stations. Represents the vth individual The Middle Route planning scheme for delivery vehicles to visit distribution stations; Step 4.2.2: Generate a random value ,like Then Assigned to the u-th child individual ,Will Assigning value to the v-th child individual If A is the crossover probability, proceed to step 4.3; otherwise, proceed to step 4.2.3; where A is the crossover probability. ; Step 4.2.3, from Randomly select the first Route planning scheme And split into two sub-path segments and ;from Randomly select the first Route planning scheme And split into two sub-path segments and ; Step 4.2.4, and Replacement body after splicing Thus, the u-th temporary individual is obtained. ;Will and Replace after splicing In Thus, the vth temporary individual is obtained. ; Step 4.2.5, Delete and The system identifies repeatedly visited delivery stations and adds any missing delivery stations to the feasible locations of each individual, thus obtaining the repaired u-th child individual. and the vth offspring individual .

4. An electronic device, comprising a memory and a processor, characterized in that, The memory is used to store a program that supports the processor in executing any of the route planning methods of claims 1-3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium storing a computer program thereon, characterized in that, The computer program, when run by a processor, performs the steps of any of the route planning methods described in claims 1-3.

Images