Time-varying green vehicle multi-target path planning method based on iterative local search

Abstract

The invention discloses a time-varying green vehicle multi-target path planning method based on iterative local search, and the method comprises the steps: 1, building a vehicle logistics-oriented time-varying green vehicle multi-target path planning model for the time-varying path planning of a fuel vehicle, and enabling the model to be used for describing the correlation condition of time-varying speed, vehicle load and fuel consumption; 2, constructing a multi-target time-varying green vehicle path planning model in combination with a time-varying driving time model and a load-speed-distance coupled fuel consumption model; and 3, solving a path planning model by using a heuristic algorithm based on iterative local search to obtain a path planning scheme. According to the method, a distribution path scheme of time-varying speed characteristics and multi-target tradeoff can be optimized, and the problem of collaborative optimization of vehicle paths and fuel consumption in a time-varying traffic environment can be solved, so that the total distance, the transportation time and the fuel consumption can be reduced, and emission is reduced.

Description

Technical Field

[0001] This invention belongs to the field of logistics route optimization and green transportation technology for fuel vehicles, specifically a time-varying green vehicle multi-objective path planning method based on iterative local search. Background Technology

[0002] Compared to vehicle routing under the traditional constant speed assumption, fuel-powered vehicles in dynamic traffic environments incur additional time and fuel costs due to time-varying speeds, and changes in vehicle load further exacerbate fuel consumption. How to rationally plan delivery routes for fuel-powered vehicles based on time-varying traffic characteristics and load dynamics, reducing total transportation distance, time, and fuel consumption, while simultaneously reducing CO2 emissions, is a pressing issue in the logistics and distribution field.

[0003] Most existing methods either optimize only a single objective or integrate multiple objectives into a single objective for path finding, ignoring the coupled impact of time-varying speed and load on fuel consumption, and failing to fully consider the conflicts between objectives, thus failing to reflect the dynamic problems of fuel vehicles in actual delivery. Summary of the Invention

[0004] This invention addresses the shortcomings of existing technologies by proposing a time-varying green vehicle multi-objective path planning method based on iterative local search. This method aims to incorporate multi-objective trade-offs into the path planning problem for fuel-powered vehicles, taking into full account the coupling relationship between time-varying speed, vehicle load changes, and fuel consumption. The goal is to establish a delivery route decision model based on time-varying speed characteristics and multi-objective trade-offs, thereby simultaneously achieving a multi-objective trade-off between distance, transportation time, and fuel consumption. This approach scientifically addresses the logistics and delivery needs in dynamic traffic environments and improves both efficiency and environmental friendliness in logistics and delivery.

[0005] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: The present invention provides a time-varying multi-objective path planning method for green vehicles based on iterative local search, characterized in that it is applied to a delivery service network consisting of one distribution center, N customer nodes, and K fuel vehicles with a capacity of Q; in the delivery service network, the distribution center is denoted as... Let the set of customer nodes consisting of N customers be denoted as . ,make This represents the node of the nth customer, where... n=1,2,...,N N Let B represent the total number of client nodes; let the set of arcs connecting N client nodes be B = {( , )| , =1,2,...,N, };in,( , ) indicates the first Nodes of individual customers With the Nodes of individual customers The arc between; From the distribution center and client node set Together they form the node set V={ }∪ ,make Let c be the c-th node, where c = 0, 1, 2, ..., N+1, where N+1 represents the total number of nodes; let the arc set A = {(i,j)|i,j} connecting the N+1 nodes. 0,1,2,...,N+1,i j}, (i,j) represents the j-th Nodes With the Nodes The arc between; Let the set of gasoline-powered vehicles be H= ,in, This represents the kth fuel-powered vehicle; This indicates the total number of gasoline-powered vehicles; Let delivery time X = ,in, This represents the m-th time period; M represents the total number of time periods. Each vehicle travels at the same speed during the same period, but at different speeds during different periods. The time-varying green vehicle multi-objective path planning method is performed according to the following steps: Step 1: Construct a time-varying green vehicle multi-objective path planning model for vehicle logistics; Step 2: Define the set of globally optimal Pareto path solutions as pa, and initialize pa as an empty set; define the number of local iterations W. max Define and initialize the counter value num=0; define and initialize the number of optimizations D=0; A heuristic algorithm is used to solve a time-varying multi-objective path planning model for green vehicles, generating an optimal path planning scheme consisting of K paths, which is then added to pa. Each path involves a fuel-powered vehicle traveling from a distribution center... After departing and passing through different customer nodes, it finally returns to the distribution center. The formed; Step 3: Construct a refined set of local search operators O and destruction-repair operators. : Step 4: Based on the set of refined local search operators , , For any optimal path planning scheme in pa, perform the Dth multi-objective optimization to obtain the local optimal path planning scheme under the Dth optimization. Step 5: Destruction-Repair Operator Based on Random Removal and Greedy Insertion By performing a fine-grained perturbation operation on the locally optimal path planning scheme under the Dth optimization, we obtain the repair path planning scheme under the Dth optimization: Step 6: Calculate the three objective values ​​for each optimal path planning scheme in pa using equations (1), (2), and (3), and then use the Pareto dominance relation to evaluate the dominance relationship between the repair path planning scheme under the Dth optimization and each optimal path planning scheme in pa: If the repair path planning scheme under the Dth optimization is not dominated by any optimal path scheme in pa, then add the repair path planning scheme under the Dth optimization as an optimal path planning scheme to pa, and remove all optimal path planning schemes in pa that are dominated by the repair path planning scheme under the Dth optimization; otherwise, keep pa unchanged. Step 7: Determine if pa has been updated. If it has, clear D to zero and return to step 4 for sequential execution. Otherwise, assign D+1 to D. If D reaches the maximum number of optimizations Dmax, calculate the three objective values ​​of each optimal path planning scheme in pa, and select the optimal path planning scheme with the three smallest objective values ​​as the global optimal path planning scheme. Otherwise, return to step 4 for sequential execution.

[0006] The multi-objective optimization method for time-varying green vehicle paths based on iterative local search described in this invention is also characterized in that step 1 includes: Step 1.1: Construct the objective function of the time-varying green vehicle multi-objective path planning model based on equations (1)-(3): z1=Min (1) z2=Min (2) z3=Min (3) In equations (1)-(3), z1 represents the total distance target, z2 represents the total time target, and z3 represents the total energy consumption target. For the i-th node To the j-th node The distance between arcs (i,j) For the vehicle starting from the i-th node To the j-th node The transport time between arcs (i,j) When the vehicle is loaded, the i-th node To the j-th node Fuel consumption between (i,j) For the kth fuel-powered vehicle Does it pass through arc (i,j)? If so, let... =1, otherwise, let =0; Step 1.2: Construct the constraints of the time-varying green vehicle multi-objective path planning model based on equations (4)-(21): =1, i≠ (4) =1, , ≠j (5) =0, ,k= K (6) = , =1,2,...,N (7) i 0,1,2,...,N+1, i≠ (8) i 0,1,2,...,N+1, ,k= K (9) =1, k= K (10) =1, k= K (11) =0, k= K (12) {0,1} (13) (1 ), (14) , (15) = / , , (16) =( b i )+( ( b i ) ) / , , +1 (17) = / , =0, , (18) b i +(( ( b i ) ) / =0, , +1 (19) = (1+(p / L) (20) = / (twenty one) In equations (4)-(21), For the kth fuel-powered vehicle From the i-th node To the j-th node The load weight, For the kth fuel-powered vehicle From the Customer Nodes To the j-th node The load weight, For the kth fuel-powered vehicle From the j-th node To the Customer Nodes The load weight, For the kth fuel-powered vehicle From the Customer Nodes to the distribution center The load weight, For the kth fuel-powered vehicle Has the i-th node been visited? To the Customer Nodes Between arcs (i, ), For the kth fuel-powered vehicle Has it been through the first Customer Nodes To the Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has the i-th node been visited? To the Nodes Between arcs (i, ), For the kth fuel-powered vehicle Has it been through the first Nodes To the Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has it been through the first Customer Nodes to the distribution center Between arcs ( , ), The kth fuel-powered vehicle Did it go through a distribution center? To the Customer Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has it been through the first Nodes to the distribution center Between arcs ( , ), For the first Customer Nodes Customer demand for goods, For the first Customer Nodes Customer demand for goods, For the first Nodes The customer's cargo demand is given by Q, where Q is the maximum load capacity of a single vehicle. For the vehicle starting from the i-th node Departure time For the vehicle in the Customer Nodes Arrival time, For the vehicle starting from the i-th node To the Customer Nodes Between arcs (i, The transportation time, The vehicle starts from the i-th node to the distribution center Between arcs (i, The transportation time, The maximum time limit for vehicle operation. For the m-th time period From the i-th node To the j-th node average driving speed For the (m+1)th time period From the i-th node To the j-th node average driving speed For the m-th time period The time limit, For the m-th time period Vehicles inside the i-th node To the j-th node fuel consumption per mile For the m-th time period Vehicles inside the i-th node To the j-th node Fuel consumption per unit time The i-th node when the vehicle is unloaded To the j-th node The fuel consumption is p, where p is the percentage increase in fuel consumption due to load, and L is the base value per unit load.

[0007] Furthermore, step 2 includes: Step 2.1: Create a route from the distribution center for each customer node. From its own customer nodes to the distribution center An independent route; Step 2.2: Calculate the first step using equation (22). Customer Nodes Independent route and Static distance savings from merging independent routes This allows for the construction of a list of distance savings values. ; = + - (twenty two) In equation (22), Indicates distribution center To the Customer Nodes The distance between them Indicates distribution center To the Customer Nodes The first generation distance between them Indicates the first Customer Nodes To the Customer Nodes The first-generation distance between them; Step 2.3, for After sorting the static distance savings values ​​in descending order, it iterates through the sorted list of distance savings values ​​to check if any two adjacent customer nodes are distribution centers that are the endpoints in the route planning scheme. The previous customer node and the starting distribution center If the next customer node is the vehicle load, calculate the sum of the customer cargo demand of the two customer nodes as the vehicle load; otherwise, continue to determine the next adjacent customer node. Step 2.4: If the vehicle load satisfies equation (9), then the independent routes of two adjacent customer nodes are merged into one route; otherwise, they are not merged, thus forming a path planning scheme from all merged routes.

[0008] Furthermore, step 3 includes: Step 3.1: Define the set of refined local search operators O = { , , },in, It is the swap operator. It is the shift operator. It is the 2-opt operator; Define the swap operator It involves randomly selecting two different client nodes within a path and swapping their access order to generate a new neighborhood path; Define the shift operator It involves selecting a client node in a path and moving it to another location on the same path or joining another path to generate a new neighborhood path. Define the 2-opt operator It is to connect two adjacent client nodes , The first in Customer Nodes With the other two adjacent client nodes , The first in Customer Nodes After swapping, reverse arrive The sequence of nodes between; where, Indicates the first One customer node; Step 3.2: Define the destruction-repair operator based on random removal and greedy insertion. , It randomly removes some customer nodes from the current path and greedily inserts them into the current path after deletion according to the principle of minimum weighted increment of multi-objectives, so as to generate a new perturbation path scheme.

[0009] 5. The multi-objective optimization method for time-varying green vehicle paths based on iterative local search according to claim 4, characterized in that step 4 includes: Step 4.1: Define W=1, and let the set of locally Pareto optimal path solutions of generation W-1 under the D-th optimization be... It is an empty set; Step 4.2: Randomly select a set of refined local search operators under the Dth optimization. , , An operator in pa performs a neighborhood operation on any optimal path planning scheme in pa, generating a neighborhood path planning scheme; Step 4.3: Execute the process of Step 4.2 U times to generate the Wth generation neighborhood path planning scheme set containing U neighborhood path planning schemes under the Dth optimization. Step 4.4: Determine whether each neighborhood path planning scheme in the W-th generation neighborhood path planning scheme set containing U neighborhood path planning schemes under the D-th optimization satisfies equations (9), (14), and (15). If all are satisfied, retain the corresponding neighborhood path planning scheme; otherwise, delete the corresponding neighborhood path planning scheme, thus obtaining the neighborhood path planning scheme set after the W-th generation selection under the D-th optimization. Let any neighborhood path planning scheme in the W-th generation selection set under the D-th optimization be... ; Step 4.5: Calculate the three objective vectors of each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of the Dth optimization according to Equations (1), (2), and (3); Step 4.6: Based on the three objective vectors of each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of optimization under the Dth optimization, use the Pareto dominance relation to evaluate the relationship between each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of optimization under the Dth optimization and the target path. Dominance relationships of each neighborhood path planning scheme: when Not If any neighborhood path planning scheme dominates, then it will join in At the same time remove All of them were The dominant neighborhood path planning scheme is used to obtain the set of Pareto local optimal path planning schemes in the Wth generation under the Dth optimization. ; Step 4.5, Judgment = Check if the condition is met. If it is met, increment the count value num by 1 and proceed to step 4.6; otherwise, clear the count value num and proceed to step 4.7. Step 4.6; Determine if num is equal to the application. If the total number of times Num_max is reached, proceed to step 4.8; otherwise, proceed to step 4.7. Step 4.7; After assigning W+1 to W, if W=W max This indicates that the Wth value obtained under the Dth optimization is obtained. max Pareto Local Optimum Path Planning Scheme Set If the above is not true, proceed to step 4.8; otherwise, return to step 4.2 and execute sequentially. Step 4.8: From The optimal path planning scheme is selected as the local optimal path planning scheme under the Dth optimization.

[0010] Furthermore, step 5 includes: Step 5.1: Randomly select from the locally optimal path planning schemes under the Dth optimization. After removing N customer nodes, a partial path planning scheme is formed under the Dth optimization, where... Damage rate; Step 5.2, Each removed customer node from the N customer nodes is sequentially inserted into the partial path planning scheme under the Dth optimization. Equations (1) to (3) are used to calculate the increments of the three objective values ​​generated when any removed customer node is inserted into a position satisfying the constraints of equations (9), (14), and (15). The position with the smallest weighted sum of the increments of the three objective values ​​is then selected for insertion, thereby... After inserting N customer nodes into the partial path planning scheme under the Dth optimization, the repair path planning scheme under the Dth optimization is obtained.

[0011] The present invention provides an electronic device, including a memory and a processor, characterized in that the memory is used to store a program that supports the processor in executing the time-varying environment green multi-objective path optimization method for vehicle logistics, and the processor is configured to execute the program stored in the memory.

[0012] The present invention discloses a computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, performs the steps of the claimed method for green multi-objective path optimization in time-varying environments for vehicle logistics.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. To address the problem that existing fuel consumption models in vehicle route planning are overly simplified and fail to reflect real-world operating conditions, this invention proposes a time-varying fuel consumption calculation model. By simultaneously introducing the coupling relationship between travel distance, vehicle load, and time-varying speed during the route planning process, fuel consumption is transformed from a traditional fixed parameter or static estimation method into a function that dynamically adjusts with changes in speed and load over time. This technique effectively overcomes the energy consumption estimation bias caused by existing models neglecting changes in traffic conditions and load effects, thereby significantly improving the accuracy of fuel consumption assessment and the practical applicability of route decisions.

[0014] 2. To address the common problem in existing green vehicle routing problems of converting multi-objective weighted optimization into a single objective, which leads to the loss of conflicting information, this invention constructs a multi-objective mathematical optimization model for multi-objective time-varying green vehicle routing problems. By simultaneously characterizing and minimizing three conflicting optimization objectives—travel distance, transportation time, and fuel consumption—it avoids the subjectivity and instability caused by manually setting weights. This technical solution can achieve more reasonable coordination and trade-offs among multiple objectives, improving the overall performance of the routing scheme in terms of efficiency, timeliness, and energy saving, thereby enhancing the model's decision robustness in complex urban traffic environments.

[0015] 3. To address the problems of high computational complexity, slow response speed, and inability to meet dynamic decision-making requirements in practical logistics scheduling scenarios, existing vehicle route optimization methods propose an efficient heuristic solution method based on iterative local search. By constructing a multi-neighborhood search structure and an iterative update strategy, computational overhead is significantly reduced while ensuring solution quality. This method can quickly generate high-quality route solutions under limited computational resources in practical application scenarios such as urban delivery, emergency material transportation, and multi-period traffic condition changes. It is suitable for large-scale customer nodes, complex road networks, and time-varying traffic environments. The application of this algorithm effectively shortens the route planning calculation time and improves the vehicle scheduling response speed, thereby enhancing the overall operational efficiency of the logistics system. Furthermore, the method of this invention has good engineering scalability and system compatibility, and can be embedded into existing intelligent logistics scheduling systems or transportation management systems. It provides technical support for enterprises to reduce fuel consumption, shorten transportation time, improve vehicle utilization, and reduce operating costs, demonstrating significant practical application value and promising prospects for widespread adoption. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] In this implementation example, a time-varying green vehicle multi-objective path planning method based on iterative local search is proposed. First, for the fuel vehicle logistics path planning problem in a time-varying environment addressed by this invention, several customized local search operators (including swap, shift, and 2-opt operators) are designed to better adapt to the complex impact of time-varying speed and load changes on path planning. Second, a multi-objective iterative optimization mechanism based on Pareto dominance is proposed. By systematically evaluating and comparing the three objectives of distance, time, and fuel consumption, the search direction is automatically guided, effectively balancing the conflicts between objectives and significantly reducing reliance on prior knowledge or human preferences. Finally, a disruptive-repair perturbation strategy is combined to help the iterative local search algorithm escape locally optimal path solutions and explore a broader and higher-quality path solution space by continuously updating the Pareto optimal path solution set, thereby achieving a synergistic improvement in algorithm performance and path solution stability in dynamic traffic environments. This method continuously optimizes the current path scheme by alternating the execution of local search operators and perturbation operations, expanding the distribution range of non-dominated path schemes, and updating the Pareto front after each iteration. This allows the model to systematically approximate the optimal set of paths with multi-objective trade-offs, ultimately improving the overall quality of the path scheme. This helps logistics companies scientifically reduce overall costs and environmental impact by adapting to time-sensitive speeds and balancing load distribution. It is key to improving logistics efficiency and environmental friendliness and has significant practical implications.

[0018] Specifically, this method is applied to a delivery service network consisting of one distribution center, N customer nodes, and K fuel-powered vehicles with a capacity of Q; in the delivery service network, the distribution center is denoted as... Let the set of customer nodes consisting of N customers be denoted as . ,make This represents the node of the nth customer, where... n=1,2,...,N N Let B represent the total number of client nodes; let the set of arcs connecting N client nodes be B = {( , )| , =1,2,...,N, };in,( , ) indicates the first Nodes of individual customers With the Nodes of individual customers The arc between; From the distribution center and client node set Together they form the node set V={ }∪ ,make Let c be the c-th node, where c = 0, 1, 2, ..., N+1, where N+1 represents the total number of nodes; let the arc set A = {(i,j)|i,j} connecting the N+1 nodes. 0,1,2,...,N+1,i j}, (i,j) represents the j-th Nodes With the Nodes The arc between; Let the set of gasoline-powered vehicles be H= ,in, This represents the kth fuel-powered vehicle; This indicates the total number of gasoline-powered vehicles; Let delivery time X = ,in, This represents the m-th time period; M represents the total number of time periods. Each vehicle travels at the same speed during the same period, but at different speeds during different periods.

[0019] like Figure 1 As shown, the time-varying green vehicle multi-objective path planning method is performed according to the following steps: Step 1: Construct a time-varying green vehicle multi-objective path planning model for vehicle logistics: Step 1.1: Construct the objective function of the time-varying green vehicle multi-objective path planning model based on equations (1)-(3): z1=Min (1) z2=Min (2) z3=Min (3) In equations (1)-(3), z1 represents the total distance target, z2 represents the total time target, and z3 represents the total energy consumption target. For the i-th node To the j-th node The distance between arcs (i,j) For the vehicle starting from the i-th node To the j-th node The transport time between arcs (i,j) When the vehicle is loaded, the i-th node To the j-th node Fuel consumption between (i,j) For the kth fuel-powered vehicle Does it pass through arc (i,j)? If so, let... =1, otherwise, let =0.

[0020] Step 1.2: Construct the constraints of the time-varying green vehicle multi-objective path planning model based on equations (4)-(21): =1, i≠ (4) =1, , ≠j (5) =0, ,k= K (6) = , =1,2,...,N (7) i 0,1,2,...,N+1, i≠ (8) i 0,1,2,...,N+1, ,k= K (9) =1, k= K (10) =1, k= K (11) =0, k= K (12) {0,1} (13) (1 ), (14) , (15) = / , , (16) =( b i )+( ( b i ) ) / , , +1 (17) = / , =0, , (18) b i +(( ( b i ) ) / =0, , +1 (19) = (1+(p / L) (20) = / (twenty one) In equations (4)-(21), For the kth fuel-powered vehicle From the i-th node To the j-th node The load weight, For the kth fuel-powered vehicle From the Customer Nodes To the j-th node The load weight, For the kth fuel-powered vehicle From the j-th node To the Customer Nodes The load weight, For the kth fuel-powered vehicle From the Customer Nodes to the distribution center The load weight, For the kth fuel-powered vehicle Has the i-th node been visited? To the Customer Nodes Between arcs (i, ), For the kth fuel-powered vehicle Has it been through the first Customer Nodes To the Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has the i-th node been visited? To the Nodes Between arcs (i, ), For the kth fuel-powered vehicle Has it been through the first Nodes To the Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has it been through the first Customer Nodes to the distribution center Between arcs ( , ), The kth fuel-powered vehicle Did it go through a distribution center? To the Customer Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has it been through the first Nodes to the distribution center Between arcs ( , ), For the first Customer Nodes Customer demand for goods, For the first Customer Nodes Customer demand for goods, For the first Nodes The customer's cargo demand is given by Q, where Q is the maximum load capacity of a single vehicle. For the vehicle starting from the i-th node Departure time, For the vehicle in the Customer Nodes Arrival time, For the vehicle starting from the i-th node To the Customer Nodes Between arcs (i, The transportation time, The vehicle starts from the i-th node to the distribution center Between arcs (i, The transportation time, The maximum time limit for vehicle operation. For the m-th time period From the i-th node To the j-th node average driving speed For the (m+1)th time period From the i-th node To the j-th node average driving speed For the m-th time period The time limit, For the m-th time period Vehicles inside the i-th node To the j-th node fuel consumption per mile For the m-th time period Vehicles inside the i-th node To the j-th node Fuel consumption per unit time The i-th node when the vehicle is unloaded To the j-th node The fuel consumption is p, where p is the percentage increase in fuel consumption due to load, and L is the base value per unit load.

[0021] Step 2: Define the set of globally optimal Pareto path solutions as pa, and initialize pa as an empty set; define the number of local iterations W. max Define and initialize the counter value num=0; define and initialize the number of optimizations D=0; A heuristic algorithm is used to solve a time-varying multi-objective path planning model for green vehicles, generating an optimal path planning scheme consisting of K paths, which is then added to pa. Each path involves a fuel-powered vehicle traveling from a distribution center... After departing and passing through different customer nodes, it finally returns to the distribution center. What is formed.

[0022] Step 2.1: Create a route from the distribution center for each customer node. From its own customer nodes to the distribution center An independent route; Step 2.2: Calculate the first step using equation (22). Customer Nodes Independent route and Static distance savings from merging independent routes This allows for the construction of a list of distance savings values. ; = + - (twenty two) In equation (22), Indicates distribution center To the Customer Nodes The distance between them Indicates distribution center To the Customer Nodes The first generation distance between them Indicates the first Customer Nodes To the Customer Nodes The first-generation distance between them.

[0023] Step 2.3, for After sorting the static distance savings values ​​in descending order, it iterates through the sorted list of distance savings values ​​to check if any two adjacent customer nodes are distribution centers that are the endpoints in the route planning scheme. The previous customer node and the starting distribution center If the next customer node is the vehicle load, calculate the sum of the customer cargo demand of the two customer nodes as the vehicle load; otherwise, continue to determine the next adjacent customer node. Step 2.4: If the vehicle load satisfies equation (9), then the independent routes of two adjacent customer nodes are merged into one route; otherwise, they are not merged, thus forming a path planning scheme from all merged routes.

[0024] Step 3: Construct a refined set of local search operators O and destruction-repair operators. : Step 3.1: Define the set of refined local search operators O = { , , },in, It is the swap operator. It is the shift operator. It is the 2-opt operator; Define the swap operator It involves randomly selecting two different client nodes within a path and swapping their access order to generate a new neighborhood path; Define the shift operator It involves selecting a client node in a path and moving it to another location on the same path or joining another path to generate a new neighborhood path. Define the 2-opt operator It is to connect two adjacent client nodes , The first in Customer Nodes With the other two adjacent client nodes , The first in Customer Nodes After swapping, reverse arrive The sequence of nodes between; where, Indicates the first One customer node.

[0025] Step 3.2: Define the destruction-repair operator based on random removal and greedy insertion. , It randomly removes some customer nodes from the current path and greedily inserts them into the current path after deletion according to the principle of minimum weighted increment of multi-objectives, so as to generate a new perturbation path scheme.

[0026] Step 4: Based on the set of refined local search operators , , For any optimal path planning scheme in pa, perform the Dth multi-objective optimization to obtain the local optimal path planning scheme under the Dth optimization. Step 4.1: Define W=1, and let the set of locally Pareto optimal path solutions of generation W-1 under the D-th optimization be... It is an empty set; Step 4.2: Randomly select a set of refined local search operators under the Dth optimization. , , An operator in pa performs a neighborhood operation on any optimal path planning scheme in pa, generating a neighborhood path planning scheme; Step 4.3: Execute the process of step 4.2 U times to generate the Wth generation neighborhood path planning scheme set containing U neighborhood path planning schemes under the Dth optimization.

[0027] Step 4.4: Determine whether each neighborhood path planning scheme in the W-th generation neighborhood path planning scheme set containing U neighborhood path planning schemes under the D-th optimization satisfies equations (9), (14), and (15). If all are satisfied, retain the corresponding neighborhood path planning scheme; otherwise, delete the corresponding neighborhood path planning scheme, thus obtaining the neighborhood path planning scheme set after the W-th generation selection under the D-th optimization. Let any neighborhood path planning scheme in the W-th generation selection set under the D-th optimization be... .

[0028] Step 4.5: Calculate the three objective vectors of each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of the Dth optimization according to Equations (1), (2), and (3); Step 4.6: Based on the three objective vectors of each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of optimization under the Dth optimization, use the Pareto dominance relation to evaluate the relationship between each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of optimization under the Dth optimization and the target path. Dominance relationships of each neighborhood path planning scheme: When a certain neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of the Dth optimization is selected... Not If any neighborhood path planning scheme dominates, then the corresponding neighborhood path planning scheme in the set of neighborhood path planning schemes selected after the Wth generation of optimization under the Dth optimization will be included. join in At the same time remove All neighboring path planning schemes The dominant neighborhood path planning scheme is used to obtain the set of Pareto local optimal path planning schemes in the Wth generation under the Dth optimization. .

[0029] Step 4.5, Judgment = Check if the condition is met. If it is met, increment the count value num by 1 and proceed to step 4.6; otherwise, clear the count value num and proceed to step 4.7. Step 4.6; Determine if num is equal to the application. If the total number of times Num_max is reached, proceed to step 4.8; otherwise, proceed to step 4.7. Step 4.7; After assigning W+1 to W, if W=W max This indicates that the Wth value obtained under the Dth optimization is obtained. max Pareto Local Optimum Path Planning Scheme Set If the above is not true, proceed to step 4.8; otherwise, return to step 4.2 and execute sequentially. Step 4.8: From The optimal path planning scheme is selected as the local optimal path planning scheme under the Dth optimization.

[0030] Step 5: Destruction-Repair Operator Based on Random Removal and Greedy Insertion By performing a fine-grained perturbation operation on the locally optimal path planning scheme under the Dth optimization, we obtain the repair path planning scheme under the Dth optimization: Step 5.1: Randomly select from the locally optimal path planning schemes under the Dth optimization. After removing N customer nodes, a partial path planning scheme is formed under the Dth optimization, where... Damage rate; Step 5.2, Each removed customer node from the N customer nodes is sequentially inserted into the partial path planning scheme under the Dth optimization. Equations (1) to (3) are used to calculate the increments of the three objective values ​​generated when any removed customer node is inserted into a position satisfying the constraints of equations (9), (14), and (15). The position with the smallest weighted sum of the increments of the three objective values ​​is then selected for insertion, thereby... After inserting N customer nodes into the partial path planning scheme under the Dth optimization, the repair path planning scheme under the Dth optimization is obtained.

[0031] Step 6: Calculate the three objective values ​​for each optimal path planning scheme in pa using equations (1), (2), and (3), and then use the Pareto dominance relation to evaluate the dominance relationship between the repair path planning scheme under the Dth optimization and each optimal path planning scheme in pa: If the repair path planning scheme under the Dth optimization is not dominated by any optimal path scheme in pa, then add the repair path planning scheme under the Dth optimization as an optimal path planning scheme to pa, and remove all optimal path planning schemes in pa that are dominated by the repair path planning scheme under the Dth optimization; otherwise, keep pa unchanged.

[0032] Step 7: Determine if pa has been updated. If it has, clear D to zero and return to step 4 for sequential execution. Otherwise, assign D+1 to D. If D reaches the maximum number of optimizations Dmax, calculate the three objective values ​​of each optimal path planning scheme in pa, and select the optimal path planning scheme with the three smallest objective values ​​as the global optimal path planning scheme. Otherwise, return to step 4 for sequential execution.

[0033] In this embodiment, an electronic device includes a memory and a processor. The memory stores a program that supports the processor in executing the methods described above, and the processor is configured to execute the program stored in the memory.

[0034] 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 time-varying multi-objective path planning method for green vehicles based on iterative local search, characterized in that, This applies to a delivery service network consisting of one distribution center, N customer nodes, and K fuel-powered vehicles with a capacity of Q; in this delivery service network, the distribution center is denoted as... Let the set of customer nodes consisting of N customers be denoted as . ,make This represents the node of the nth customer, where... n=1,2,...,N N Let B represent the total number of client nodes; let the set of arcs connecting N client nodes be B = {( , )| , =1,2,...,N, };in,( , ) indicates the first Nodes of individual customers With the Nodes of individual customers The arc between; From the distribution center and client node set Together they form the node set V={ }∪ ,make Let c be the c-th node, where c = 0, 1, 2, ..., N+1, where N+1 represents the total number of nodes; let the arc set A = {(i,j)|i,j} connecting the N+1 nodes. 0,1,2,...,N+1,i j}, (i,j) represents the j-th Nodes With the Nodes The arc between; Let the set of gasoline-powered vehicles be H= ,in, This represents the kth fuel-powered vehicle; This indicates the total number of gasoline-powered vehicles. Let delivery time X = ,in, This represents the m-th time period; M represents the total number of time periods. Each vehicle travels at the same speed during the same period, but at different speeds during different periods. The time-varying green vehicle multi-objective path planning method is performed according to the following steps: Step 1: Construct a time-varying green vehicle multi-objective path planning model for vehicle logistics; Step 2: Define the set of globally optimal Pareto path solutions as pa, and initialize pa as an empty set; define the number of local iterations W. max Define and initialize the counter value num=0; define and initialize the number of optimizations D=0; A heuristic algorithm is used to solve a time-varying multi-objective path planning model for green vehicles, generating an optimal path planning scheme consisting of K paths, which is then added to pa. Each path involves a fuel-powered vehicle traveling from a distribution center... After departing and passing through different customer nodes, it finally returns to the distribution center. The formed; Step 3: Construct a refined set of local search operators O and destruction-repair operators. : Step 4: Based on the set of refined local search operators , , For any optimal path planning scheme in pa, perform the Dth multi-objective optimization to obtain the local optimal path planning scheme under the Dth optimization. Step 5: Destruction-Repair Operator Based on Random Removal and Greedy Insertion By performing a fine-grained perturbation operation on the locally optimal path planning scheme under the Dth optimization, we obtain the repair path planning scheme under the Dth optimization: Step 6: Calculate the three objective values ​​for each optimal path planning scheme in pa using equations (1), (2), and (3), and then use the Pareto dominance relation to evaluate the dominance relationship between the repair path planning scheme under the Dth optimization and each optimal path planning scheme in pa: If the repair path planning scheme under the Dth optimization is not dominated by any optimal path scheme in pa, then add the repair path planning scheme under the Dth optimization as an optimal path planning scheme to pa, and remove all optimal path planning schemes in pa that are dominated by the repair path planning scheme under the Dth optimization; otherwise, keep pa unchanged. Step 7: Determine if pa has been updated. If it has, clear D to zero and return to step 4 for sequential execution. Otherwise, assign D+1 to D. If D reaches the maximum number of optimizations Dmax, calculate the three objective values ​​of each optimal path planning scheme in pa, and select the optimal path planning scheme with the three smallest objective values ​​as the global optimal path planning scheme. Otherwise, return to step 4 for sequential execution.

2. The multi-objective optimization method for time-varying green vehicle paths based on iterative local search according to claim 1, characterized in that, Step 1 includes: Step 1.1: Construct the objective function of the time-varying green vehicle multi-objective path planning model based on equations (1)-(3): z1=Min (1) z2=Min (2) z3=Min (3) In equations (1)-(3), z1 represents the total distance target, z2 represents the total time target, and z3 represents the total energy consumption target. For the i-th node To the j-th node The distance between arcs (i,j) For the vehicle starting from the i-th node To the j-th node The transport time between arcs (i,j) When the vehicle is loaded, the i-th node To the j-th node Fuel consumption between (i,j) For the kth fuel-powered vehicle Does it pass through arc (i,j)? If so, let... =1, otherwise, let =0; Step 1.2: Construct the constraints of the time-varying green vehicle multi-objective path planning model based on equations (4)-(21): =1 ， ,i≠ (4) =1， , ≠j (5) =0, ,k= K (6) = ， =1,2,...,N (7) ,i 0,1,2,...,N+1, ,i≠ (8) ,i 0,1,2,...,N+1, ,k= K (9) =1,k= K (10) =1,k= K (11) =0,k= K (12) {0,1}， (13) (1 )， (14) ， (15) = / ， , (16) =( b i )+( ( b i ) ) / , , +1 (17) = / ， =0， , (18) b i +(( ( b i ) ) / =0, , +1 (19) = (1+(p / L) ) (20) = / (21) In equations (4)-(21), For the kth fuel-powered vehicle From the i-th node To the j-th node The load weight, For the kth fuel-powered vehicle From the Customer Nodes To the j-th node The load weight, For the kth fuel-powered vehicle From the j-th node To the Customer Nodes The load weight, For the kth fuel-powered vehicle From the Customer Nodes to the distribution center The load weight, For the kth fuel-powered vehicle Has the i-th node been visited? To the Customer Nodes Between arcs (i, ), For the kth fuel-powered vehicle Has it been through the first Customer Nodes To the Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has the i-th node been visited? To the Nodes Between arcs (i, ), For the kth fuel-powered vehicle Has it been through the first Nodes To the Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has it been through the first Customer Nodes to the distribution center Between arcs ( , ), The kth fuel-powered vehicle Did it go through a distribution center? To the Customer Nodes Between arcs ( , ), For the kth fuel-powered vehicle Has it been through the first Nodes to the distribution center Between arcs ( , ), For the first Customer Nodes Customer demand for goods, For the first Customer Nodes Customer demand for goods, For the first Nodes The customer's cargo demand is given by Q, where Q is the maximum load capacity of a single vehicle. For the vehicle starting from the i-th node Departure time, For the vehicle in the Customer Nodes Arrival time, For the vehicle starting from the i-th node To the Customer Nodes Between arcs (i, The transportation time, The vehicle starts from the i-th node to the distribution center Between arcs (i, The transportation time, The maximum time limit for vehicle operation. For the m-th time period From the i-th node To the j-th node average driving speed For the (m+1)th time period From the i-th node To the j-th node average driving speed For the m-th time period The time limit, For the m-th time period Vehicles inside the i-th node To the j-th node fuel consumption per mile For the m-th time period Vehicles inside the i-th node To the j-th node Fuel consumption per unit time The i-th node when the vehicle is unloaded To the j-th node The fuel consumption is p, where p is the percentage increase in fuel consumption due to load, and L is the base value per unit load.

3. The multi-objective optimization method for time-varying green vehicle paths based on iterative local search according to claim 2, characterized in that, Step 2 includes: Step 2.1: Create a route from the distribution center for each customer node. From its own customer nodes to the distribution center An independent route; Step 2.2: Calculate the first step using equation (22). Customer Nodes Independent route and Static distance savings from merging independent routes This allows for the construction of a list of distance savings values. ; = + - (22) In equation (22), Indicates distribution center To the Customer Nodes The distance between them Indicates distribution center To the Customer Nodes The first generation distance between them Indicates the first Customer Nodes To the Customer Nodes The first-generation distance between them; Step 2.3, for After sorting the static distance savings values ​​in descending order, it iterates through the sorted list of distance savings values ​​to check if any two adjacent customer nodes are distribution centers that are the endpoints in the route planning scheme. The previous customer node and the starting distribution center If the next customer node is the vehicle load, calculate the sum of the customer cargo demand of the two customer nodes as the vehicle load; otherwise, continue to determine the next adjacent customer node. Step 2.4: If the vehicle load satisfies equation (9), then the independent routes of two adjacent customer nodes are merged into one route; otherwise, they are not merged, thus forming a path planning scheme from all merged routes.

4. The multi-objective optimization method for time-varying green vehicle paths based on iterative local search according to claim 1, characterized in that, Step 3 includes: Step 3.1: Define the set of refined local search operators O = { , , },in, It is the swap operator. It is the shift operator. It is the 2-opt operator; Define the swap operator It involves randomly selecting two different client nodes within a path and swapping their access order to generate a new neighborhood path; Define the shift operator It involves selecting a client node in a path and moving it to another location on the same path or joining another path to generate a new neighborhood path. Define the 2-opt operator It is to connect two adjacent client nodes , The first in Customer Nodes With the other two adjacent client nodes , The first in Customer Nodes After swapping, reverse arrive The sequence of nodes between; where, Indicates the first One customer node; Step 3.2: Define the destruction-repair operator based on random removal and greedy insertion. , It randomly removes some customer nodes from the current path and greedily inserts them into the current path after deletion according to the principle of minimum weighted increment of multi-objectives, so as to generate a new perturbation path scheme.

5. The multi-objective optimization method for time-varying green vehicle paths based on iterative local search according to claim 4, characterized in that, Step 4 includes: Step 4.1: Define W=1, and let the set of locally Pareto optimal path solutions of generation W-1 under the D-th optimization be... It is an empty set; Step 4.2: Randomly select a set of refined local search operators under the Dth optimization. , , An operator in pa performs a neighborhood operation on any optimal path planning scheme in pa, generating a neighborhood path planning scheme; Step 4.3: Execute the process of Step 4.2 U times to generate the Wth generation neighborhood path planning scheme set containing U neighborhood path planning schemes under the Dth optimization. Step 4.4: Determine whether each neighborhood path planning scheme in the W-th generation neighborhood path planning scheme set containing U neighborhood path planning schemes under the D-th optimization satisfies equations (9), (14), and (15). If all are satisfied, retain the corresponding neighborhood path planning scheme; otherwise, delete the corresponding neighborhood path planning scheme, thus obtaining the neighborhood path planning scheme set after the W-th generation selection under the D-th optimization. Let any neighborhood path planning scheme in the W-th generation selection set under the D-th optimization be... ; Step 4.5: Calculate the three objective vectors of each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of the Dth optimization according to Equations (1), (2), and (3); Step 4.6: Based on the three objective vectors of each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of optimization under the Dth optimization, use the Pareto dominance relation to evaluate the relationship between each neighborhood path planning scheme in the neighborhood path planning scheme set after the Wth generation of optimization under the Dth optimization and the target path. Dominance relationships of each neighborhood path planning scheme: when Not If any neighborhood path planning scheme dominates, then it will join in At the same time remove All of them were The dominant neighborhood path planning scheme is used to obtain the set of Pareto local optimal path planning schemes in the Wth generation under the Dth optimization. ; Step 4.5, Judgment = Check if the condition is met. If it is met, increment the count value num by 1 and proceed to step 4.6; otherwise, clear the count value num and proceed to step 4.

7. Step 4.6; Determine if num is equal to the application. If the total number of times Num_max is reached, proceed to step 4.8; otherwise, proceed to step 4.

7. Step 4.7; After assigning W+1 to W, if W=W max This indicates that the Wth value obtained under the Dth optimization is obtained. max Pareto Local Optimum Path Planning Scheme Set If the above is not true, proceed to step 4.8; otherwise, return to step 4.2 and execute sequentially. Step 4.8: From The optimal path planning scheme is selected as the local optimal path planning scheme under the Dth optimization.

6. The multi-objective optimization method for time-varying green vehicle paths based on iterative local search according to claim 5, characterized in that, Step 5 includes: Step 5.1: Randomly select from the locally optimal path planning schemes under the Dth optimization. After removing N customer nodes, a partial path planning scheme is formed under the Dth optimization, where... Damage rate; Step 5.2, Each removed customer node from the N customer nodes is sequentially inserted into the partial path planning scheme under the Dth optimization. Equations (1) to (3) are used to calculate the increments of the three objective values ​​generated when any removed customer node is inserted into a position satisfying the constraints of equations (9), (14), and (15). The position with the smallest weighted sum of the increments of the three objective values ​​is then selected for insertion, thereby... After inserting N customer nodes into the partial path planning scheme under the Dth optimization, the repair path planning scheme under the Dth optimization is obtained.

7. 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 the time-varying environment green multi-objective path optimization method for vehicle logistics as described in any one of claims 1-6, and the processor is configured to execute the program stored in the memory.

8. A computer-readable storage medium storing a computer program thereon, characterized in that, The computer program, when run by the processor, executes the steps of the green multi-objective path optimization method for time-varying environments oriented towards vehicle logistics as described in any one of claims 1-6.

Images