Emergency logistics two-stage path optimization method based on truck-unmanned aerial vehicle cooperation

Abstract

The invention discloses an emergency logistics two-stage path optimization method based on truck-unmanned aerial vehicle cooperation, and relates to the technical field of emergency logistics path optimization, the method is based on a truck-unmanned aerial vehicle cooperation two-stage path optimization algorithm, the method is composed of a self-adaptive K-means clustering algorithm and a variable neighborhood simulated annealing algorithm, the calculation efficiency is high, the solving quality is high, and the method is suitable for large-scale popularization and application. And the method has a relatively high practical value for the emergency logistics transportation optimization problem with relatively strict timeliness requirements. The method has high universality and transportability, can better deal with the scene that the road is interrupted and emergency materials are affected to arrive in time under the emergency situation, can provide reference for the government and decision consultation department to formulate an emergency logistics transportation mode, improves the emergency rescue efficiency, and ensures the credibility of decision.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency logistics path optimization, and in particular to a two-stage emergency logistics path optimization method based on truck-unmanned aerial vehicle collaboration. Background Art

[0002] The main natural disasters in my country throughout the year are floods, typhoons, earthquakes and geological disasters. Droughts, wind and hail, low temperatures and freezing, snowstorms, sandstorms and forest and grassland fires also occur to varying degrees. Although the destructive effects of natural disasters are inevitable, the impact of disasters on people's lives and property can be mitigated through rapid response and well-prepared emergency logistics systems. However, due to the damage of natural disasters to infrastructure and transportation networks, emergency logistics transportation has brought huge challenges. Therefore, how to optimize emergency logistics transportation routes to ensure that emergency materials are accurately and quickly delivered to the disaster-stricken areas is a top priority that needs to be urgently studied in emergency relief work, and it also provides theoretical support for the design of emergency material transportation plans.

[0003] For emergency logistics, the problems that need to be solved can be summarized as: site selection of emergency logistics facilities, route optimization, and site selection-route optimization. First, in terms of emergency logistics facilities, it is divided into deterministic facility site selection, dynamic facility site selection, random facility site selection, and robust facility site selection; second, in terms of emergency logistics route optimization, it is necessary to consider factors such as the time window for material transportation, the transportation method and distance between nodes, and the choice of transportation tools; third, in emergency logistics site selection-route optimization, it is necessary to consider the site selection of emergency warehouses, transportation safety, transportation timeliness, accessibility, and other factors at the same time.

[0004] For the above-mentioned emergency logistics path optimization, precise algorithms and heuristic algorithms are usually used for solving. One is the precise algorithm: the solution quality is good but the solution algorithm is slow and requires more computing resources; the other is the heuristic algorithm: the solution speed is fast but it is easy to fall into the local optimal solution and it is difficult to ensure the solution quality. However, it is difficult to solve it using only a single heuristic algorithm. As the scale of the problem becomes larger and larger, it is difficult to obtain a satisfactory solution in a short time using a precise algorithm. In addition, in the process of emergency response, it is often required to complete the transportation of emergency materials in the shortest possible time. Therefore, there is an urgent need for a transportation path optimization method that can quickly solve and ensure the solution quality.

[0005] To address the above problems, an improved two-stage path optimization method for emergency logistics based on truck-UAV collaboration is designed. Summary of the invention

[0006] The purpose of the present invention is to provide a two-stage path optimization method for emergency logistics based on truck-UAV collaboration to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The two-stage path optimization method for emergency logistics based on truck-UAV collaboration includes the following steps:

[0009] Step 1: The nodes in the emergency logistics network are composed of distribution centers and demand points. For the nodes in the emergency logistics network, an adaptive K-means clustering method is designed. The inertia_ index of the cluster·KMeans module of the Skleans library is used to evaluate the clustering effect and determine the number of cluster centers and cluster radius.

[0010] According to the number of cluster centers and cluster radius, the number of clusters is adaptively optimized, and the distribution center is initially selected as the cluster center. Each demand point is clustered to determine the truck-UAV collaborative transportation node set, and there is no path intersection between the sets.

[0011] Step 2: For each truck-UAV collaborative transport node set, use the ant colony algorithm to solve the pure truck emergency logistics transport path within the set;

[0012] According to the pure truck transportation paths generated within the set, a variable neighborhood simulated annealing algorithm is used to design truck-UAV collaborative transportation paths for the pure truck transportation paths within the set;

[0013] At the truck route nodes, the drone routes are continuously adjusted to optimize the truck-drone collaborative transportation routes to minimize costs while ensuring the timeliness of emergency logistics transportation;

[0014] Step 3: By setting different maximum number of iterations for different examples, determine whether the number of algorithm iterations is greater than the maximum number of iterations or whether the global optimal solution is found. If not, re-enter step 2. Otherwise, output the algorithm solution result and the optimal truck-UAV collaborative transportation path to complete the task.

[0015] As a further solution of the present invention: in step 2, the distribution centers and demand points in the emergency logistics network are all coded with natural numbers, and then an initial path is generated.

[0016] As a further solution of the present invention: in step 2, a neighborhood structure design is adopted, and the neighborhood structure adopted is cyclic exchange.

[0017] As a further solution of the present invention: in step 2, the perturbation process adopts insertion and exchange operations, and a large number of random solutions are obtained through the insertion and exchange operations to expand the solution space.

[0018] As a further solution of the present invention: Insertion operation: refers to removing the sub-path formed by several consecutive shelter nodes on a path and inserting it into another path, and inserting the numbers 1 and 3 in the truck path behind the number 0 in the drone path.

[0019] Transformation operation: refers to exchanging the sub-path formed by several consecutive shelter nodes on a path with another path, exchanging the numbers 9 and 12 in the truck path with the numbers 13 and 15 in the drone path.

[0020] As a further solution of the present invention: In step 2, the 2-opt rule is used to locally optimize the neighborhood solution generated by the disturbance process. The nodes numbered 1 and 7 in the truck path are reversed and exchanged, and the nodes numbered 0 and 12 in the drone path are reversed and exchanged, and the two paths are exchanged.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] The invention has fast calculation efficiency and high solution quality, and has strong practical value for emergency logistics transportation optimization problems with strict timeliness requirements. A two-stage path optimization algorithm based on truck-UAV collaboration is designed in the algorithm, which consists of an adaptive K-means clustering algorithm and a variable neighborhood simulated annealing algorithm.

[0023] The present invention has strong versatility and portability, and can better cope with scenarios where sudden situations lead to road interruptions and affect the timely arrival of emergency supplies. It can provide a reference for governments and decision-making advisory departments to formulate emergency logistics transportation modes, improve the efficiency of emergency rescue, and ensure the credibility of decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart of the two-stage path optimization algorithm in the present invention.

[0025] Figure 2 This is a schematic diagram of the initial path in the present invention.

[0026] Figure 3 Schematic diagram of the cyclic exchange of the neighborhood structure in the present invention.

[0027] Figure 4 The following is a schematic diagram of inserting operation in the present invention.

[0028] Figure 5 It is a schematic diagram of the transformation operation in the present invention.

[0029] Figure 6 It is a schematic diagram of the local search optimization operation in the present invention.

[0030] Figure 7 This is a schematic diagram of pure truck transportation in the present invention.

[0031] Figure 8 Schematic diagram of truck-UAV collaborative transportation in the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] See also Figure 1-Figure 8 In an embodiment of the present invention, a two-stage path optimization method for emergency logistics based on truck-UAV collaboration includes the following steps:

[0034] Step 1: The nodes in the emergency logistics network are composed of distribution centers and demand points. For the nodes in the emergency logistics network, an adaptive K-means clustering method is designed. The inertia_ index of the cluster·KMeans module of the Skleans library is used to evaluate the clustering effect and determine the number of cluster centers and cluster radius.

[0035] According to the number of cluster centers and cluster radius, the number of clusters is adaptively optimized, and the distribution center is initially selected as the cluster center. Each demand point is clustered to determine the truck-UAV collaborative transportation node set, and there is no path intersection between the sets.

[0036] The specific algorithm is shown in Table 1.

[0037] Table 1: Adaptive K-means clustering algorithm

[0038]

[0039]

[0040] Step 2: For each truck-UAV collaborative transport node set, use the ant colony algorithm to solve the pure truck emergency logistics transport path within the set;

[0041] According to the pure truck transportation paths generated within the set, a variable neighborhood simulated annealing algorithm is used to design truck-UAV collaborative transportation paths for the pure truck transportation paths within the set;

[0042] At the truck route nodes, the drone routes are continuously adjusted to optimize the truck-drone collaborative transportation routes to minimize costs while ensuring the timeliness of emergency logistics transportation;

[0043] Step 2.1 Node encoding and initial path generation:

[0044] The distribution centers and demand points in the emergency logistics network are all coded with natural numbers.

[0045] According to the node set obtained by the adaptive K-means clustering algorithm, the "truck-UAV collaborative transportation path" is designed. The variable neighborhood simulated annealing algorithm is used to design the truck-UAV collaborative transportation path between the distribution center and the demand point, and the ant colony algorithm is used to find the optimal truck transportation path within the clustering range of each distribution center; the truck path node is used as the UAV launch center, and the variable neighborhood simulated annealing algorithm is used to continuously optimize the UAV transportation path to jointly form a truck-UAV collaborative transportation path, such as Figure 2 shown.

[0046] Step 2.2 Neighborhood structure design:

[0047] The neighborhood structure used is cyclic exchange, such as Figure 3 shown.

[0048] Step 2.3 Perturbation process:

[0049] The perturbation process uses insertion and exchange operations to obtain a large number of random solutions to expand the solution space.

[0050] Step 2.3.1 Insert operation:

[0051] It means removing the sub-path formed by several consecutive shelter nodes on a path and inserting them into another path, such as Figure 4 As shown, numbers 1 and 3 in the truck path are inserted after number 0 in the drone path.

[0052] Step 2.3.2 Transformation operation:

[0053] It means exchanging the sub-path formed by several consecutive shelter nodes on a path with another path, such as Figure 5 As shown, numbers 9 and 12 in the truck path are exchanged with numbers 13 and 15 in the drone path.

[0054] Step 2.4 Local search:

[0055] The 2-opt rule is used to locally optimize the neighborhood solutions generated by the disturbance process. The nodes between 1 and 7 in the truck path are reversed and exchanged, and the nodes between 0 and 12 in the drone path are reversed and exchanged, and the two paths are exchanged, such as Figure 6 shown.

[0056] Step 3: By setting different maximum number of iterations for different examples, determine whether the number of algorithm iterations is greater than the maximum number of iterations or whether the global optimal solution is found. If not, re-enter step 2. Otherwise, output the algorithm solution result and the optimal truck-UAV collaborative transportation path to complete the task. Specific embodiments

[0058] In order to demonstrate the superiority of the algorithm proposed in this technical solution, seven examples of different scales are used for verification and explanation:

[0059] Step 1: Generate an initial case

[0060] Based on the 7 examples A-n38-k5, A-n38-k5, A-n44-k6, A-n53-k7, A-n61-k9, A-n69-k9, and A-n80-k10 in the CVRP standard example library on the website https: / / www.bernabe.dorronsoro.es / vrp / , the dataset was improved in combination with the research scenario and named M-n32, M-n38, M-n44, M-n53, M-n61, M-n69, and M-n80, respectively.

[0061] Step 2: Example parameter setting

[0062] The parameter settings are shown in Table 1 below:

[0063] Table 1 Related parameter settings

[0064] parameter Numeric Maximum truck load 80kg Maximum payload of drone 30kg Maximum flight range of drone 40km Truck Mileage Cost 1.25 yuan / km Drone Mileage Cost 0.15 yuan / km Truck speed 40km / h Drone speed 60km / h

[0065] Step 3: Analysis of algorithm calculation results

[0066] First, taking M-n32 as an example, the results before and after optimization are compared, as shown in Table 2. The schematic diagram of the transportation path before and after algorithm optimization is shown in Figure 1 , 2 As shown in the figure, after the algorithm optimization, the truck mileage, total transportation cost and transportation time can be effectively reduced, which further proves the effectiveness of the algorithm. Furthermore, the two-stage path optimization algorithm designed by the present invention is used to solve the 7 examples of the above transformation, and each example is run 10 times separately to obtain the optimal result. The algorithm calculation results are shown in Table 3 below.

[0067] It can be seen from the calculation results in Table 3 that after algorithm optimization, the truck mileage, transportation cost and transportation time can be effectively reduced; at the same time, compared with the pure truck transportation route, the truck mileage reduction ratio is 40.41%, 38.22%, 46.21%, 41.96%, 41.96%, 49.01% and 44.34%, respectively, with an average reduction of 43.15%; the transportation cost reduction ratio is 22.05%, 21.78%, 29.21%, 27.01%, 25.80%, 33.90% and 25.74%, respectively, with an average reduction of 26. 49%; the transportation time reduction ratios are 6.88%, 8.20%, 12.83%, 15.62%, 10.72%, 18.91% and 11.38% respectively, with an average reduction of 12.06%; it can be shown that after adding drones to emergency logistics transportation, the algorithm is more inclined to use drones for emergency material transportation due to their advantages such as flexibility and high speed, which can effectively reduce the situation where emergency materials cannot be delivered in time due to destruction or damage to transportation infrastructure. This is also in line with the trend of drones being widely used in emergency rescue in the low-altitude field in recent years.

[0068] Table 2 Comparison of results before and after optimization of M-n32 algorithm

[0069] index Initial Generation Algorithm optimization reduce(%) Number of truck routes 4 4 0 Number of drone transport routes 0 13 - Number of nodes the truck passes 32 11 65.6 Number of nodes the drone passes 0 32 - Transportation cost (yuan) 875.46 682.36 22.05 Transportation time (h) 20.18 18.79 6.9

[0070] Table 3 Comparison of the results of the two-stage path optimization algorithm for seven examples

[0071] Calculation example <![CDATA[R t ]]> <![CDATA[R d ]]> <![CDATA[L 1 ]]> <![CDATA[L 2 ]]> Δ <![CDATA[C 1 ]]> <![CDATA[C 2 ]]> Δ <![CDATA[T 1 ]]> <![CDATA[T 2 ]]> Δ M-n32 4 13 574 342 40.41 875.46 682.36 22.05 20.18 18.79 6.88 M-n38 4 11 620 383 38.22 932.95 729.70 21.78 21.56 19.79 8.20 M-n44 5 14 792 426 46.21 1147.95 812.55 29.21 26.72 23.29 12.83 M-n53 5 14 753 437 41.96 1099.20 802.30 27.01 25.55 21.56 15.62 M-n61 6 18 784 455 41.96 1137.95 844.30 25.80 26.48 23.64 10.72 M-n69 6 19 1059 540 49.01 1481.70 979.35 33.90 34.73 28.16 18.91 M-n80 8 23 1283 714 44.34 1761.70 1308.10 25.74 41.45 36.73 11.38 average 43.15 26.49 12.06

[0072] Note: R t , R d are the number of truck and drone paths respectively; L 1 , L 2 are the mileage of trucks and trains before and after optimization (km); C 1 , C 2 are the transportation costs before and after optimization (yuan); T 1 , T 2 The transportation time before and after was optimized respectively (h); Δ was the reduction ratio (%).

[0073] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

Claims

1. A two-stage path optimization method for emergency logistics based on truck-UAV collaboration, characterized by: The following steps are involved: Step 1: The nodes in the emergency logistics network are composed of distribution centers and demand points. For the nodes in the emergency logistics network, an adaptive K-means clustering method is designed. The inertia_ index of the cluster·KMeans module of the Skleans library is used to evaluate the clustering effect and determine the number of cluster centers and cluster radius. According to the number of cluster centers and cluster radius, the number of clusters is adaptively optimized, and the distribution center is initially selected as the cluster center. Each demand point is clustered to determine the truck-UAV collaborative transportation node set, and there is no path intersection between the sets. Step 2: For each truck-UAV collaborative transport node set, use the ant colony algorithm to solve the pure truck emergency logistics transport path within the set; According to the pure truck transportation paths generated within the set, a variable neighborhood simulated annealing algorithm is used to design truck-UAV collaborative transportation paths for the pure truck transportation paths within the set; At the truck route nodes, the drone routes are continuously adjusted to optimize the truck-drone collaborative transportation routes to minimize costs while ensuring the timeliness of emergency logistics transportation; Step 3: By setting different maximum number of iterations for different examples, determine whether the number of algorithm iterations is greater than the maximum number of iterations or whether the global optimal solution is found. If not, re-enter step 2. Otherwise, output the algorithm solution result and the optimal truck-UAV collaborative transportation path to complete the task.

2. The two-stage path optimization method for emergency logistics based on truck-UAV collaboration according to claim 1 is characterized in that: In step 2, the distribution centers and demand points in the emergency logistics network are coded with natural numbers, and then the initial path is generated.

3. The two-stage path optimization method for emergency logistics based on truck-UAV collaboration according to claim 1 is characterized in that: In step 2, a neighborhood structure design is adopted, and the neighborhood structure adopted is cyclic exchange.

4. The two-stage path optimization method for emergency logistics based on truck-UAV collaboration according to claim 1 is characterized in that: In step 2, the perturbation process uses insertion and exchange operations to obtain a large number of random solutions to expand the solution space.

5. The two-stage path optimization method for emergency logistics based on truck-UAV collaboration according to claim 4 is characterized in that: Insertion operation: refers to removing the sub-path formed by several consecutive shelter nodes on a path and inserting it into another path, inserting the number 1 and 3 in the truck path behind the number 0 in the drone path; Transformation operation: refers to exchanging the sub-path formed by several consecutive shelter nodes on a path with another path, exchanging the numbers 9 and 12 in the truck path with the numbers 13 and 15 in the drone path.

6. The two-stage path optimization method for emergency logistics based on truck-UAV collaboration according to claim 1 is characterized in that: In step 2, the 2-opt rule is used to locally optimize the neighborhood solutions generated by the disturbance process. The nodes between 1 and 7 in the truck path are reversed and exchanged, the nodes between 0 and 12 in the drone path are reversed and exchanged, and the two paths are exchanged.

Images