Vehicle-unmanned aerial vehicle distribution path optimization method considering customer distribution urgent degree

Abstract

The invention discloses a vehicle-unmanned aerial vehicle distribution path optimization method considering customer distribution urgent degree, and relates to the field of logistics distribution path optimization. The method comprises the following steps: firstly, establishing a customer distribution urgent coefficient by considering a sensitive degree parameter of a customer for receiving time; secondly, a distribution time efficiency coefficient is established based on the customer distribution urgency coefficient, and a vehicle-unmanned aerial vehicle cooperative distribution path optimization model is constructed by taking the minimum distribution time efficiency coefficient as an optimization target; then, aiming at a dynamic order (a customer cancels or adds an order), calculating a dynamic deviation coefficient, and re-planning a path by taking minimization of the coefficient as a target; and finally, solving the established multi-target path planning optimization model by adopting a mode of combining a kmeans clustering algorithm, a greedy algorithm and a genetic algorithm. The method can effectively plan the collaborative distribution route of the logistics vehicle and the unmanned aerial vehicle, adapts to dynamic order changes, improves the distribution efficiency and customer satisfaction, and remarkably improves the overall service level of urban terminal distribution.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics distribution path optimization, and particularly relates to a vehicle-unmanned aerial vehicle (UAV) distribution path optimization method considering customer distribution urgency. BACKGROUND

[0002] With the rapid development of e-commerce, the logistics distribution industry is facing increasing distribution demand and complex distribution scenarios. Traditional logistics distribution methods are often inefficient in dealing with large-scale, diversified orders and complex urban traffic environments, making it difficult to meet customer requirements for delivery timeliness and accuracy.

[0003] In recent years, the development of unmanned aerial vehicle technology has provided a new solution for logistics distribution. Through the collaborative distribution of logistics vehicles and unmanned aerial vehicles, the advantages of both can be fully utilized to improve distribution efficiency and reduce distribution costs. However, in practical applications, the collaborative distribution of logistics vehicles and unmanned aerial vehicles faces many challenges. For example, how to reasonably plan the distribution paths of logistics vehicles and unmanned aerial vehicles while considering the range and weight limitations of unmanned aerial vehicles to ensure that each express delivery point is delivered in a timely and accurate manner; how to deal with dynamic orders (new or canceled orders) caused by changes in customer demand, minimize the impact on the original distribution plan when re-planning the path, and ensure the timeliness and customer satisfaction of overall distribution.

[0004] Currently, existing vehicle path optimization methods have certain limitations in dealing with these complex problems, making it difficult to meet the needs of actual logistics distribution scenarios. Therefore, there is an urgent need for a more efficient and intelligent vehicle and unmanned aerial vehicle collaborative path optimization method to address various challenges in logistics distribution and improve the overall service level of logistics distribution. SUMMARY

[0005] The purpose of the present application is to provide a vehicle-unmanned aerial vehicle distribution path optimization method considering customer distribution urgency, aiming to solve problems such as path planning and dynamic order processing in logistics distribution, and improve distribution efficiency and customer satisfaction.

[0006] To achieve the above purpose, the present application adopts the following technical solutions:

[0007] A vehicle-unmanned aerial vehicle distribution path optimization method considering customer distribution urgency, comprising the following steps:

[0008] Step 1: considering the sensitivity parameter of customers to delivery time, establishing a customer distribution urgency coefficient;

[0009] Step 2: establishing a distribution timeliness coefficient based on the customer distribution urgency coefficient, taking the minimum distribution timeliness coefficient as the optimization objective, and constructing a vehicle-unmanned aerial vehicle collaborative distribution path optimization model to improve the timeliness of end logistics distribution.

[0010] Step 3: For dynamic orders (customer cancellation or new orders), a dynamic deviation coefficient is calculated to re-plan the path with the goal of minimizing the coefficient, and a multi-objective path planning optimization model is constructed;

[0011] Step 4: The multi-objective path planning optimization model is solved using the kmeans clustering algorithm combined with the greedy algorithm and genetic algorithm.

[0012] To optimize the above technical solutions, specific measures / limitations are also taken, including:

[0013] In step 1, the urgency coefficient ugi is considered as follows:

[0014] The customer delivery urgency coefficient ugi is considered comprehensively according to the category, value, and customer consumption history evaluation of the goods, and is used to determine the customer's sensitivity to the delivery time. ugi is defined in the range of 0 to 1, and the larger ugi represents the higher sensitivity of the customer to the delivery time, and the higher the urgency.

[0015] In step 2, the delivery timeliness coefficient is defined as follows:

[0016] The vehicle-unmanned aerial vehicle collaborative delivery path planning is performed with the goal of minimizing the delivery timeliness coefficient Tic. The delivery timeliness coefficient Tic is the sum of the timeliness of all express points, and is defined as follows:

[0017]

[0018] Where Tic is the delivery timeliness coefficient, ari is the actual delivery time of express point i, ugi is the urgency coefficient of express point i, and toi is the tolerance delivery time of express point i. If the actual delivery time ari of a single express is less than the tolerance delivery time toi, the timeliness of the express is 0, and if ari is greater than toi, the difference between the two is multiplied by the urgency coefficient ugi to obtain the timeliness of the express;

[0019] The optimization goal is to minimize the delivery timeliness coefficient, and the objective function is defined as follows:

[0020] minf(x)=Tic

[0021] The parameters such as the logistics vehicle speed cv=40km / h, the unmanned aerial vehicle speed dv=60km / h, the maximum load of the unmanned aerial vehicle dl_m=12kg, and the maximum range ds=8km are obtained from the platform database, and the model constraint conditions are established.

[0022] At the same time, in the path planning process, the cargo capacity of the unmanned aerial vehicle should be less than its maximum load, and the longest distance of a single flight of the unmanned aerial vehicle should be less than its maximum range, etc.

[0023] In step 3, when a dynamic order (customer cancels an order or adds a new order) occurs, the dynamic increase point set add_eps and the dynamic cancellation point set can_eps are obtained.

[0024] To meet customer demand, the dynamic deviation coefficient Dyd is defined as follows:

[0025] Dyd = 0.4 * Tic * + 0.6 * |Δar i | * ug i

[0026] Where Tic * represents the new delivery timeliness coefficient affected by demand changes, |Δar i | * ug i represents the original customer delivery time change, and the dynamic deviation coefficient Dyd is the weighted sum of the new delivery timeliness coefficient Tic * and the original customer delivery time change. The timeliness coefficient of the new route (step 2) has a weight of 0.4, and the delivery experience of the original customer has a weight of 0.6.

[0027] The new delivery timeliness coefficient after the change is the sum of all the delivery timeliness after the change, and the calculation method is similar to the initial calculation of the delivery timeliness coefficient, which is defined as follows:

[0028]

[0029] Where eps is the current remaining undelivered customer point, and the new delivery point set m represents the sum of the current undelivered customer points and the new point set, and the point set whose demand has been canceled is removed.

[0030] The dynamic route construction aims to minimize the dynamic deviation coefficient Dyd to ensure that the new route has the least impact on the delivery time of the unaffected customers and meets all the constraints. The objective function is defined as follows:

[0031] min f(x) = Tic *

[0032] In step 4, the method of kmeans clustering is used to classify the delivery points according to geographical location distribution. The center point and the secondary center point are selected according to the relative position while clustering. Then a greedy algorithm is used to generate some legal initial population. Finally, a genetic algorithm is used to optimize the route, which is divided into the following four steps:

[0033] Step 4.1: Classify customer points using kmeans, encode all delivery points into a series of clusters according to geographical features, where the distance meets the UAV range as the UAV delivery point, and the cluster center as the vehicle delivery point. Each cluster contains at least one vehicle delivery point and two UAV delivery points.

[0034] Step 4.2: Use the greedy algorithm to generate the initial closed-loop vehicle path with the shortest vehicle running time as the constraint.

[0035] Step 4.3: According to the generated vehicle path, use the greedy algorithm to generate the initial UAV path with the range, load, task completion of two UAVs, and the UAV and vehicle convergence conditions as constraints. If the UAV task point in a cluster cannot be delivered by the UAV, it will be inserted into the vehicle path and re-operated according to step 2.

[0036] Step 4.4: Use genetic algorithm to generate new paths by crossing part of the delivery tasks of UAVs and vehicles, and check whether the new paths meet the constraints of vehicle and UAV operation, and repair the paths that do not meet the constraints. The newly generated paths and the top 20% elite individuals in the original path are proportionally formed into a new population, and the dynamic bias coefficient in step 3 is used as the optimization target to optimize the vehicle-UAV collaborative delivery path. When all paths are constructed, the method of eliminating sub-loops is used to check and optimize the planned path to ensure that there are no sub-loops in all paths. The method of checking sub-loops is as follows:

[0037]

[0038]

[0039]

[0040]

[0041]

[0042] where cs eps represents the set of vehicle delivery points, represents the set of two UAV delivery points, and eps represents the set of all delivery points. A, B represent any set, and in the formula, cs A , cs B , represents the number of all elements in the subscript set. cr eps (a, b) represents the set of edges passed by the vehicle from delivery point a to delivery point b. represents the set of all delivery edges in the jth mission of the i-th drone.

[0043] This formula means that for any proper subset S of a vertex set, if the sum of the number of edges connecting the vertices is less than or equal to the number of vertices minus 1, then the path does not contain subcircuits.

[0044] At the same time, it is also necessary to ensure that the path points where the logistics vehicle completes one delivery and the path points where the two drones complete all deliveries are the sum of all the path points, and there is no overlap between the path points passed by the logistics vehicle and the path points passed by the drones during delivery.

[0045] In the process of solving the model, it is also necessary to ensure that:

[0046] (1) The drone can be launched from any delivery point of the logistics vehicle, carry some goods, and go to several delivery points for delivery. After completing the delivery, it will go to the meeting point to wait for the logistics vehicle;

[0047] (2) The logistics vehicle immediately replaces the battery of the drone after picking it up;

[0048] (3) The time it takes for the drone to be launched and meet up with the logistics vehicle must be less than the time it takes for the logistics vehicle to travel from the launch point to the meeting point.

[0049] (4) For new orders in the dynamically added point set add_eps, it is necessary to reasonably arrange for delivery by logistics vehicles or drones based on their express weight, delivery location and other information, and include them in the path planning calculation;

[0050] (5) For the canceled orders in the dynamic cancellation point set can_eps, the relevant path points and edges should be removed from the original path planning, and the relevant parameters should be recalculated.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. The vehicle path optimization method proposed in this invention comprehensively considers the coordinated delivery of logistics vehicles and drones, as well as factors such as the range and load restrictions of drones. It can rationally plan delivery paths, improve delivery efficiency, and reduce delivery costs.

[0053] 2. By introducing a dynamic deviation coefficient, we effectively address the dynamic order problem caused by changes in customer demand, minimize the impact on the original delivery plan when replanning the route, and ensure the timeliness of overall delivery and customer satisfaction.

[0054] 3. Ensure that a single drone mission starts from the upstream point of the vehicle path and converges at the downstream point. There is no sub-loop, which conforms to real logic and greatly improves the utilization rate of drones.

[0055] 4. The vehicle path optimization method considers the "urgency coefficient", an empirical coefficient, which reflects the customer's historical habits, making the entire model closer to customer needs.

[0056] 5. The vehicle path optimization system has clear division of labor among modules, and the modules work together to achieve efficient and intelligent path planning and distribution management, has good practicability and expansibility, provides an effective solution for the logistics and distribution industry, and provides a realistic case basis for the development of low-altitude economy. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 is a flowchart of the method of the present application.

[0058] Figure 2 is an optimal static route planning graph of the method case of the present application.

[0059] Figure 3 is an optimal route planning graph after the method case of the present application occurs dynamic change. DETAILED DESCRIPTION

[0060] The above content of the present application will be further described in the form of examples, but this should not be understood as limiting the scope of the above subject matter of the present application to the following examples. Any technology realized based on the above content of the present application belongs to the scope of the present application.

[0061] The present application proposes a vehicle-unmanned aerial vehicle collaborative distribution path optimization method considering the urgency of customer distribution and dynamic demand, and the customer distribution graph is as shown in Figure 1 The method comprises the following steps:

[0062] (1) The customer distribution urgency coefficient ugi is considered as follows:

[0063] The customer distribution urgency coefficient ugi is comprehensively considered according to the category, value, customer consumption history evaluation and other parameters of the goods, and is used to determine the sensitivity of the customer to the receipt time. The range of ugi is defined as between 0 and 1, and the greater ugi represents the more sensitive the customer is to the receipt time, and the higher the urgency.

[0064] (2) A distribution timeliness coefficient based on the customer distribution urgency coefficient is established, the minimum distribution timeliness coefficient is taken as the optimization target, a vehicle-unmanned aerial vehicle collaborative distribution path optimization model is constructed, and the timeliness of the terminal logistics distribution is improved.

[0065] In step (2), the minimum timeliness coefficient Tic is taken as the target, the data and parameters obtained are used to perform path planning calculation by using a mathematical model. The timeliness coefficient Tic is the sum of all express timeliness, and the formula is as follows:

[0066]

[0067] Wherein, Tic is the timeliness coefficient, ari is the actual delivery time of the pickup point i, ugi is the urgency coefficient of the pickup point i, toi is the tolerant delivery time of the pickup point i. If the actual delivery time ari of a single express point is less than the tolerant delivery time toi, the express timeliness is 0, if ari is greater than toi, the difference between the two is multiplied by the urgency coefficient ugi as the timeliness of the express point.

[0068] The objective function should select the route with the minimum timeliness coefficient, defined as follows:

[0069] minf(x)=Tic

[0070] The parameters such as the logistics vehicle speed cv=40km / h, the unmanned aerial vehicle speed dv=60km / h, the maximum load of the unmanned aerial vehicle dl_m=12kg, the maximum range ds=8km, etc. are obtained from the platform database, and the model constraint conditions are established.

[0071] At the same time, in the process of path planning, the cargo capacity of the unmanned aerial vehicle should be less than its maximum load, and the longest distance of the unmanned aerial vehicle in a single flight should be less than its maximum range, etc.

[0072] (3) For dynamic orders (customers cancel or add orders), the dynamic deviation coefficient is calculated, and the path is re-planned with the minimum coefficient as the target

[0073] The dynamic deviation coefficient Dyd is composed of the weighted sum of the timeliness coefficient after the change and the customer delivery experience coefficient not affected by the demand, defined as follows:

[0074] Dyd=0.4*Tic * +0.6*|Δar i |*ug i

[0075] Wherein, Tic * represents the new delivery timeliness coefficient affected by the change of demand, |Δar i |*ug i represents the change of original customer delivery time, and the dynamic deviation coefficient Dyd is the weighted sum of the new delivery timeliness coefficient Tic * and the change of original customer delivery time; the timeliness coefficient weight of the new route is 0.4, and the customer delivery experience weight is 0.6.

[0076] The new delivery timeliness coefficient after the change is the sum of all express delivery timeliness after the change, and the calculation method is similar to the calculation of the delivery timeliness coefficient at the beginning, defined as follows:

[0077]

[0078] Wherein, eps is the current remaining undelivered customer points, the new delivery point set m is expressed as the sum of the current undelivered customer points and the new point set, and the point set whose demand has been cancelled is removed.

[0079] The dynamic route construction takes the minimum dynamic deviation coefficient Dyd as the target, ensures that the new route has the minimum impact on the delivery time of the unaffected customers, and meets all the constraint conditions. The objective function is defined as follows:

[0080] minf(x) = Tic *

[0081] (4) Using the method of kmeans clustering, the delivery points are classified according to the geographical location distribution, and the center point and the secondary center point are selected according to the relative position while clustering. Then, a greedy algorithm is used to generate some legal initial population, and finally, a genetic algorithm is used to optimize the route, which is divided into the following four steps:

[0082] Step 4.1: Use kmeans to classify customer points, and according to the geographical location characteristics, all delivery points are coded into a series of clusters, wherein the distance satisfies the UAV range as the UAV delivery point, and the cluster center is as the vehicle delivery point. Each cluster contains at least one vehicle delivery point and two UAV delivery points.

[0083] Step 4.2: Use the greedy algorithm to generate the initial closed-loop vehicle path with the shortest vehicle running time as the constraint.

[0084] Step 4.3: According to the generated vehicle path, use the greedy algorithm to generate the initial UAV path with the range, load, task completion of two UAVs, and the convergence conditions of UAV and vehicle as constraints. If the UAV task point in a cluster cannot be delivered by the UAV, it will be inserted into the vehicle path, and the operation will be re-operated according to step 2.

[0085] Step 4.4: Use the genetic algorithm to generate new paths by crossing part of the delivery tasks of UAVs and vehicles, and check whether the new path meets the various constraints of vehicle and UAV operation, and repair the path that does not meet the constraint conditions. The newly generated path and the top 20% elite individuals in the original path are proportionally formed into a new population, and the dynamic deviation coefficient in step 3 is used as the optimization target to optimize the vehicle-UAV collaborative delivery path. When all the paths are constructed, the method of eliminating sub-loops is used to check and optimize the planned path to ensure that there is no sub-loop in all the paths, and the formula is expressed as follows:

[0086]

[0087]

[0088]

[0089]

[0090]

[0091] Among them, cs eps represents the set of vehicle delivery points, represents the set of two drone delivery points, and eps represents the set of all delivery points. A and B represent any set, which in the formula represent the proper subset of the vehicle delivery point set and the proper subset of the drone delivery point set. A 、S B 、 Indicates the number of all elements in the subscript set. cr eps (a,b) represents the set of edges that a vehicle passes through when traveling from the delivery point numbered a to the delivery point numbered b. represents the set of all delivery edges in the jth mission of the i-th drone.

[0092] This formula means that for any proper subset S of a vertex set, if the sum of the number of edges connecting the vertices is less than or equal to the number of vertices minus 1, then the path does not contain subcircuits.

[0093] At the same time, it is also necessary to ensure that the path points where the logistics vehicle completes one delivery and the path points where the two drones complete all deliveries are the sum of all the path points, and there is no overlap between the path points passed by the logistics vehicle and the path points passed by the drones during delivery.

[0094] The technical solution of the present invention is further illustrated by a specific embodiment below:

[0095] We selected 20 points in the dense data set from the Solomon standard data set as the research objects. The distribution of demand distribution points and warehouse nodes are as follows: Figure 2 As shown in the figure, the package number, package weight, package delivery location (x,y coordinates), urgency coefficient, and delivery tolerance are determined. Five additional points are identified as dynamic addition points, and three of the known points are selected as dynamic cancellation points to simulate a dynamic task. Parameters such as the speed of the logistics vehicle and the speed of the drone are determined, as well as the maximum payload and maximum range of the drone. The set of express delivery points (eps), the set of route points for the logistics vehicle to complete a delivery (cs_eps), and the set of edges along the route for the logistics vehicle to complete a delivery (cr_{eps}) are initialized.

[0096] The initial task allocation is completed by using a task allocation model with timeliness coefficient as the target. Then, the dynamic change is simulated by using the selected increase points and cancellation points. The task state is adjusted according to the increase points and cancellation points, the path allocation mode after the dynamic change is calculated by using a dynamic task processing model with dynamic deviation coefficient as the target, and the operation time is recorded respectively.

[0097] After the model calculation, the result of the customer demand point in the case is calculated as shown in Figure 2 and Figure 3 Compared with the optimization algorithm Gurobi, the heuristic model greatly reduces the calculation time under the premise of ensuring the timeliness coefficient and the dynamic deviation coefficient, and has application value.

[0098] Comparison of experimental results of the optimization model and the Gurobi algorithm

[0099]

[0100] After actual test, the vehicle path optimization method can effectively improve the distribution efficiency, reduce the timeliness coefficient, quickly adjust the path when processing the dynamic order, minimize the influence on the original distribution plan, and has good application effect.

[0101] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Any skilled person in the art can make any simple modification, equivalent replacement and improvement to the above embodiment according to the technical essence of the present application without departing from the technical solution of the present application, and all of them still belong to the protection scope of the present application.

Claims

1. A vehicle-drone delivery path optimization method considering the urgency of customer delivery, characterized by: The following steps are involved: Step 1: Consider the customer's sensitivity to delivery time and establish the customer delivery urgency coefficient; Step 2: Establish a delivery time efficiency coefficient based on the customer's delivery urgency coefficient. Taking the minimum delivery time efficiency coefficient as the optimization goal, build a vehicle-drone collaborative delivery path optimization model to improve the timeliness of terminal logistics delivery; Step 3: For dynamic orders, the dynamic deviation coefficient is calculated, and the path is replanned with the goal of minimizing the dynamic deviation coefficient, thereby building a multi-objective path planning optimization model; Step 4: Use kmeans clustering algorithm combined with greedy algorithm and genetic algorithm to solve the established multi-objective path planning optimization model.

2. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 1 is characterized in that: The specific process of step 1 is as follows: The customer delivery urgency coefficient ugi is used to determine the customer's sensitivity to the delivery time based on a comprehensive consideration of the goods' category, value, and customer consumption history evaluation parameters. The range of ugi is defined between 0 and 1. The larger the ugi, the more sensitive the customer is to the delivery time and the higher the degree of urgency.

3. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 1 is characterized in that: The delivery efficiency coefficient in step 2 is defined as follows: Among them, Tic is the delivery timeliness coefficient, ari is the actual delivery time of receiving point i, ugi is the urgency coefficient of receiving point i, and toi is the tolerated delivery time of receiving point i; the delivery timeliness coefficient Tic is the sum of the delivery timeliness of all express deliveries. If the actual delivery time ari of a single express is less than the tolerated delivery time toi, the timeliness is 0. If it is greater, the product of the difference between the two and the urgency coefficient ugi is used as the timeliness.

4. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 1 is characterized in that: The specific process of the vehicle-UAV collaborative delivery path optimization model in step 2 is as follows: Establish a function with the minimum delivery timeliness coefficient Tic as the optimization objective, and select the route with the minimum delivery timeliness coefficient: minf(x)=Tic At the same time, the constraints that the drone's cargo capacity must be less than its maximum load and the drone's longest single flight distance must be less than its maximum range must be met.

5. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 1 is characterized in that: The specific process of defining the dynamic deviation coefficient in step 3 is as follows: When a dynamic order occurs, that is, when a customer cancels an order or places a new order, the dynamic deviation coefficient Dyd is calculated: Dyd=0.4*Tic * +0.6*|Δar i |*and i Among them, Tic * represents the new delivery timeliness coefficient affected by demand changes, |Δar i |*ug i Indicates the change in the original customer's delivery time. The dynamic deviation coefficient Dyd is the new delivery timeliness coefficient Tic * and the weighted sum of the changes in the original customer's delivery time; and based on dynamic demand, the vehicle-UAV dynamic path replanning is performed with the minimum dynamic deviation coefficient Dyd as the optimization goal. The objective function is defined as follows: minf(x)=Tic * 。 6. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 5, characterized in that: New delivery timeliness coefficient Tic * is defined as follows: Among them, the new delivery point set m represents the sum of the current undelivered customer points and the newly added point set, and removes the point set whose demand has been cancelled; add eps Indicates dynamically adding a point set, can eps Represents the dynamic cancellation point set, and eps represents the current remaining undelivered customer points.

7. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 1, characterized in that: In step 4, the kmeans clustering method is used to classify the receiving points according to their geographical distribution. During clustering, the center point and secondary center point are selected according to their relative positions. Then, a greedy algorithm is used to generate some legal initial populations. Finally, a genetic algorithm is used to optimize the route. The specific steps are as follows: Step 4.1: Use kmeans to classify customer points. Based on their geographic location, encode all delivery points into a series of clusters. Points within the range of the drone are considered drone delivery points, and points at the center of the cluster are considered vehicle delivery points. Each cluster contains at least one vehicle delivery point and two drone delivery points. Step 4.2: Generate the initial closed-loop vehicle path using the greedy algorithm with the shortest vehicle running time as the constraint; Step 4.3: Based on the generated vehicle path, use a greedy algorithm to generate an initial drone path, with the two drones' range, payload, mission completion status, and drone-vehicle convergence conditions as constraints. If a drone mission point in a cluster cannot be delivered by the drone, insert it into the vehicle path and recalculate according to step 2. Step 4.4: Use a genetic algorithm to generate new routes by crossing parts of the delivery tasks of the drone and vehicle, check whether the new routes meet the various constraints of the vehicle and drone operations, and repair the routes that do not meet the constraints; the newly generated routes and the top 20% elite individuals in the original routes are proportionally combined to form a new population, and the dynamic deviation coefficient in step 3 is used as the optimization target to optimize the vehicle-drone collaborative delivery route.

8. The vehicle-drone delivery path optimization method considering the urgency of customer delivery according to claim 1 is characterized in that: When constructing all paths, use the method of eliminating sub-loops to check and optimize the planned paths to ensure that no sub-loops appear in any path. The method of checking sub-loops is as follows: Among them, cs eps represents the set of vehicle delivery points, represents the set of two UAV delivery points, eps represents the set of all delivery points; A, B represent any set, representing the proper subset of the vehicle delivery point set and the proper subset of the UAV delivery point set; S A 、S B 、 Indicates the number of all elements in the subscript set; cr eps (a,b) represents the set of edges that a vehicle passes through when traveling from the delivery point numbered a to the delivery point numbered b; represents the set of all delivery edges in the jth mission of the i-th drone.

Images