Urban-end co-delivery center site selection method and system based on air-ground coordination

Abstract

The present application belongs to the technical field of co-delivery center site selection. Disclosed are an urban-end co-distribution center site selection method and system based on air-ground coordination, which are used for solving the technical problems of an existing air-ground coordination co-delivery center site selection model being insufficiently integrated with an unmanned aerial vehicle delivery technique, failing to take into consideration constraint conditions on unmanned aerial vehicles, and the solving efficiency being low. The method comprises: establishing a planned layout set of an air-ground coordination co-delivery center; taking the minimization of the overall delivery time and the minimization of the total network operation cost as objectives to construct an air-ground coordination layout planning model; setting constraint conditions of the air-ground coordination layout planning model; by means of a grey wolf population algorithm, generating an initial site selection solution set for a pre-selected point set of the co-delivery center; filtering out invalid solutions in the initial site selection solution set on the basis of the constraint conditions; and performing an individual iterative update on the filtered initial site selection solution set, and performing a fitness assessment and no-fly zone crossing determination on a new solution set obtained after each iteration, so as to obtain a final site selection scheme.

Description

An urban terminal co-distribution center site selection method and system based on air-ground cooperation

[0001] The present application claims priority to the Chinese patent application No. 202410814057.9, filed on June 24, 2024, and entitled "An urban terminal co-distribution center site selection method and system based on air-ground cooperation", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the technical field of co-distribution center site selection, in particular to an urban terminal co-distribution center site selection method and system based on air-ground cooperation. BACKGROUND

[0003] With the rapid development of online retail industry and e-commerce, the demand for express delivery continues to surge. Traditional logistics distribution methods are difficult to meet the needs of users for delivery timeliness and delivery volume. In recent years, with the rapid development of unmanned aerial vehicle delivery technology, its application in the field of express delivery is becoming more and more widespread. The air-ground cooperative co-distribution network composed of unmanned aerial vehicles and vehicles is one of the effective ways to alleviate the pressure of express delivery.

[0004] However, the construction of air-ground cooperative co-distribution network, the site selection of co-distribution center is a key link, and the site selection result has a direct impact on the delivery efficiency and cost of the overall co-distribution network. Currently, some scholars have carried out research on site selection problems, but there are still some problems, such as: (1) The cooperative operation mechanism of co-distribution network is not well studied, and only economic and timeliness are considered when establishing the site selection model, without considering how to integrate with unmanned aerial vehicle delivery technology. In the actual delivery process, the flight of unmanned aerial vehicles is restricted by many conditions such as weather, no-fly zones, flight range, etc. If these restrictive conditions are ignored in the modeling stage, the final site selection result will be quite different from the actual situation, and it will be difficult to put into production. (2) The solution efficiency of the site selection model is not high, and the solution quality is low. These problems lead to low solution efficiency of traditional co-distribution center site selection methods, and the solution is far from the actual application scenario, and the practicality is poor. SUMMARY

[0005] The embodiments of the present application provide an urban terminal co-distribution center site selection method and system based on air-ground cooperation, which is used to solve the following technical problems: the existing air-ground cooperative co-distribution center site selection model is not well integrated with unmanned aerial vehicle delivery technology, the restrictive conditions of unmanned aerial vehicles are not considered, and the solution efficiency is not high, resulting in poor final site selection effect.

[0006] The embodiments of the present application adopt the following technical solutions:

[0007] In one aspect, the embodiments of the present application provide an urban terminal co-distribution center site selection method based on air-ground cooperation, the method comprising:

[0008] establish a set of planning layouts of the air-ground collaborative co-deployment center based on the principle of full coverage of demand points, the set of planning layouts comprising: a set of customer demand points, a set of alternative co-deployment centers, a set of co-deployment center classifications, a set of unmanned aerial vehicles, and a set of vehicles;

[0009] build an air-ground collaborative layout planning model according to the set of planning layouts, with the minimum overall delivery time and the minimum total network operation cost as the targets;

[0010] set constraint conditions of the air-ground collaborative layout planning model based on flight constraints of the unmanned aerial vehicles;

[0011] determine a set of co-deployment center preselected points based on the set of customer demand points;

[0012] generate a set of site selection initial solutions for the set of co-deployment center preselected points by using a grey wolf swarm algorithm, and filter invalid solutions in the set of site selection initial solutions according to the constraint conditions;

[0013] build a fitness function according to the air-ground collaborative layout planning model;

[0014] perform individual iterative updating on the filtered set of site selection initial solutions, perform fitness evaluation and forbidden flight zone crossing discrimination on the new solution set after each iteration, and obtain a final site selection scheme.

[0015] In a feasible implementation, an air-ground collaborative layout planning model is built with the minimum overall delivery time and the minimum total network operation cost as the targets, comprising:

[0016] build a delivery time function based on the time for the unmanned aerial vehicles and the vehicles to complete a delivery task;

[0017] build a cost function based on the construction cost of the co-deployment center and the delivery operation cost;

[0018] build the air-ground collaborative layout planning model with the minimum values of the delivery time function and the cost function as the targets: min Z = α∑Z time + β∑Z cost ; where Z is a site selection evaluation target value, Z time is the delivery time function, Z cost is the cost function, α is a delivery time function weight, and β is a cost function weight.

[0019] In a feasible implementation, a delivery time function is built based on the time for the unmanned aerial vehicles and the vehicles to complete a delivery task, and a cost function is built based on the construction cost of the co-deployment center and the delivery operation cost, comprising:

[0020] build the delivery time function according to ​

[0021] wherein a is the number of UAVs in the set of UAVs, b is the number of vehicles in the set of vehicles; u is the UAV index, k is the vehicle index; n is the number of alternative co-fulfillment centers in the set of alternative co-fulfillment centers, m is the number of customer demand points in the set of customer demand points, i is the alternative co-fulfillment center index, and j is the customer demand point index; is the flight distance of the UAV u from node i to node j, is the driving distance of the vehicle k from node i to node j; V u is the average delivery speed of the UAV u, k is the average delivery speed of the vehicle k; is whether the demand point j is delivered by the UAV u from the i point, and is 1 if yes, otherwise 0; is whether the demand point j is delivered by the vehicle k from the i point, and is 1 if yes, otherwise 0;

[0022] According to construct the cost function;

[0023] wherein s is the co-fulfillment center type index in the set of co-fulfillment center classifications; C s is the construction cost of the s-type co-fulfillment center; x is is whether the alternative landing site i is constructed as an s-type co-fulfillment center, and is 1 if yes, otherwise 0; is the delivery operation cost of the UAV co-fulfillment center; is the delivery operation cost of the vehicle co-fulfillment center.

[0024] In a feasible implementation, constraint conditions of the air-ground collaborative layout planning model are set based on flight constraints of the UAVs, including:

[0025] Flight constraints of the UAVs are analyzed, and constraint conditions of the air-ground collaborative layout planning model are set according to the set of planning layouts; wherein the constraint conditions at least include: landing point quantity constraint, co-fulfillment center type constraint, supply-demand point matching constraint, capacity-demand matching constraint, no-fly zone constraint, UAV performance constraint, delivery vehicle performance constraint, service distance constraint, and delivery relationship constraint.

[0026] In a feasible implementation, the set of co-fulfillment center preselected points is determined based on the set of customer demand points, including:

[0027] In the set of customer demand points, a demand point is randomly selected as the first co-fulfillment center preselected point and stored in the set of co-fulfillment center preselected points;

[0028] Calculate the minimum distance between each remaining demand point in the customer demand point set and the pre-selected point of the selected co-distribution center;

[0029] Based on the minimum distance, calculate the probability that each demand point will be selected as the next cluster center, and select the demand point with the highest probability as the new co-matching center pre-selection point, and store it in the co-matching center pre-selection point set;

[0030] Once the number of pre-selected co-matching centers reaches the target number of classifications, the clustering operation is stopped, and the final set of pre-selected co-matching centers is obtained.

[0031] In one feasible implementation, an initial solution set for site selection is generated for the pre-selected co-mating center point set using the gray wolf population algorithm, including:

[0032] According to the address classification formula:

[0033] The pre-selected points of the co-location centers in the pre-selected point set are classified by address; where k∈n′, and n′ is the total number of pre-selected points of the co-location centers in the pre-selected point set;

[0034] Based on the address classification formula, a single pre-selected co-mating center point I is randomly generated using the gray wolf population algorithm. k Initial addressing solution at time t0: I k (t0) = rand(1,2,3); where 1 represents that the current pre-selected point will not build a shared distribution center, 2 represents that the current pre-selected point will build a shared distribution center for drones, and 3 represents that the current pre-selected point will build a shared distribution center for vehicles; rand is a random operator;

[0035] Based on a single co-location center pre-selected point I k The initial location solution at time t0 yields the initial location solution set corresponding to the pre-selected points of the co-location center: I(t0) = [I1(t0), I2(t0), ..., I k (t0),...,I n (t0)].

[0036] In one feasible implementation, filtering invalid solutions in the initial location solution set according to the constraints includes:

[0037] Traverse the initial solution set for site selection and determine whether the initial solution for each pre-selected point of the co-location center satisfies the constraint conditions.

[0038] Check if any of the remaining initial addressing solutions are classified as I. k Pre-selected points for co-location centers with (t0) = 1;

[0039] If so, randomly select one that satisfies I. kThe point k′ of (t0) = 1 is reassigned to I. k′ (t0) = rand(2,3), re-evaluate whether point k′ satisfies the constraints;

[0040] If not, then randomly select one that satisfies I. k The point k″ of (t0) = 2 is reassigned to I. k″ (t0) = 3, re-evaluate whether point k″ satisfies the constraint conditions;

[0041] Initial location solutions that do not meet the constraints are removed from the initial location solution set to filter out invalid solutions.

[0042] In one feasible implementation, a fitness function is constructed based on the air-ground coordinated layout planning model, including:

[0043] according to Construct a time fitness function; the time fitness function is a normalized value of the delivery time function;

[0044] In the time fitness function, when all distribution centers use vehicles for delivery and each demand point is delivered by the furthest distribution center, the maximum delivery time maxZ is taken. time At this point, the minimum delivery time is min Z. time =0, satisfying f time ∈(0,1];

[0045] according to Construct a cost fitness function; the cost fitness function is the normalized value of the cost function;

[0046] In the cost fitness function, the maximum cost, max Z, is taken when all co-distribution centers are drone delivery centers and each demand point is delivered by the furthest co-distribution center. cost At this point, the minimum cost is min Z. cost =0, satisfying f cost ∈(0,1];

[0047] Based on the goal of minimizing both delivery time and operating costs, an overall fitness function is constructed: fitness = ω t f time +ω c f cost ; where ω t ω is the time scaling factor. c This is the cost ratio coefficient.

[0048] In an implementable embodiment, the filtered initial solution set of the site selection is updated iteratively, and the fitness of the new solution set after each iteration is evaluated and the crossing of the forbidden zone is determined to obtain the final site selection scheme, and the method further comprises:

[0049] The filtered initial solution set of the site selection is updated iteratively by the grey wolf swarm algorithm;

[0050] According to the fitness function, the fitness of the new solution set obtained after each iteration is calculated.

[0051] If there is any polygon forbidden zone in the distribution range, the direction number of any line segment in the polygon forbidden zone is calculated according to the position of the unmanned aerial vehicle distribution center and the position of the demand point in the new solution set.

[0052] According to the product of the direction numbers and the determination formula, it is determined whether the line connecting the position of the unmanned aerial vehicle distribution center and the position of the demand point crosses the polygon forbidden zone; if so, the current new solution set is abandoned, and the individual iterative update is continued.

[0053] When the fitness and the forbidden zone determination both meet the preset requirements, the new solution set at this time is determined as the final site selection scheme.

[0054] On the other hand, the embodiment of the present application also provides a city end joint distribution center site selection system based on air-ground cooperation, which comprises:

[0055] A model construction module is configured to establish a planning layout set of air-ground cooperative joint distribution centers based on a demand point full coverage principle; the planning layout set comprises a customer demand point set, a candidate joint distribution center set, a joint distribution center classification set, an unmanned aerial vehicle set and a vehicle set; an air-ground cooperative layout planning model is constructed based on the planning layout set, with the minimum overall distribution time and the minimum total network operation cost as the target; constraint conditions of the air-ground cooperative layout planning model are set based on the flight constraints of the unmanned aerial vehicle; and a joint distribution center preselected point set is determined based on the customer demand point set.

[0056] A model solution module is configured to generate an initial solution set of site selection for the joint distribution center preselected point set by a grey wolf swarm algorithm; to filter invalid solutions in the initial solution set of site selection based on the constraint conditions; to construct a fitness function based on the air-ground cooperative layout planning model; to update the filtered initial solution set of site selection iteratively, and to evaluate the fitness of the new solution set after each iteration and determine the crossing of the forbidden zone to obtain the final site selection scheme.

[0057] Compared with the prior art, the city end joint distribution center site selection method and system based on air-ground cooperation provided by the embodiment of the present application has the following beneficial effects:

[0058] Based on the in-depth analysis of the mechanism of vehicle transportation and unmanned aerial vehicle transportation, the application proposes a "air-ground" two-type collaborative center layout mode, takes the minimum of the overall operation cost and distribution time of the collaborative network as the target, constructs a collaborative center layout planning model with the actual flight conditions of the unmanned aerial vehicle as the constraint, and then solves the high-dimensional complex model of air-ground collaboration by using an improved grey wolf algorithm. On the one hand, the reasonable fitness function is set by comprehensively considering the distribution time and construction cost, so as to reduce the construction and operation cost as much as possible and improve the distribution efficiency. On the other hand, the invalid solution is effectively filtered out and the solution quality and solving efficiency of the algorithm are improved by adding the initial solution screening strategy and the no-fly zone discrimination. BRIEF DESCRIPTION OF DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without any creative effort. In the drawings:

[0060] FIG. 1 is a flowchart of a city end collaborative center site selection method based on air-ground collaboration provided by an embodiment of the present application;

[0061] FIG. 2 is a schematic diagram of air-ground collaborative center planning layout provided by an embodiment of the present application;

[0062] FIG. 3 is a schematic diagram of no-fly zone crossing judgment provided by an embodiment of the present application;

[0063] FIG. 4 is a schematic diagram of the structure of a city end collaborative center site selection system based on air-ground collaboration provided by an embodiment of the present application. DETAILED DESCRIPTION

[0064] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without any creative effort should be within the scope of protection of the present application.

[0065] The embodiment of the present application provides a city end collaborative center site selection method based on air-ground collaboration, as shown in FIG. 1, the city end collaborative center site selection method based on air-ground collaboration includes steps S101-S107:

[0066] S101, based on the demand point full coverage principle, the planning layout set of air-ground collaborative co-configuration center is established.

[0067] Further, the planning layout set includes: a customer demand point set, a candidate co-configuration center set, a co-configuration center classification set, a UAV set and a vehicle set.

[0068] As a feasible implementation, within the customer demand range, two types of co-configuration centers, "air" and "ground", are preset to complete certain distribution tasks by UAVs and vehicles respectively, and through the connection between the co-configuration center and the terminal relay station (or other intelligent receiving point), the full coverage of regional distribution demand points is realized. Fig. 2 is a schematic diagram of air-ground collaborative co-configuration center planning layout, as shown in Fig. 2, the candidate co-configuration center set is I = {1, 2, 3, ……n}, the customer demand range also contains the customer demand point set J = {1, 2, 3, ……m}, and the co-configuration center classification set S = {1, 2}, the co-configuration center types include two types of UAV co-configuration center and vehicle co-configuration center. The UAV set in the UAV co-configuration center is U = {1, 2, 3, ……u}, and the vehicle set is K = {1, 2, 3, ……k}. In addition, there can be one or more flight restricted areas in the customer demand range, or there can be no flight restricted areas, depending on the actual situation.

[0069] S102, taking the minimum overall distribution time and the minimum total network operation cost as the target, an air-ground collaborative layout planning model is constructed according to the planning layout set.

[0070] Further, first, based on the time of UAVs and vehicles completing distribution tasks, a distribution time function is constructed:

[0071] Wherein, a is the number of UAVs in the UAV set, b is the number of vehicles in the vehicle set; u is the UAV label, k ∈ K is the vehicle label; n is the number of candidate co-configuration centers in the candidate co-configuration center set, m is the number of customer demand points in the customer demand point set, i ∈ I is the candidate co-configuration center label, j ∈ J is the customer demand point label; is the flight distance of UAV u from node i to node j, is the driving distance of vehicle k from node i to node j; V u is the average distribution speed of UAV u, V k is the average distribution speed of vehicle k; is whether demand point j is distributed by UAV u taking off from i point, 1 if yes, otherwise 0; is whether demand point j is distributed by vehicle k taking off from i point, 1 if yes, otherwise 0.

[0072] Further, based on the co-configuration center construction cost and the distribution operation cost, a cost function is constructed:

[0073] wherein s e S is the co-configuration center type label in the co-configuration center classification set; C s is the construction cost of the s-type co-configuration center; x is is whether the alternative landing site i is constructed s-type co-configuration center, 1 if yes, otherwise 0; is the distribution operation cost of the unmanned aerial vehicle co-configuration center; is the distribution operation cost of the vehicle co-configuration center.

[0074] Further, the air-ground collaborative layout planning model is constructed with the minimum value of the distribution time function and the cost function as the goal: min Z = a∑Z time + b∑Z cost ; wherein Z is the site selection evaluation target value, Z time is the distribution time function, Z cost is the cost function, a is the distribution time function weight, and b is the cost function weight. The distribution time function weight and the cost function weight can be obtained according to the expert experience, or can be adjusted.

[0075] S103, based on the flight restriction condition of the unmanned aerial vehicle, the constraint condition of the air-ground collaborative layout planning model is set.

[0076] Further, the flight restriction condition of the unmanned aerial vehicle is analyzed, and the constraint condition of the air-ground collaborative layout planning model is set according to the planning layout set; wherein the constraint condition at least includes: landing site quantity constraint, co-configuration center type constraint, supply-demand point matching constraint, capacity-demand matching constraint, no-fly zone constraint, unmanned aerial vehicle performance constraint, distribution vehicle performance constraint, service distance constraint and distribution relationship constraint.

[0077] As a feasible implementation manner,

[0078] 1) Landing site quantity constraint: the total number of co-configuration centers cannot exceed the number of alternative co-configuration centers:

[0079] 2) Type constraint: any alternative co-configuration center is not allowed to set multiple types:

[0080] 3) Supply-demand point matching constraint: for any demand point, there is only one co-configuration center responsible for distribution, and the situation that the demand point is not distributed or multiple landing sites are distributed is not allowed:

[0081] 4) Capacity-demand matching constraint: if the alternative co-configuration center is selected, its capacity must meet the demand, that is, the distribution business volume: wherein E s is the service capacity of the co-fulfillment center i (pieces), h j is the distribution demand of demand point j.

[0082] 5) No-fly zone constraint: if the straight flight path between the candidate co-fulfillment center and the demand point crosses a no-fly zone, the distribution relationship does not exist: wherein A ij is whether the distribution path of the UAV co-fulfillment center i and the demand point j crosses a no-fly zone.

[0083] 6) UAV performance constraint: the actual range must not exceed the maximum range, and the distribution weight must not exceed the maximum load: wherein D u is the maximum range of the UAV (km) ; Q u is the maximum load of the UAV (kg) ; is the single distribution load of the UAV (kg) ; u e U is the UAV label.

[0084] 7) Distribution vehicle performance constraint: the distribution weight must not exceed the maximum load of the logistics vehicle: wherein is the single distribution load of the vehicle (kg), Q k is the maximum load of the vehicle (kg).

[0085] 8) Service distance constraint: the actual distribution distance of any type of co-fulfillment center is less than its maximum service distance R s , and less than half of the performance mileage of this distribution mode: wherein R u is the maximum service distance of the UAV co-fulfillment center (km) ; R k is the maximum service distance of the vehicle co-fulfillment center (km) ; D k is the maximum driving distance of the vehicle (km).

[0086] 9) Distribution relationship constraint: when a point is selected as a co-fulfillment center, there can be a distribution relationship between the co-fulfillment center and the demand point:

[0087] S104, based on the set of customer demand points, determine the set of co-fulfillment center pre-selected points.

[0088] Further, in the set of customer demand points, a demand point is randomly selected as the first co-fulfillment center pre-selected point, and stored in the set of co-fulfillment center pre-selected points. The minimum distance between each demand point in the set of customer demand points and the selected co-fulfillment center pre-selected point is calculated.

[0089] Further, according to the minimum distance, the probability of each demand point being selected as the next clustering center is calculated, and the demand point with the maximum probability is selected as the new co-distribution center preselected point and stored in the co-distribution center preselected point set. After the number of co-distribution center preselected points reaches the target classification number, the clustering operation is stopped, and the final co-distribution center preselected point set is obtained.

[0090] As a feasible implementation, a point is randomly selected from all demand points as the first distribution center preselected point: I1=Random(J), J={1, 2, 3, …, m}. In the formula, I n is the nth point of the distribution center preselected point set, and J is the demand point set.

[0091] Then the minimum distance of each remaining demand point from the selected distribution center preselected point set is calculated: In the formula, r is the target classification number of the clustering algorithm.

[0092] Further, the probability of each demand point being selected as the next clustering center is calculated: and the demand point with the maximum probability is selected as the new distribution center preselected point: wherein, j=j+1. Until j=r, the operation is terminated, and the is the final co-distribution center preselected point set; otherwise, the operation continues.

[0093] S105, through the grey wolf population algorithm, an initial solution set of site selection is generated for the co-distribution center preselected point set; and according to the constraint condition, invalid solutions in the initial solution set of site selection are filtered.

[0094] Further, according to the address classification formula:

[0095] The co-distribution center preselected points in the co-distribution center preselected point set are classified by address; wherein, k∈n′, n′ is the total number of co-distribution center preselected points in the co-distribution center preselected point set.

[0096] Further, according to the address classification formula, a single co-distribution center preselected point I k is randomly generated by the grey wolf population algorithm. k (t0)=rand(1, 2, 3); wherein, 1 represents that the current preselected point does not construct a co-distribution center, 2 represents that the current preselected point constructs a UAV co-distribution center, and 3 represents that the current preselected point constructs a vehicle co-distribution center; and rand is a random operator.

[0097] Further, according to the single co-distribution center preselected point I kThe initial solution of site selection at t0, the initial solution set of site selection corresponding to the pre-selected point set of co-located center is obtained: I(t0)=[I1(t0), I2(t0),…,I k (t0),…,I n (t0)].

[0098] Further, the initial solution set of site selection is filtered in the following way:

[0099] First, the initial solution set of site selection is traversed to determine whether the initial solution of site selection of each pre-selected point of co-located center meets the constraint condition. Then it is checked whether there is a pre-selected point of co-located center whose address is classified as I k (t0)=1 in the remaining initial solution of site selection; if so, a point k' meeting I k (t0)=1 is randomly selected, and I k′ (t0) is re-assigned as rand(2,3), and it is re-determined whether the point k' meets the constraint condition; if not, a point k'' meeting I k (t0)=2 is randomly selected, and I k″ (t0) is re-assigned as 3, and it is re-determined whether the point k'' meets the constraint condition.

[0100] Finally, the initial solution of site selection that does not meet the constraint condition is removed from the initial solution set of site selection, so as to filter out the invalid solution in the initial solution set of site selection.

[0101] S106, according to the air-ground collaborative layout planning model, a fitness function is constructed.

[0102] Further, a time fitness function is constructed according to ; the time fitness function is a normalized value of the distribution time function. In the time fitness function, when all co-located centers are vehicle distribution centers and each demand point is distributed by the farthest co-located center, the maximum distribution time max Z time is taken; at this time, the minimum distribution time min Z time =0, and f time ∈(0,1].

[0103] Further, a cost fitness function is constructed according to ; the cost fitness function is a normalized value of the cost function. In the cost fitness function, when all co-located centers are unmanned aerial vehicle distribution centers and each demand point is distributed by the farthest co-located center, the maximum cost max Z cost is taken; at this time, the minimum cost min Z cost =0, and f cost ∈(0,1].

[0104] Further, based on the objective of simultaneous minimization of delivery time and operation cost, a whole fitness function is constructed: fitness = ω t f time + ω c f cost ; wherein ω t is a time proportion coefficient, and ω c is a cost proportion coefficient.

[0105] S107, individual iterative updating is performed on the filtered initial solution set of site selection, and fitness evaluation and crossing forbidden zone discrimination are performed on the new solution set after each iteration to obtain a final site selection scheme.

[0106] Further, the gray wolf population algorithm is used to perform individual iterative updating on the filtered initial solution set of site selection.

[0107] As a feasible implementation manner, in the traditional gray wolf population algorithm, all calculations are performed in the real number field, and generally, the solution is continuous in the real number space. However, in the site selection problem, if a non-integer solution is generated when updating the wolf group position, the solution cannot be directly used for address classification calculation. Therefore, the position vector of each solution is integerized according to the rounding method to generate a new position vector

[0108] Then, it is judged whether each element in satisfies the address classification formula, and for the solution that does not satisfy the formula, further truncation processing is performed according to the following formula:

[0109] wherein A is the whole wolf group in the gray wolf algorithm.

[0110] Further, according to the fitness function, the fitness of the new solution set obtained after each iteration is calculated.

[0111] Further, if there is any polygon forbidden zone in the delivery range, then according to the position of the unmanned aerial vehicle distribution center and the position of the demand point in the new solution set, the direction number of any line segment in the polygon forbidden zone is calculated. According to the product of the direction numbers and the judgment formula, it is judged whether the line connecting the position of the unmanned aerial vehicle distribution center and the position of the demand point crosses the polygon forbidden zone; if so, the current new solution set is abandoned, and individual iterative updating is continued.

[0112] When the fitness and the forbidden zone discrimination both meet the preset requirements, the new solution set at this time is determined as the final site selection scheme.

[0113] As a feasible implementation, the no-fly zone can be regarded as a closed figure of a plurality of line segments surrounded by a set of vertices. Whether the delivery path of the UAV crosses the no-fly zone can be determined by calculating the number of directions of the line segments to check whether the straight line connecting the distribution center and the demand point has intersection points with the boundary line segments of the no-fly zone. FIG. 3 is a schematic diagram of a no-fly zone crossing determination provided by an embodiment of the present application. As shown in FIG. 3, for the distribution center and the demand point whether the line connecting the distribution center and the demand point passes through the arbitrary polygon no-fly zone can be converted into determining whether the line connecting the distribution center and the demand point intersects with the edges of the polygon, and the direction number method can be used.

[0114] For any line segment in the no-fly zone four direction numbers are calculated according to the following formula:

[0115] The product of the direction numbers has the following cases:

[0116] Among the product of the direction numbers, only when the product of the direction numbers is 0, it cannot be directly determined whether the line segments are parallel or intersecting. At this time, there are two cases: the two line segments are collinear, or the end point of one line segment is on the straight line where the other line segment is located. At this time, the following formula can be used to determine whether the bounding box of and overlaps on the x-axis and y-axis. If both formulas are satisfied, it is proved that the two line segments partially overlap, that is, the two line segments intersect. If one formula is not satisfied, it is proved that the two line segments are parallel:

[0117] Formula 1: max(min(r q (x),r q+1 (x)),min(I A (x),J l (x)))≤min(max(r q (x),r q+1 (x)),max(I A (x),J l (x)));

[0118] Formula 2: max(min(r q (y),r q+1 (y)),min(I A (y),J l (y)))≤min(max(r q (y),r q+1 (y)),max(I A (y),J l (y)))。

[0119] In addition, the embodiment of the present application further provides a city end co-distribution center site selection system based on air-ground cooperation, as shown in FIG. 4, the city end co-distribution center site selection system 400 based on air-ground cooperation comprises:

[0120] A model construction module 410 is configured to establish a planning layout set of the air-ground cooperation co-distribution center based on a demand point full coverage principle; the planning layout set comprises a customer demand point set, a candidate co-distribution center set, a co-distribution center classification set, a UAV set and a vehicle set; a air-ground cooperation layout planning model is constructed according to the planning layout set, with the minimum overall distribution time and the minimum total network operation cost as the target; constraint conditions of the air-ground cooperation layout planning model are set based on flight restriction conditions of the UAV; a co-distribution center preselected point set is determined based on the customer demand point set;

[0121] A model solution module 420 is configured to generate a site selection initial solution set for the co-distribution center preselected point set through a grey wolf swarm algorithm; invalid solutions in the site selection initial solution set are filtered according to the constraint conditions; a fitness function is constructed according to the air-ground cooperation layout planning model; the filtered site selection initial solution set is subjected to individual iterative updating, and the new solution set after each iteration is subjected to fitness evaluation and forbidden flight area discrimination, to obtain a final site selection scheme.

[0122] Each of the embodiments in the present application is described in a progressive manner, and the same or similar parts of each embodiment can be referred to each other. Each embodiment mainly describes the difference from other embodiments. Especially, for the device, equipment and nonvolatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the related parts can be referred to the part of the method embodiment.

[0123] The above describes the specific embodiments of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0124] The above only describes the embodiments of the present application and is not used to limit the present application. The embodiments of the present application can have various changes and variations for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiments of the present application shall be included in the protection scope of the present application.

Claims

1. A city end joint distribution center site selection method based on air-ground cooperation, wherein, The method comprises: Based on the demand point full coverage principle, a planning layout set of air-ground collaborative co-distribution center is established, the planning layout set comprises: a customer demand point set, a candidate co-distribution center set, a co-distribution center classification set, a UAV set and a vehicle set; Taking the minimum overall distribution time and the minimum total network operation cost as the target, an air-ground collaborative layout planning model is constructed according to the planning layout set, comprising: Based on the time of completing the distribution task by the UAV and the vehicle, a distribution time function is constructed; based on the construction cost of the co-distribution center and the distribution operation cost, a cost function is constructed, comprising: According to constructing the delivery time function; wherein a is the number of unmanned aerial vehicles in the unmanned aerial vehicle set, b is the number of vehicles in the vehicle set; u is the unmanned aerial vehicle label, k is the vehicle label; n is the number of alternative co-fulfillment centers in the alternative co-fulfillment center set, m is the number of customer demand points in the customer demand point set, i is the alternative co-fulfillment center label, and j is the customer demand point label; the flight distance of the drone u from node i to node j, is the travel distance for vehicle k from node i to node j; V u is the average delivery speed for drone u, V k is the average delivery speed for vehicle k; whether the demand point j is delivered by the UAV u from the i point, and if yes, the value is 1, otherwise, the value is 0; For whether the demand point j is distributed by the vehicle k from the i point, if yes, the value is 1, otherwise, 0; According to constructing the cost function; wherein s is a co-location center type label in the co-location center classification set; C s is the construction cost of the s-type co-location center; x is is whether the alternative landing site i constructs the s-type co-location center, and is 1 if yes, otherwise 0; For the unmanned aerial vehicle distribution center distribution operation cost; The distribution operation cost of the vehicle co-distribution center; A site selection evaluation model is constructed with the objective of minimizing the distribution time function and the cost function simultaneously: min Z = a∑Z time + b∑Z cost ; wherein Z is a site selection evaluation target value, Z time is the distribution time function, Z cost is the cost function, a is a distribution time function weight, and b is a cost function weight. Based on the flight constraint condition of the UAV, the constraint condition of the air-ground collaborative layout planning model is set, comprising: The flight constraint condition of the UAV is analyzed, and the constraint condition of the air-ground collaborative layout planning model is set according to the planning layout set; wherein the constraint condition at least comprises: take-off and landing point quantity constraint, co-distribution center type constraint, supply and demand point matching constraint, capacity demand matching constraint, no-fly zone constraint, UAV performance constraint, distribution vehicle performance constraint, service distance constraint and distribution relationship constraint; Based on the customer demand point set, a co-distribution center preselected point set is determined, comprising: In the customer demand point set, a demand point is randomly selected as the first co-distribution center preselected point and stored in the co-distribution center preselected point set; the minimum distance between each demand point in the remaining customer demand point set and the selected co-distribution center preselected point is calculated; according to the minimum distance, the probability of each demand point being selected as the next clustering center is calculated, and the demand point with the maximum probability is selected as the new co-distribution center preselected point and stored in the co-distribution center preselected point set; when the number of co-distribution center preselected points reaches the target classification number, the clustering operation is stopped, and the final co-distribution center preselected point set is obtained; An initial solution set of site selection is generated for the co-distribution center preselected point set through the grey wolf population algorithm; invalid solutions in the initial solution set of site selection are filtered according to the constraint condition; According to the air-ground collaborative layout planning model, an fitness function is constructed; The individual iterative update is performed on the filtered initial solution set of site selection, and the fitness evaluation and no-fly zone discrimination are performed on the new solution set after each iteration to obtain the final site selection scheme, comprising: The individual iterative update is performed on the filtered initial solution set of site selection through the grey wolf population algorithm; the fitness of the new solution set obtained after each iteration is calculated according to the fitness function; If there is any polygon no-fly zone in the distribution range, the direction number of any line segment in the polygon no-fly zone is calculated according to the position of the UAV distribution center and the position of the demand point in the new solution set; According to the product of the direction number and the judgment formula, it is judged whether the line connecting the position of the UAV distribution center and the position of the demand point crosses the polygon no-fly zone; if yes, the current new solution set is abandoned, and the individual iterative update is continued; When the fitness and the no-fly zone discrimination both meet the preset requirements, the new solution set at this time is determined as the final site selection scheme.

2. The urban end-to-end co-location center site selection method based on air-ground collaboration according to claim 1, wherein, The grey wolf population algorithm is used to generate a site selection initial solution set for the co-configuration center preselected point set, including: According to the address classification formula: address classification is performed on the co-configuration center preselected points in the co-configuration center preselected point set; wherein k∈n′, n′ is the total number of co-configuration center preselected points in the co-configuration center preselected point set; According to the address classification formula, a single collocation center preselected point I is randomly generated by a gray wolf swarm algorithm k The initial solution of site selection at t0: I k (t0) = rand(1, 2, 3); wherein 1 represents that the current preselected point does not construct a collocation center, 2 represents that the current preselected point constructs a UAV collocation center, and 3 represents that the current preselected point constructs a vehicle collocation center; and rand is a random operator. According to the single co-configuration center preselected point I k The initial solution of site selection at t0, get the initial solution set of site selection corresponding to the co-configuration center preselected point set: I(t0)=[I1(t0),I2(t0),…,I k (t0),…,I n (t0)] 3. The urban end-to-end co-location center site selection method based on air-ground collaboration according to claim 2, wherein, invalid solutions in the site selection initial solution set are filtered according to the constraint condition, including: whether the site selection initial solution of each co-configuration center preselected point meets the constraint condition is judged by traversing the site selection initial solution set; checking whether there is an address classified as I in the remaining sited initial solution k co-sited preselected points of (t0) = 1; If yes, randomly select a point k' satisfying I k (t0) = 1, and reassign I k′ (t0) = rand(2,3), and rejudge whether the point k' satisfies the constraint condition. If not, randomly select a point k" satisfying I k (t0) = 2, and reassign I k″ (t0) = 3, and rejudge whether the point k" satisfies the constraint condition. site selection initial solutions that do not meet the constraint condition are removed from the site selection initial solution set to filter invalid solutions in the site selection initial solution set.

4. The urban end-to-end co-location center site selection method based on air-ground collaboration according to claim 1, wherein, According to the air-ground collaborative layout planning model, a fitness function is constructed, including: According to a time fitness function is constructed; the time fitness function is a normalized value of a distribution time function; In the time fitness function, when all the co-located centers are vehicle distribution, and each demand point is distributed by the farthest co-located center, the maximum distribution time maxZ is taken time ; At this time, the minimum distribution time minZ time = 0, satisfies f time ∈(0, 1]; According to a cost fitness function is constructed; the cost fitness function is a normalized value of a cost function; In the cost fitness function, when all the co-fulfillment centers are unmanned delivery centers, and each demand point is delivered by the farthest co-fulfillment center, the maximum cost max Z cost ; at this time, the minimum cost min Z cost = 0, satisfies f cost ∈(0, 1]; Based on the target of synchronously minimizing the delivery time and the operation cost, the overall fitness function is constructed: fitness=ω t f time +ω c f cost ; wherein ω t is the time proportion coefficient, and ω c is the cost proportion coefficient.

5. A site selection system for a city end co-distribution center based on air-ground collaboration according to any one of claims 1-4, wherein, The system includes: A model construction module is configured to establish a planning layout set of air-ground collaborative co-configuration centers based on a demand point full coverage principle; the planning layout set includes a customer demand point set, a candidate co-configuration center set, a co-configuration center classification set, a UAV set, and a vehicle set; an air-ground collaborative layout planning model is constructed based on the planning layout set, with the minimum overall distribution time and the minimum total network operation cost as the target; constraint conditions of the air-ground collaborative layout planning model are set based on flight constraints of the UAV; and a co-configuration center preselected point set is determined based on the customer demand point set. A model solution module is configured to generate a site selection initial solution set for the co-configuration center preselected point set by using the grey wolf population algorithm; to filter invalid solutions in the site selection initial solution set according to the constraint condition; to construct a fitness function according to the air-ground collaborative layout planning model; and to perform individual iterative updating on the filtered site selection initial solution set, fitness evaluation on a new solution set after each iteration, and forbidden flight area discrimination, to obtain a final site selection scheme.

Images