A method for planning a transport route that prioritizes outlier service points

Abstract

The application relates to a transport route planning method for preferentially selecting an outlying service point, which comprises the following steps: determining the distance sum between each service point and its adjacent service points up to a preset number according to the distance information between all the service points in pairs, and screening out the outlying service points according to the distance sum ranking relationship; constructing adjacent edges by taking the screened out outlying service points and their adjacent service points as two ends of the adjacent edges; marking the two end service points of the adjacent edges as visited service points and / or merged service points according to the degree of the two end service points of the adjacent edges, so as to build an outlying path line; selecting any one of the unvisited service points as a starting point, and traversing all the unvisited service points respectively, so as to obtain a plurality of complete paths containing the outlying path line; and selecting the shortest path from the plurality of complete paths as an optimal complete path. The application can efficiently and accurately plan a short-distance route for a transport task in the presence of an outlying service point.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of route planning, in particular to a transport route planning method for prioritizing outlier service points. BACKGROUND

[0002] With the rapid development of economy, the trade scale of the world is also expanding, and people's demand for efficient and low-cost logistics transportation is becoming more and more strong. Saving logistics cost and shortening transportation time is the most important link in the whole logistics system. In a large-scale service node, quickly planning a transportation route is related to transportation time and fuel consumption, and transportation route planning has a direct and important influence on logistics efficiency and cost.

[0003] In the traditional method for planning the shortest non-repeated access path, the distance between each target point is mainly taken as a reference element. Therefore, the traditional method often adopts a greedy strategy, that is, the node closest to the current node is selected as the next node. The above algorithm can quickly establish a path. However, in the real world, there is often one or several sparse service nodes, and the service nodes are far away from the dense service area. The node with these characteristics is called an outlier service node.

[0004] The traditional algorithm is not suitable for handling the outlier service point, which causes the constructed path to be long. Therefore, how to efficiently and accurately plan a short route for a transportation task in the presence of an outlier service point is a problem to be solved. SUMMARY

[0005] Therefore, it is necessary to provide a transport route planning method for prioritizing outlier service points to solve the technical problem of the prior art that the outlier service point is not handled properly, causing the constructed path to be long.

[0006] The present application provides a transport route planning method for prioritizing outlier service points, comprising: obtaining position information of all logistics service points and determining distance information between all logistics service points;

[0007] According to the distance information between all logistics service points, the sum of distances between each service point and its adjacent service points up to a preset number is determined, and the outlier service points are screened out based on the ordering relationship of the sum of distances;

[0008] Each node in the outlier service point is connected with its nearest neighbor node to obtain a plurality of adjacent edges;

[0009] According to the degree of the two end nodes of the plurality of adjacent edges, the two end nodes are determined as visited nodes and / or merged nodes, and an outlier path line is constructed according to the visited nodes and / or merged nodes;

[0010] Select any one of all unvisited service nodes as a starting point, and traverse all unvisited nodes respectively to obtain a plurality of complete paths containing outlier path lines, and determine the shortest path in the plurality of complete paths as an optimal complete path.

[0011] In some possible implementation manners, the sum of distances between each service node in the service nodes and its adjacent service nodes up to a preset number is determined according to distance information between the service nodes two by two, and an outlier service node is screened out based on a ranking relationship of the sum of distances.

[0012] A preset number of adjacent nodes are selected for each node in the service nodes to form an adjacent node set.

[0013] The sum of distances of each node in the service nodes to the adjacent node set of the node is calculated to form a distance sum set.

[0014] The median of the distance sum set after ranking is taken, and the median is used to build a control filtering strength.

[0015] The outlier service node after filtering is obtained according to a size relationship between each element in the distance sum set and the control filtering strength.

[0016] In some possible implementation manners, each node in the outlier service node is connected with its nearest neighbor node to obtain a plurality of adjacent edges, including:

[0017] S201: selecting any one of the outlier service nodes as a starting node;

[0018] S202: judging the degree of the starting node, if the degree of the starting node is greater than 1, performing step S201, if the degree of the starting node is less than or equal to 1, entering S203;

[0019] S203: selecting any one of the adjacent node set of the any one node as a reaching node according to a first node probability selection formula;

[0020] S204: judging whether the starting node and the reaching node simultaneously satisfy the following conditions: no adjacent edge is constructed between the starting node and the reaching node; the degree of the reaching node is not equal to 2; the degree of the starting node and the degree of the reaching node are not simultaneously equal to 1, if simultaneously satisfying, constructing an adjacent edge between the starting node and the reaching node, and adding 1 to the degrees of the starting node and the reaching node respectively; if not simultaneously satisfying, performing step S203.

[0021] S205: judging whether all outlier service nodes have constructed adjacent edges, if yes, ending; if not, continuing to execute step S201.

[0022] In some possible implementation manners, the first node selection probability formula is:

[0023]

[0024] wherein, p ij represents a probability value of a route between service nodes v i and v j being selected, represents a neighbor node set of service node i, η ij represents an inverse of a distance between two service nodes η ij = 1 / d ij , τ ij represents cumulative experience information between two service nodes i and j, and α and β are control parameters respectively.

[0025] In some possible implementation manners, the two end nodes of the adjacent edge are determined to be visited nodes and / or merged nodes according to degrees of the two end nodes, and an outlier path line is constructed according to the visited nodes and / or the merged nodes, comprising:

[0026] S301: selecting a service node with a degree of 1 in the two end nodes of the adjacent edge as a head node, adding a service node connected to the head node to a path and taking the service node as a temporary tail node, and marking the head node and the temporary tail node as merged;

[0027] S302: judging whether a degree of the temporary tail node is equal to 2, if yes, marking the current temporary tail node as visited and merged, and taking a service node connected to the temporary tail node as a new temporary tail node, and continuing to execute step S302; if not, executing step S303;

[0028] S303: judging whether all the two end nodes of the adjacent edge are marked as merged, if yes, ending; if not, executing step S301.

[0029] In some possible implementation manners, any one of all unvisited service nodes is selected as a starting point respectively, and all unvisited nodes are traversed respectively, and a plurality of complete paths containing outlier path lines are obtained respectively, comprising:

[0030] S401: selecting any one of all unvisited service nodes as a starting point respectively, and marking the starting node as visited;

[0031] S402: obtaining a next service node of the current departure node according to a second node selection probability formula respectively, if the current next service node is an end node of an outlier path line, marking head and tail end nodes in the outlier path line as visited, adding all nodes of the outlier path line to a complete path being constructed, taking the tail end node as a departure node, otherwise marking the current next service node as visited and taking the next service node as a departure node, until all service nodes in unvisited service nodes are marked as visited;

[0032] S403: obtaining a plurality of complete paths respectively.

[0033] In some possible implementation manners, the second node selection probability formula is:

[0034]

[0035] wherein, p wu represents a probability value of a route between a service node v w and a service node v u , represents an unvisited node set of a kth virtual trolley car k , η wu represents an inverse of a distance between two nodes η wu = 1 / d wu , τ wu represents cumulative experience information between two service nodes v w and v u , and δ and ε are control parameters respectively.

[0036] In some possible implementation manners, determining a shortest path in the plurality of complete paths as an optimal complete path comprises:

[0037] attenuating cumulative experience values of edges between all the logistics service nodes;

[0038] selecting a complete path with a shortest distance from the plurality of complete paths;

[0039] enhancing cumulative experience values of edges of the shortest complete path to an experience strength.

[0040] In some possible implementation manners, cumulative experience values of the plurality of complete paths are attenuated, and an attenuated cumulative experience value formula is:

[0041]

[0042] wherein, τ ij represents cumulative experience information between two service nodes i and j, and ρ represents a historical selection information attenuation strength, τmin represents the minimum value of preset historical experience information.

[0043] In some possible implementation manners, a calculation formula of the cumulative experience value of the edge of the shortest complete path is enhanced by experience strength is as follows:

[0044]

[0045] wherein, τ xy represents the cumulative experience information representation between two service nodes x and y, L best represents the length of the shortest complete path.

[0046] The beneficial effects of the above embodiment are that the method for planning a transport route with priority to an outlying service node provided by the application divides outlying service nodes by the nearest neighbor distance and the median value coefficient method, selects neighbor nodes for the outlying service nodes in the nearest neighbor distance node set and constructs the adjacent edges therebetween, then determines two end nodes as visited nodes and / or merged nodes according to the degrees of the two end nodes of the adjacent edges, obtains an outlying path line, and finally traverses the unvisited nodes to obtain multiple complete paths, and further determines the shortest path as the optimal complete path. The application efficiently and accurately plans a shorter route for a transport task in the presence of an outlying service node. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure.

[0048] Figure 2 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure. Figure 1 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure.

[0049] Figure 3 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure. Figure 1 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure.

[0050] Figure 4 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure. Figure 1 The flowchart of one embodiment of the method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure. DETAILED DESCRIPTION

[0051] The preferred embodiments of the application are specifically described below with reference to the drawings, wherein the drawings form a part of the application and are used together with the embodiments of the application to explain the principles of the application, but are not used to limit the scope of the application.

[0052] Figure 1 The method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure. Figure 1 The method for planning a transport route with priority to an outlying service node provided by the application is shown in the figure.

[0053] S101, acquiring all logistics service node position information and determining distance information between all logistics service nodes;

[0054] S102, determining the sum of distances between each service node and its adjacent service nodes up to a preset number of service nodes according to the distance information between all logistics service nodes, and screening out outlier service nodes based on the ordering relationship of the sum of distances;

[0055] S103, connecting each node in the outlier service nodes with its nearest neighbor node to obtain a plurality of adjacent edges;

[0056] S104, determining the two end nodes of the adjacent edges as visited nodes and / or merged nodes according to the degree of the two end nodes, and constructing an outlier path line according to the visited nodes and / or merged nodes;

[0057] S105, selecting any one of all unvisited service nodes as a starting point, respectively, and traversing all unvisited nodes to obtain a plurality of complete paths containing outlier path lines, respectively, and determining the shortest path in the plurality of complete paths as the optimal complete path.

[0058] Compared with the prior art, the transport route planning method for prioritizing outlier service points provided by the embodiment divides outlier service nodes by the nearest neighbor distance and median value coefficient method, selects adjacent nodes for the outlier service nodes in the nearest neighbor distance node set and constructs adjacent edges therebetween, then determines the two end nodes of the adjacent edges as visited nodes and / or merged nodes to obtain an outlier path line, and finally traverses unvisited nodes to obtain a plurality of complete paths, and then determines the shortest path as the best complete path. The application efficiently and accurately plans a shorter route for a transport task in the presence of outlier service nodes.

[0059] In the step S101, the position information data of all service nodes can be collected, the distance between two service nodes can be calculated, and the calculated distance result can be expressed in a matrix A d storage. The expression of the matrix A d is as follows:

[0060]

[0061] In the formula, d ij represents the distance between two service nodes and, and n represents the number of service nodes.

[0062] In practice, the path walked according to experience obtains a cumulative experience value, and the cumulative experience value of each edge between service nodes is stored in a matrix τ MIn the formula, τ M The expression is as follows:

[0063]

[0064] In the formula, τ ij represents the cumulative experience value of two service nodes v i and v j , and the initial experience value of τ ij is set as τ0.

[0065] It should be noted that there are some remote outlier service nodes in the transport task.

[0066] In order to divide the remote outlier service nodes, in some embodiments of the present application, in step S102, the sum of distances between each service node and its adjacent service nodes up to a preset number is determined according to the distance information between all the logistics service nodes, and the outlier service nodes are screened based on the ordering relationship of the sum of distances, including:

[0067] A preset number of adjacent nodes are selected for each node in all the logistics service nodes to form a set of adjacent nodes;

[0068] The sum of distances from each node in all the logistics service nodes to its set of adjacent nodes is calculated to form a set of distance sums;

[0069] The median of the sorted set of distance sums is taken to build a control screening intensity;

[0070] The screened outlier service nodes are obtained according to the size relationship between each element in the set of distance sums and the control screening intensity.

[0071] Specifically, according to the distance matrix A d , n i nearest neighbor nodes are selected for each node v closet as a candidate queue, and the n closet nearest neighbor nodes are stored in the set , and the nearest neighbor distance sum f i of the service node v i is calculated, as follows:

[0072]

[0073] In the formula, i∈[1,n],

[0074] Further, the f values of all the service nodes are sorted, the median M of the sorted values is taken out, and the outlier service nodes are selected according to the following formula,

[0075] f i ≥ M·γ

[0076] wherein γ is a parameter for controlling the screening intensity, γ≥1. If the nearest neighbor distance of node i and f i ≥ M·γ, node i is regarded as a candidate outlier service node.

[0077] Further, the number of outlier service nodes is controlled according to the following formula,

[0078] 0|U outlier |≤ ω·n

[0079] wherein U outlier represents the set of outlier service nodes, and ω represents a control parameter for the maximum percentage of outlier service nodes in the total number of target points.

[0080] It should be noted that the degree of a node represents the number of edges connected to each node. In order to reduce the influence of outlier service nodes on the route, in some embodiments of the present application, in step S103, each node in the outlier service nodes is connected with its nearest neighbor node to obtain a plurality of adjacent edges, including:

[0081] S201: selecting an arbitrary node from the outlier service nodes as a starting node;

[0082] S202: judging the degree of the starting node, if the degree of the starting node is greater than 1, executing step S201, if the degree of the starting node is less than or equal to 1, entering S203.

[0083] S203: selecting an arbitrary node from the set of adjacent nodes of the arbitrary node as a destination node according to a first node selection probability formula;

[0084] S204: judging whether the starting node and the destination node simultaneously satisfy the following conditions: no adjacent edge is constructed between the starting node and the destination node; the degree of the destination node is not equal to 2; the degree of the starting node and the degree of the destination node are not simultaneously equal to 1, if simultaneously satisfied, constructing an adjacent edge between the starting node and the destination node, and adding 1 to the degrees of the starting node and the destination node respectively; if not simultaneously satisfied, executing step S203;

[0085] S205: judging whether all outlier service nodes have constructed adjacent edges, if yes, ending; if not, continuing to execute step S201.

[0086] In order to give a larger selection probability to the path with a short distance and a strong accumulated experience value. In some embodiments of the present application, the first node selection probability formula is:

[0087]

[0088] where p ij denotes the probability value of the route selected between service nodes v i and service nodes v j , denotes the neighbor set of service node i, η ij denotes the inverse of the distance between two service nodes η ij = 1 / d ij , τ ij denotes the cumulative experience information between two service nodes i and j, and α and β are control parameters respectively. The values of α and β can be determined according to actual conditions.

[0089] In order to optimize the out-of-cluster path line constructed for the out-of-cluster service node, the out-of-cluster node path is merged. In some embodiments of the present application, in step S104, the two end nodes of the adjacent edge are determined to be visited nodes and / or merged nodes according to the degrees of the two end nodes, and the out-of-cluster path line is constructed according to the visited nodes and / or merged nodes, including:

[0090] S301: selecting a service node with a degree of 1 in the two end nodes of the adjacent edge as a head node, adding a service node connected to the head node to the path and taking the service node as a temporary tail node, and marking the head node and the temporary tail node as merged;

[0091] S302: judging whether the degree of the temporary tail node is equal to 2, if the degree of the temporary tail node is equal to 2, marking the current temporary tail node as visited and merged, and taking a service node connected to the temporary tail node as a new temporary tail node, and continuing to execute step S302; if the degree of the temporary tail node is not equal to 2, executing step S303;

[0092] S303: judging whether all the two end nodes of the adjacent edge are marked as merged, if all the two end nodes of the adjacent edge are marked as merged, ending; otherwise, executing step S301.

[0093] It should be noted that the principle of the second node selection probability formula is the same as that of the first node selection probability formula, and only the range of the set is different. In the first node selection probability formula, the value range of the arrival node is the neighbor set of the starting node, and in the second node selection probability formula, the value range of the arrival node is the node in the unvisited node. In some embodiments of the present application, in step S105, any one of all unvisited service nodes is selected as a starting point, and all unvisited nodes are traversed respectively to obtain a plurality of complete paths containing out-of-cluster path lines, and a plurality of complete paths are obtained respectively, including:

[0094] S401: Select any one of all unvisited service nodes as a starting point respectively, and mark the starting node as visited;

[0095] S402: Obtain a next service node of the starting node according to a second node selection probability formula respectively, if the next service node is an end node of an outlier path line, mark the head and tail end nodes in the outlier path line as visited, add all nodes in the outlier path line to a complete path being constructed, and take the tail end node as a starting node, otherwise mark the next service node as visited and take the next service node as a starting node, until all service nodes in all unvisited service nodes are marked as visited;

[0096] S403: Obtain a plurality of complete paths respectively.

[0097] In specific embodiments of the present application, a plurality of virtual trolleys are used to construct a global path.

[0098] In some embodiments of the present application, the second node selection probability formula is:

[0099]

[0100] In the formula, p wu represents a probability value of a route between service nodes v w and v u being selected, represents an unvisited node set of the kth virtual trolley car k , η wu represents an inverse distance between two nodes η wu = 1 / d wu , τ wu represents cumulative experience information between two service nodes v w and v u , and δ and ε are control parameters respectively. The values of δ and ε can be determined according to actual conditions.

[0101] In some embodiments of the present application, the shortest path in the plurality of complete paths is determined as an optimal complete path, comprising:

[0102] Decaying the cumulative experience value of an edge between all logistics service nodes;

[0103] Selecting a complete path with the shortest distance from the plurality of complete paths;

[0104] Enhancing the cumulative experience value of an edge of the shortest complete path to an experience strength.

[0105] In order to avoid falling into a local optimal solution, in some embodiments of the present application, the accumulated experience value of the plurality of complete paths is decayed, and the decayed accumulated experience value is calculated according to the following formula:

[0106]

[0107] In the formula, τ ij represents the accumulated experience information between two service nodes i and j, and ρ represents the decay strength of the historical selection information, and τ min represents the preset minimum value of the historical experience information.

[0108] In some embodiments of the present application, the calculation formula of the accumulated experience value of the edge of the shortest complete path is enhanced by the experience strength, and is calculated according to the following formula:

[0109]

[0110] In the formula, τ xy represents the accumulated experience information between two service nodes x and y, and L best represents the length of the shortest complete path.

[0111] Embodiment 2:

[0112] According to the technical solution in embodiment 1, simulation experiments are performed on service node numbers n = 101, 198, 225, 280, 318, 417, 439, 442, 493, 654, 1323, 1400 in the TSPLIB test set. In the experiment, the number of virtual cars ncar = 100; the control parameters α = 1.0, β = 3.0, γ = 2.0, ω = 0.1; the initial path experience value τ0 = 1.0; the lower limit of the experience value τmin = 1.0e-8, the decay parameter ρ = 0.5; the number of nearest neighbor nodes ncloset = 8; the number of iterations Max_Iteration = 600;

[0113] 100 independent experiments are performed on each test case in the present embodiment, and the best result, the worst result, the average result and the standard deviation are calculated. The number after the name of each test case represents the number of service nodes, and the experimental results are shown in Table 1:

[0114]

[0115]

[0116] As can be seen from Table 1, the technical solution in embodiment 1 can effectively achieve the effect of the present technical solution.

[0117] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiment methods can be completed by instructing the relevant hardware by a computer program, and the program can be stored in a computer readable storage medium. The computer readable storage medium is a disk, an optical disk, a read-only memory, a random access memory, etc.

[0118] The above description is merely preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A transport route planning method that prioritizes outlier service points, characterized by, The method comprises the following steps: acquiring all logistics service node position information and determining distance information between all logistics service nodes; determining the sum of distances between each service node and its adjacent service nodes up to a preset number of service nodes based on the distance information between all logistics service nodes, and screening out outlier service nodes based on the ordering relationship of the sum of distances; connecting each of the outlier service nodes with its nearest neighbor node to obtain a plurality of adjacent edges; determining the two end nodes of the adjacent edges as visited nodes and / or merged nodes based on the degree of the two end nodes, and constructing outlier path lines based on the visited nodes and / or merged nodes; selecting any one of all unvisited service nodes as a starting point, respectively, and traversing all unvisited nodes and all outlier path lines to obtain a plurality of complete paths, respectively, and determining the shortest path in the plurality of complete paths as an optimal complete path; determining the two end nodes of the adjacent edges as visited nodes and / or merged nodes based on the degree of the two end nodes, and constructing outlier path lines based on the visited nodes and / or merged nodes, comprising: S301: selecting a service node with a degree of 1 among the two end nodes of all adjacent edges as a head node, adding the service node connected to the head node to the path and taking the service node as a temporary tail node, and marking the head node and the temporary tail node as merged; S302: determining whether the degree of the temporary tail node is equal to 2, if the degree of the temporary tail node is equal to 2, marking the current temporary tail node as visited, and taking the service node connected to the temporary tail node as a new temporary tail node, and continuing to execute step S302; if the degree of the temporary tail node is not equal to 2, executing step S303; S303: determining whether all adjacent edge end nodes are marked as merged, if all adjacent edge end nodes are marked as merged, ending; otherwise, executing step S301.

2. The method of claim 1, wherein, determining the sum of distances between each service node and its adjacent service nodes up to a preset number of service nodes based on the distance information between all logistics service nodes, and screening out outlier service points based on the ordering relationship of the sum of distances, comprising: selecting a preset number of adjacent nodes for each node in all logistics service nodes to form an adjacent node set; calculating the sum of distances from each node in all logistics service nodes to its adjacent node set to form a distance sum set; taking the median of the sorted distance sum set to construct a control filtering strength; obtaining the filtered outlier service nodes based on the size relationship between each element in the distance sum set and the control filtering strength.

3. The method of claim 1, wherein, connecting each of the outlier service nodes with its nearest neighbor node to obtain a plurality of adjacent edges, comprising: S201: selecting any one node from the outlier service nodes as a starting node; S202: determining the degree of the starting node, if the degree of the starting node is greater than 1, executing step S201, if the degree of the starting node is less than or equal to 1, entering S203; S203: selecting an arbitrary node from the neighbor node set of the arbitrary node as a destination node according to a first node selection probability formula; S204: judging whether the departure node and the destination node satisfy the following conditions simultaneously: no adjacent edge is constructed between the departure node and the destination node; the degree of the destination node is not equal to 2; the degrees of the departure node and the destination node are not equal to 1 simultaneously, if the conditions are satisfied simultaneously, constructing an adjacent edge between the departure node and the destination node, and adding 1 to the degrees of the departure node and the destination node respectively, if the conditions are not satisfied simultaneously, executing step S203; S205: judging whether all the outlier service nodes have constructed adjacent edges, if the adjacent edges are constructed, ending; if the adjacent edges are not constructed, continuing to execute step S201.

4. The method of claim 3, wherein, The first node selection probability formula is: wherein p ij denotes a service node v i and a service node v j between which a route is selected, denotes a service node i adjacent to the service node η ij denotes the inverse of the distance between two service nodes η ij =1 / d ij , denotes the accumulated experience information between two service nodes i and j , α and β are control parameters, respectively.

5. The method of claim 1, wherein, Selecting an arbitrary node from all the unvisited service nodes as a starting point respectively, and traversing all the unvisited nodes and the all-outlier path lines respectively, a plurality of complete paths are obtained, including: S401: selecting an arbitrary node from all the unvisited service nodes as a departure node respectively, and marking the departure node as visited; S402: obtaining the next service node of the current starting point according to a second node selection probability formula respectively, and marking the next service node as visited and as a departure node until all the service nodes in the unvisited service nodes are marked as visited; S403: obtaining a plurality of complete paths respectively.

6. The method of claim 5, wherein the method further comprises: The second node selection probability formula is: wherein denotes a service node v w and a service node v u between which a route is selected, denotes the set of unvisited nodes of the k virtual trolley car k , η wu denotes the inverse of the distance between two nodes η wu =1 / d wu , denotes the cumulative empirical information between two service nodes v w and v u , δ and ε are control parameters, respectively.

7. The method of claim 1, wherein, Determining the shortest path in the plurality of complete paths as an optimal complete path, including: Decaying the cumulative experience values of the plurality of complete paths; Selecting the shortest complete path from the plurality of complete paths after the decay; Enhancing the cumulative experience values of the edges of the shortest complete path to experience strength, and obtaining an optimal complete path.

8. The method of claim 7, wherein, The decayed cumulative experience value formula of the plurality of complete paths is: wherein represents accumulated experience information between two service nodes i and j represents accumulated experience information between two service nodes ρ represents a history selection information decay strength, represents a preset minimum value of the history experience information.

9. The method of claim 7, wherein the priority outlier service point is a service point that is not on the route. The calculation formula of the cumulative experience values of the edges of the shortest complete path to experience strength is: wherein denotes the accumulated experience information between two service nodes x and y denotes the accumulated experience information between two service nodes L best denotes the length of the shortest complete path.

Images