A shared bicycle recycling and allocation method

Abstract

The present application provides a shared bicycle recycling and allocation method, which belongs to the field of shared bicycle technology. The embodiment of the present application aims to minimize the total cost, and constructs an exponentially initialized uniform estimation distribution algorithm based on the estimation distribution algorithm to iteratively solve the mathematical model, and finally outputs the optimal solution of the shared bicycle recycling and allocation model. In the process of iterative solution in the embodiment of the present application, the same truck simultaneously recycles damaged bicycles and rebalances normal bicycles, which improves the utilization rate of the truck capacity, effectively reduces the cost brought by the recycling and allocation process, and makes the recycling and rebalancing plan formulated based on the model more economical; the embodiment of the present application provides the location of the recycling center, the number of trucks, the truck route planning, and the decision of the load distribution of the demand point, so that the established model can more truly reflect the actual situation of shared bicycle recycling and rebalancing, and thus the recycling and allocation plan formulated based on the model is more reliable.

Description

Technical Field

[0001] The present application relates to the field of shared bicycles, and in particular to a shared bicycle recycling and distribution method. Background Art

[0002] As an important part of the sharing economy, the shared bicycle system has broad market prospects and significant prospects for rapid development. However, with the development of the shared bicycle system, two major problems, namely, damaged bicycles and unbalanced distribution of available bicycles, have begun to emerge. On the one hand, shared bicycles will be damaged during riding. If there are too many damaged bicycles that cannot be unlocked, the needs of riding customers may not be met, and the broken bicycles scattered on the road will also leave a bad impression on citizens and tourists. On the other hand, due to the different riding needs in different locations, as time goes by, the available bicycles tend to be concentrated at the riding destination, and the available bicycles at the riding starting point will become fewer and fewer. Without an appropriate rebalancing strategy, even if there are enough functional bicycles, the needs of users cannot be met.

[0003] Although shared bicycle platform operators have been recycling damaged bicycles and rebalancing shared bicycles for a long time, due to the high recycling cost and the lack of efficient recycling distribution strategy, there are still a large number of damaged bicycles on the streets, and the uneven distribution of bicycles also occurs from time to time. Therefore, the research on resource recycling and distribution of shared bicycle systems is of great significance. Summary of the invention

[0004] The present application provides a shared bicycle recycling and distribution method to solve the problem of recycling damaged bicycles and the problem of rebalancing shared bicycles.

[0005] In order to solve the above problems, this application adopts the following technical solutions:

[0006] The present application provides a method, which includes:

[0007] Step S1: Establishing a shared bicycle recycling allocation model, the shared bicycle recycling allocation model includes decision factors, an objective function of the total cost and constraints; wherein the decision factors include the location distribution of the recycling center, the number distribution of trucks, the truck route planning and the load demand distribution of the demand point; the total cost includes the fixed cost of the recycling center, the fixed cost of the truck, the transportation cost and the income brought by meeting the load demand distribution of the demand point;

[0008] Step S2: Based on the estimated distribution algorithm, an exponentially initialized uniform estimated distribution algorithm is constructed to solve the shared bicycle recycling allocation model and obtain a feasible solution. Among them, the feasible solution It includes the number of recycling centers, the locations of recycling centers, the number of trucks, the path information of each truck, and the load allocated to each demand point, wherein the load allocated to the demand point includes the number of damaged bicycles at the demand point, the number of normal bicycles missing at the demand point, and the number of normal bicycles in excess at the demand point;

[0009] Step S3: Repeat step S2 to obtain a feasible solution set Among them, m represents the iteration number, Represents all feasible solutions obtained after the mth iteration;

[0010] Step S4: When the number of feasible solutions reaches a threshold, the feasible solution set is sorted in ascending order according to the objective function. Among all feasible solutions in , the top s% of feasible solutions are selected to form the set of surviving solutions. And the surviving solutions are The feasible solution that minimizes the objective function is taken as the optimal solution.

[0011] Step S5: According to the surviving solution set The average number of trucks in all feasible solutions to generate a distribution of trucks According to the surviving solution set The frequency with which trucks are assigned to recycling centers generates the distribution of recycling centers According to the surviving solution set The frequency of each demand point being visited in the distribution of route planning is generated According to the surviving solution set The load information at each demand point in the , generates the load distribution corresponding to each demand point

[0012] Step S6: Output the optimal solution of the shared bicycle recycling allocation model Among them, the optimal solution Including distribution of trucks Location of recycling centers Distribution of bicycle positions And the load distribution corresponding to the demand point

[0013] In an embodiment of the present application, the specific steps of step S1 include:

[0014] Step S1-1: The truck starts from the recycling center k, and when passing through each demand point, it takes away the damaged bicycles at the demand point and / or takes away the excess normal bicycles and / or brings the missing normal bicycles, until the total load of the damaged bicycles on the truck reaches the threshold or the overall rebalancing demand reaches the upper limit, and then returns to the recycling center k; wherein the overall rebalancing demand reaches the upper limit means that the excess normal bicycles of all demand points are transferred to the demand points where the normal bicycles are missing;

[0015] Step S1-2: Define the following variables:

[0016]

[0017]

[0018]

[0019] Step S1-3: Construct the objective function of the total cost:

[0020]

[0021] The goal is to minimize the total cost; where f k represents the fixed cost of the recycling center; g n represents the fixed cost of truck n; t ij Represents the distance between demand point i and demand point j; express or in represents the profit from recycling a damaged bicycle at demand point i, represents the profit brought by bringing a missing normal bicycle to the demand point i, represents the profit brought by taking away a redundant normal bicycle at demand point i; express or in, represents the number of damaged bicycles that need to be recovered by truck n at demand point i, represents the number of normal bicycles that need to be brought to the demand point i by truck n, Indicates the number of normal bicycles that need to be taken away by truck n at demand point i;

[0022] Step S1-4: Establish the following constraints:

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041] In the formula, I represents the set of demand points, N represents the set of trucks, K represents the set of recycling centers, and c n represents the capacity of truck n, M is a preset constant, represents the number of damaged bicycles at demand point i, represents the number of normal bicycles missing at demand point i, represents the normal number of excess bicycles at demand point i, represents the load of the damaged single truck of truck n at demand point i, represents the load of truck n on the rebalancing vehicle at demand point i, represents the load change of the damaged single truck of truck n from demand point i to demand point j, represents the load change of truck n when rebalancing from demand point i to demand point j, u inEliminate sub-loop redundant variables; where, formula (4) restricts trucks to be assigned only to open recycling centers; formula (5) ensures that each truck can be assigned to the corresponding recycling center at most once; formula (6) restricts designated trucks to travel between demand points; formula (7) controls that trucks cannot take away bicycles that exceed demand; formulas (8) and (9) restrict trucks from recovering damaged bicycles or rebalancing normal bicycles if they do not pass through demand points; formulas (10) and (11) eliminate sub-loops, prevent trucks from passing through the same demand point multiple times, and ensure that the paths of all trucks include their corresponding recycling centers; formula (12) Formula (14) and (15) restrict the starting and ending points of the trucks to the designated recycling centers; Formula (16) is a flow balance constraint, which is used to associate the paths of the trucks to ensure that the trucks entering a demand point will exit from the demand point; Formula (17) indicates that the trucks cannot return to the recycling center before completing the rebalancing of the normal single vehicle; Formula (18) restricts the load of the trucks to not exceed their capacity; Formulas (19) and (20) define the load changes of the trucks at the demand points; Formula (21) defines the load changes of each truck between the demand points; Formula (22) restricts the value range of the variables in the formula.

[0042] In an embodiment of the present application, the method further includes:

[0043] Follow these steps to generate an initial solution to the shared bike recycling allocation model:

[0044] Step SA-1: determining the joint cost of each recycling center according to the fixed cost of the recycling center, the positional relationship between the recycling center and the demand point, and the load demand of the demand point, and arranging the recycling centers in ascending order according to the size of the joint cost;

[0045] Step SA-2: Set the probability of assigning the truck to recycling center k to be Generate an initial solution for the recycling center Among them, k represents the kth recycling center in ascending order;

[0046] Step SA-3: Determine the number of trucks based on the total demand of the demand point and generate an initial solution for the trucks. The total demand refers to the total number of damaged bicycles that need to be recycled at all demand points;

[0047] Step SA-4: Based on the initial solution of the recycling center Assign trucks to the recycling center and set out from the recycling center. Set each truck to have a 50% probability of passing through each demand point. Each time the truck passes through a demand point, a uniformly distributed load is allocated to the truck. After the truck returns to the recycling center, delete the demand points that the truck has passed but has not been allocated a load.

[0048] Step SA-5: Generate an initial solution for route planning based on the route planning of all trucks Generate an initial solution of the load demand at the demand point based on the loads of all trucks

[0049] In one embodiment of the present application, in step S2, the path information of each truck is obtained by the following steps:

[0050] Step S2-1: Determine whether the total load of the damaged bicycles on the truck has reached a threshold or whether the overall rebalancing demand has reached an upper limit. If so, proceed to step S2-5; otherwise, proceed to step S2-2;

[0051] Step S2-2: Determine whether there is a redundant normal bicycle on the truck, if so, proceed to step S2-3; otherwise, proceed to step S2-4;

[0052] Step S2-3: Generate a first random number between [0, 1) by using a first random function; when the first random number belongs to [0, 0.7), go to the demand point where normal bicycles are missing, and distribute the bicycles according to the route planning. Generate the next path; when the first random number belongs to [0.7, 1), go to the demand point with normal bicycle surplus, and according to the distribution of route planning Generate the next path;

[0053] Step S2-4: The truck randomly goes to a demand point and distributes according to the route planning. Generate the next path;

[0054] Step S2-5: Determine whether there are redundant normal bicycles on the truck; if so, go to the demand point where the normal bicycle is missing and distribute it according to the route planning. Generate the next path; otherwise, return the corresponding recycling center.

[0055] In one embodiment of the present application, in step S2, the load allocated to each truck at each demand point is determined by the following steps:

[0056] Step S2-6: Generate a second random number between [0, 1) by a second random function, and determine whether the second random number belongs to [0, 0.7). If so, proceed to step S2-7; otherwise, proceed to step S2-8;

[0057] Step S2-7: Load distribution corresponding to the current demand point Allocate the damaged bicycle corresponding to the current demand point to the truck;

[0058] Step S2-8: Determine whether the current demand point is a demand point with normal bicycle surplus. If so, proceed to step S2-9; otherwise, proceed to step S2-10:

[0059] Step S2-9: Load distribution of excess normal bicycles corresponding to the current demand point Allocate normal single vehicles that are in excess of the current demand point to the truck;

[0060] Step S2-10: Load distribution of missing normal bicycles corresponding to the current demand point Allocate the normal single vehicle missing from the current demand point to the truck;

[0061] Step S2-11: Load distribution of damaged bicycles corresponding to the current demand point Assign the damaged bicycle corresponding to the current demand point to the truck.

[0062] In an embodiment of the present application, between step S5 and step S6, the method further includes:

[0063] Step SE: Determine whether the iteration number m is less than or equal to the maximum iteration number L and whether the fluctuation of five consecutive iterations is greater than or equal to the preset minimum fluctuation rate. If so, proceed to step S3; otherwise, proceed to step S6.

[0064] Compared with the prior art, this application has the following advantages:

[0065] The embodiment of the present application transforms the recycling and rebalancing problem of shared bicycles into a shared bicycle recycling allocation model, which provides decisions on the location of the recycling center, the number of trucks, truck route planning, and load distribution at the demand point, so that the established model can more truly reflect the actual situation of shared bicycle recycling and rebalancing, and thus the recycling allocation plan formulated based on the model is more reliable; the embodiment of the present application takes minimizing the total cost as the goal, and constructs an exponentially initialized uniform estimation distribution algorithm based on the estimation distribution algorithm to iteratively solve the mathematical model. During the solution process, the same truck can simultaneously recycle damaged bicycles and rebalance normal bicycles, thereby improving the utilization rate of the truck capacity, and while recycling more damaged bicycles, effectively reducing the cost of recycling damaged bicycles and rebalancing normal bicycles, making the recycling and rebalancing plan formulated based on the model more economical. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0067] Figure 1 This is a flowchart of the steps of a shared bicycle recycling and allocation method in an embodiment of the present application;

[0068] Figure 2 This is a schematic diagram of a shared bicycle coordination network in an embodiment of the present application;

[0069] Figure 3 is a schematic diagram of performance analysis of different demand levels in an embodiment of the present application;

[0070] Figure 4 is a schematic diagram of performance analysis of different road network capacity levels in an embodiment of the present application;

[0071] Figure 5 2 is a schematic diagram of performance analysis of different rebalancing priorities in an embodiment of the present application. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0073] Reference Figure 1 , showing a shared bicycle recycling and allocation method of the present application, the method may include the following steps:

[0074] Step S1: Establish a shared bicycle recycling allocation model, which includes decision factors, objective functions of total cost and constraints; the decision factors include the location distribution of recycling centers, the number distribution of trucks, truck route planning and the load demand distribution of demand points; the total cost includes the fixed cost of the recycling center, the fixed cost of the truck, the transportation cost and the income brought by meeting the load demand distribution of the demand points.

[0075] It should be noted that this implementation is based on the following assumptions:

[0076] (1) Trucks must depart and return to their designated recycling centers, i.e., the closed-loop assumption of the truck routing;

[0077] (2) A truck can pass through the same demand location at most once, that is, sub-loops are prohibited on the truck path; at the same time, different trucks are allowed to visit the same demand location so that the load demand of the demand point can be effectively met;

[0078] (3) Temporary inventory is not allowed, that is, the recovery of damaged bicycles and the rebalancing of normal bicycles of each truck are independent of each other, and the situation where multiple trucks transit and deliver through a certain demand point is not considered;

[0079] (4) No available normal bicycles can be taken back to the recycling center to ensure that the number of available normal bicycles in the shared bicycle system does not change.

[0080] Based on the above assumptions (1) to (4), a shared bicycle recycling allocation model is established, which can specifically include the following steps:

[0081] Step S1-1: The truck starts from the recycling center k, and when passing through each demand point, it takes away the damaged bicycles at the demand point and / or takes away the excess normal bicycles and / or brings the missing normal bicycles, until the total load of the damaged bicycles on the truck reaches a threshold or the overall rebalancing demand reaches an upper limit, and then returns to the recycling center k; wherein, the overall rebalancing demand reaches an upper limit means that the excess normal bicycles of all demand points are transferred to the demand points where normal bicycles are missing.

[0082] In this embodiment, the recovery of damaged bikes and the rebalancing of normal bikes are performed simultaneously. Specifically, in the shared bike recovery allocation model, the locations of several recovery centers and the number of trucks allocated to each recovery center are first determined; then, the route of each truck is determined, on which the truck will recover damaged bikes and rebalance normal bikes at each demand point. It should be noted that the rebalancing of normal bikes refers to transporting normal bikes that are in excess at the first demand point to the second demand point where normal bikes are missing.

[0083] It should be noted that in order to meet the needs of cycling users, the truck needs to transport all normal bicycles on the truck to the demand point where normal bicycles are in short supply before returning to the recycling center. Figure 2, shows a schematic diagram of the shared bicycle coordination network in an embodiment of the present application. In the figure, circles represent demand points, triangles represent recycling centers, different arrows represent the driving paths of different trucks, and the numbers in brackets represent the recycling or rebalancing needs met by the corresponding truck at the demand point (representing the number of damaged bicycles recovered by the truck at the demand point, the number of missing normal bicycles brought to the demand point, and the number of excess normal bicycles taken away for the demand point, respectively). It should be noted that "surplus" is a concept relative to the demand point, and "redundancy" is a concept relative to the truck, that is, after the "surplus" normal bicycles at the demand point enter the truck, these "surplus" normal bicycles at the demand point are "redundant" for the truck. Taking Truck No. 2 as an example, when it drives to the first demand point (this demand point is a demand point with excess normal bicycles) according to the route of Truck No. 2, the truck takes away the 12 excess normal bicycles at the first demand point and drives to the second demand point (this demand point is a demand point with excess normal bicycles), and takes away 58 damaged bicycles and 3 excess normal bicycles at the second demand point (at this time, for Truck No. 2, there are a total of 58 damaged bicycles and 15 redundant normal bicycles on the vehicle), and then continues to drive to the third demand point (this demand point is a demand point where normal bicycles are missing), takes away 37 damaged bicycles and puts down 15 normal bicycles at the third demand point. At this time, there are no redundant bicycles on the truck and a total of 95 damaged bicycles have been collected, that is, the recovery of damaged bicycles and the rebalancing of normal bicycles are completed at the same time, and finally returns to the recycling center.

[0084] It should be further explained in this embodiment that although the crossing route of Truck No. 2 is not the shortest and most preferred in the traditional vehicle routing problem, it is reasonable and optimal in the shared bicycle recycling allocation model because the truck's route must satisfy the rebalancing constraint, that is, all redundant normal bicycles on the truck must be transported before returning to the recycling center. On this basis, the objective function of the total cost is established, and decisions are made with the goal of minimizing the objective function, so that the recycling rebalancing plan formulated based on this model is more economical.

[0085] Step S1-2: Define the following variables:

[0086]

[0087]

[0088]

[0089] Step S1-3: Construct the objective function of the total cost:

[0090]

[0091] The goal is to minimize the total cost; where f krepresents the fixed cost of the recycling center; g n represents the fixed cost of truck n; t ij Represents the distance between demand point i and demand point j; express or in represents the profit from recycling a damaged bicycle at demand point i, represents the profit brought by bringing a missing normal bicycle to the demand point i, represents the profit brought by taking away a redundant normal bicycle at demand point i; express or in, represents the number of damaged bicycles that need to be recovered by truck n at demand point i, represents the number of normal bicycles that need to be brought to the demand point i by truck n, It represents the number of normal bicycles that need to be taken away by truck n at demand point i.

[0092] Step S1-4: Establish the following constraints:

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] In the formula, I represents the set of demand points, N represents the set of trucks, K represents the set of recycling centers, and c n represents the capacity of truck n, M is a preset constant, represents the number of damaged bicycles at demand point i, represents the number of normal bicycles missing at demand point i, represents the normal number of excess bicycles at demand point i, represents the load of the damaged single truck of truck n at demand point i, represents the load of truck n on the rebalancing vehicle at demand point i, represents the load change of the damaged single truck of truck n from demand point i to demand point j, represents the load change of truck n when rebalancing from demand point i to demand point j, u in Eliminate sub-loop redundant variables; where, formula (4) restricts trucks to be assigned only to open recycling centers; formula (5) ensures that each truck can be assigned to the corresponding recycling center at most once; formula (6) restricts designated trucks to travel between demand points; formula (7) controls that trucks cannot take away bicycles that exceed demand; formulas (8) and (9) restrict trucks from recovering damaged bicycles or rebalancing normal bicycles if they do not pass through demand points; formulas (10) and (11) eliminate sub-loops, prevent trucks from passing through the same demand point multiple times, and ensure that the paths of all trucks include their corresponding recycling centers; formula (12) Formula (14) is a flow balance constraint, which is used to associate the paths of trucks to ensure that a truck entering a demand point will leave from that demand point; Formula (15) indicates that the truck cannot return to the recycling center before completing the rebalancing of the normal single vehicle; Formula (16) limits the load of the truck to not exceed its capacity; Formulas (17) and (18) define the load variation of the truck at the demand point; Formulas (19) and (20) define the load variation of each truck between demand points; Formula (21) limits the range of the variables in the formula.

[0112] In this embodiment, in order to enable the shared bicycle recycling allocation model established based on the above constraints to be directly solved by a commercial solver such as the CPLEX solver, constraints (17), (18) and (20) are linearized. Among them, constraint (17) can be linearized into constraints (22), (23) and (24):

[0113]

[0114]

[0115]

[0116] Constraint (18) can be linearized into constraints (25), (26) and (27):

[0117]

[0118]

[0119]

[0120] Constraint (20) can be linearized into constraints (28) and (29):

[0121]

[0122]

[0123] Step S2: Based on the estimated distribution algorithm, an exponentially initialized uniform estimated distribution algorithm is constructed to solve the shared bicycle recycling allocation model and obtain a feasible solution. Among them, the feasible solution It includes the number of recycling centers, the locations of recycling centers, the number of trucks, the route information of each truck, and the load allocated to each demand point. The load allocated to the demand point includes the number of damaged bicycles at the demand point, the number of normal bicycles missing at the demand point, and the number of normal bicycles in excess at the demand point.

[0124] It should be noted that the Estimation of Distribution Algorithm (EDA) is a meta-heuristic algorithm. The EDA algorithm gradually improves the estimated distribution based on the characteristics of the selected dominant solution, while ensuring that the convergence direction points to the optimal solution, while still retaining the possibility of global search. At the same time, it does not store the information of each solution, which greatly reduces storage and CPU time.

[0125] In this embodiment, although a solution can be generated according to the estimated distribution based on the traditional EDA algorithm, the solution generation efficiency is low, and there is a possibility of infeasible solutions that violate the constraints, such as violating the rebalancing constraint condition formula (15) and the capacity constraint condition formula (16). For this reason, this embodiment constructs an exponentially initialized uniform estimation distribution algorithm (EIUEDA, Exponential-Initialization Uniform Estimation of Distribution Algorithm) based on the EDA algorithm. In the EIUEDA algorithm, an improved path selection algorithm and an improved load distribution algorithm are designed to adjust the potential infeasible solutions generated by the estimated distribution so that they can efficiently generate feasible solutions.

[0126] In one feasible implementation, in step S2, the path information of each truck is obtained by the following path selection algorithm:

[0127] Step S2-1: Determine whether the total load of the damaged bicycles on the truck has reached a threshold or whether the overall rebalancing demand has reached an upper limit. If so, proceed to step S2-5; otherwise, proceed to step S2-2.

[0128] It should be noted that the total rebalancing demand reaching the upper limit means that all excess normal bicycles of all demand points are transferred to the demand points where normal bicycles are missing. In this embodiment, in order to prevent the violation of the rebalancing constraint condition (15) in the process of generating the solution, the minimum value of the total surplus demand and the total demand is set as the upper limit of the overall rebalancing demand. For example, if there are 1000 redundant normal bicycles in a certain area (i.e., the total surplus demand) and 7000 normal bicycles that need truck rebalancing (i.e., the total surplus demand), 7000 are set as the upper limit of the overall rebalancing demand.

[0129] It should be further explained that the total load of damaged bicycles on a truck refers to the ratio of the number of damaged bicycles recovered by the truck to the total capacity of the truck, wherein the threshold value can be set to 90%.

[0130] For example, if the number of recovered damaged bicycles accounts for 90% of the total capacity of the truck or the overall rebalancing demand reaches the upper limit, the iteration of the route is terminated and step S2-5 is entered to transport the excess normal bicycles on the truck to the demand points where normal bicycles are missing until no excess normal bicycles are carried by any truck, and then return to the designated center; otherwise, step S2-2 is entered to recover the damaged bicycles and rebalance the normal bicycles until the number of recovered damaged bicycles accounts for 90% of the total capacity of the truck or the overall rebalancing demand reaches the upper limit.

[0131] Step S2-2: Determine whether there is a redundant normal bicycle on the truck. If so, proceed to step S2-3; otherwise, proceed to step S2-4.

[0132] In this embodiment, if there are redundant normal bicycles on the truck, it means that there is a greater chance of first going to the demand point where normal bicycles are missing, and then going to the demand point where normal bicycles are in excess; if there are no redundant normal bicycles on the truck, the truck will randomly go to a demand point.

[0133] Step S2-3: Generate a first random number between [0, 1) through a first random function; when the first random number is between [0, 0.7), go to the demand point where a normal bicycle is missing, and according to the distribution of route planning Generate the next path; when the first random number belongs to [0.7, 1), go to the demand point with normal bicycle surplus and distribute the bicycles according to the route planning. Generate the next path.

[0134] In this implementation, the demand points are set into two groups: (the set of demand points with normal bicycle surplus) and (The set of demand points where normal bicycles are missing.) If the trucks carry redundant normal bicycles, they are more likely to go to the demand locations in M2, which is set to a 70% probability in this embodiment, and less likely to go to the demand locations in M1, which is set to a 30% probability in this embodiment.

[0135] Step S2-4: The truck randomly goes to a demand point and distributes according to the route planning. Generate the next path.

[0136] Step S2-5: Determine whether there are redundant normal bicycles on the truck; if so, go to the demand point where the normal bicycle is missing and distribute it according to the route planning. Generate the next path; otherwise, return the corresponding recycling center.

[0137] In one feasible implementation, in step S2, after the correct path information is generated by the path selection algorithm (steps S2-1 to S2-5), the path information of each truck can be obtained by the following load distribution algorithm:

[0138] Step S2-6: Generate a second random number between [0, 1) by a second random function, and determine whether the second random number belongs to [0, 0.7). If so, proceed to step S2-7; otherwise, proceed to step S2-8;

[0139] Step S2-7: Load distribution corresponding to the current demand point Assign the damaged bike corresponding to the current demand point to the truck;

[0140] Step S2-8: Determine whether the current demand point is a normal bicycle surplus demand point. If so, proceed to step S2-9; otherwise, proceed to step S2-10:

[0141] Step S2-9: Load distribution of excess normal bicycles corresponding to the current demand point Assign trucks to normal single vehicles that are in excess of the current demand point;

[0142] Step S2-10: Load distribution of missing normal bicycles corresponding to the current demand point Assign normal single vehicles that are missing from the current demand point to the truck;

[0143] Step S2-11: Load distribution of damaged bicycles corresponding to the current demand point Assign the damaged bike corresponding to the current demand point to the truck.

[0144] In the load distribution algorithm (steps S2-6 to S2-11), it is randomly decided whether to recover the damaged bike first or to rebalance the normal bike first, and then at the demand point where the damaged bike exists, the load distribution of the damaged bike corresponding to the demand point is calculated. Assign a damaged load to the truck (i.e., the number of damaged bikes corresponding to the demand point); if the truck goes to a demand point in M2 (the set of demand points with missing normal bikes), the load distribution of the missing normal bikes corresponding to the demand point is Assign the missing bike load to the truck (i.e. the number of normal bikes missing at the demand point); if the truck goes to a demand point in M1 (a set of demand points with excess normal bikes), the load distribution of the excess normal bikes corresponding to the demand point is Assign redundant vehicle loads (i.e., the number of normal vehicles in excess of the demand point) to trucks.

[0145] It should be noted that in this implementation mode, when allocating damaged load, missing vehicle load and redundant vehicle load to a truck, the upper limit of the number of corresponding allocated vehicles is set to the minimum value of the current demand and the available capacity of the truck. In an example, a truck travels to a demand point with excess normal vehicles, where there are 50 excess normal vehicles, but the available capacity of the truck is 30, and the number of redundant vehicles allocated is 30.

[0146] Step S3: Repeat step S2 to obtain a feasible solution set Among them, m represents the iteration number, Represents all feasible solutions obtained after the mth iteration.

[0147] Step S4: When the number of feasible solutions reaches the threshold, the feasible solution set is sorted in ascending order according to the objective function. Among all feasible solutions in , the top s% of feasible solutions are selected to form the set of surviving solutions. And the surviving solutions are The feasible solution that minimizes the objective function is taken as the optimal solution.

[0148] In this embodiment, the objective function is sorted in ascending order, and the top s% of feasible solutions are selected, that is, a part of the solutions that make the total cost too high is deleted, and a part of the solutions that save more cost is left as the surviving solution set. And the surviving solutions are The feasible solution that minimizes the objective function is taken as the optimal solution in the current iterative solution process.

[0149] Step S5: According to the surviving solution set The average number of trucks in all feasible solutions to generate a distribution of trucks According to the surviving solution set The frequency with which trucks are assigned to recycling centers generates the distribution of recycling centers According to the surviving solution set The frequency of each demand point being visited in the distribution of route planning is generated According to the surviving solution set The load information at each demand point in the , generates the load distribution corresponding to each demand point

[0150] It should be noted that the load distribution corresponding to each demand point is Including the load distribution of the damaged bicycle at this demand point The load distribution of the missing normal bicycle at this demand point and the excess normal bike load distribution at the demand point

[0151] Step S6: Output the optimal solution of the shared bicycle recycling allocation model Among them, the optimal solution Including distribution of trucks Location of recycling centers Distribution of bicycle positions And the load distribution corresponding to the demand point

[0152] The embodiment of the present application transforms the recycling and rebalancing problem of shared bicycles into a shared bicycle recycling allocation model, which provides decisions on the location of the recycling center, the number of trucks, truck route planning, and load distribution at the demand point, so that the established model can better reflect the actual situation of shared bicycle recycling and rebalancing, and the recycling allocation plan formulated based on the model is more reliable; the embodiment of the present application takes minimizing the total cost as the goal, and constructs an exponentially initialized uniform estimation distribution algorithm based on the estimation distribution algorithm to iteratively solve the mathematical model. During the solution process, the same truck can simultaneously recycle damaged bicycles and rebalance normal bicycles, thereby improving the utilization rate of the truck capacity, and while recycling more damaged bicycles, effectively reducing the cost of recycling damaged bicycles and rebalancing normal bicycles, making the recycling and rebalancing plan formulated based on the model more economical.

[0153] In a feasible implementation, considering that an appropriate initial solution can significantly improve the efficiency of the EIUEDA algorithm, the initial solution of the shared bicycle recycling allocation model can be generated by the following initial solution generation algorithm:

[0154] Step SA-1: Determine the joint cost of each recycling center based on the fixed cost of the recycling center, the location relationship between the recycling center and the demand point, and the load demand of the demand point, and arrange them in ascending order according to the size of the joint cost;

[0155] In this implementation, based on the principle of shortest route, the fixed cost of the recycling center, the positional relationship between the recycling center and the demand point, and the load demand of the demand point are comprehensively considered to determine the joint cost of each recycling center. It should be noted that the joint cost is not the total cost represented by the objective function in the embodiment of the present application, but a reasonable indicator generated in the process of generating the initial solution to determine the initial position and arrangement order of the recycling center.

[0156] Step SA-2: Set the probability of assigning a truck to recycling center k to be Generate an initial solution for the recycling center Among them, k represents the kth recycling center in ascending order;

[0157] In this embodiment, since the recycling centers are arranged in ascending order, that is, the smaller the serial number, the lower the joint cost, the probability of assigning a truck to the recycling center k is arranged in descending order, that is, the recycling center with a smaller serial number and a lower joint cost will have a higher probability of being assigned a truck.

[0158] Step SA-3: Determine the number of trucks based on the total demand at the demand point and generate an initial solution for the trucks The total demand refers to the total number of damaged bicycles that need to be recycled at all demand points;

[0159] In this embodiment, the initial solution for the truck is generated. When generating a truckload, the number of trucks is based on the principle of meeting 70% to 130% of the total demand.

[0160] For example, if the total demand is 100 and the space per truck is 10, a number of trucks between 7 and 13 may be generated.

[0161] Step SA-4: Initial solution based on recycling center Assign trucks to recycling centers and start from the recycling centers. Set each truck to have a 50% probability of passing through each demand point. Each time a truck passes through a demand point, evenly distributed loads are assigned to the truck. After the truck returns to the recycling center, delete the demand points that the truck passed but did not assign loads to.

[0162] In this embodiment, the probability of a truck driving to each demand point is random and equal in probability, and when passing each demand point, a uniformly distributed load is allocated to the truck, and the number of allocated loads ranges from 0 to the current available capacity of the truck and the minimum value that meets the demand. It should be noted that the uniformly distributed load means that when the truck distributes the load, the number of allocated loads is randomly allocated with equal probability within the range of 0 to the current available capacity of the truck and the minimum value that meets the demand. Among them, the load includes the number of damaged bicycles at the demand point, the number of normal bicycles missing at the demand point, and the number of normal bicycles in excess at the demand point. It should be noted that the order in which these three types of loads appear is random. For example, at a certain demand point, 15 damaged bicycles need to be recovered, but the current available capacity of the truck is 10, then the number of damaged bicycles allocated ranges from 0 to 10; in another example, there are 15 excess normal bicycles at the demand point, and the current available capacity of the truck is 50, then the number of excess normal bicycles allocated ranges from 0 to 15.

[0163] In this embodiment, when the truck is fully loaded (i.e., the current available capacity of the truck is 0) without passing through the pre-assigned demand point, it will directly return to the recycling center along the preset route without further load allocation, and the demand points that it has passed but have not been allocated load will be deleted.

[0164] Step SA-5: Generate an initial solution for route planning based on the route planning of all trucks Generate an initial solution of the load demand at the demand point based on the loads of all trucks .

[0165] It should be noted that the initial solution of the load demand at the demand point is Initial load distribution including the damaged bike at this demand point Initial load distribution of the missing normal bicycle at this demand point and the initial load distribution of the excess normal bicycles at the demand point

[0166] In this embodiment, the initial solution generation algorithm (step SA-1 to step SA-5) is used to reasonably customize the initial solution generation, which can effectively improve the computational efficiency and convergence speed of the EIUEDA algorithm.

[0167] In a feasible implementation, between step S5 and step S6, the method further includes:

[0168] Step SE: Determine whether the iteration number m is less than or equal to the maximum number of iterations L and whether the fluctuation of five consecutive iterations is greater than or equal to the preset minimum fluctuation rate. If so, proceed to step S3; otherwise, proceed to step S6.

[0169] In this embodiment, the optimal solution of the shared bicycle recycling allocation model is output. Previously, by further setting the termination condition of the iterative solution (iteration number m is greater than the maximum number of iterations L or the fluctuation of five consecutive iterations is less than the preset minimum fluctuation rate), the fluctuation between the most recently generated optimal solution and the optimal solution generated in the previous four times is less than the preset minimum fluctuation rate, so as to ensure the output of a stable optimal solution. The reliability of the final recycling allocation plan is further improved.

[0170] The effect of the embodiment of the present application can be further illustrated by the following numerical experiments:

[0171] It should be noted that all the following numerical experiments are conducted on Windows 10, using an Intel i5 core and 8BG RAM, and solved using the CPLEX 12.8 solver. In addition, the EIUEDA algorithm is coded by Python 3.6. All demand data are generated based on operational data in the bike-sharing platform. According to the Annual Report on the Development of China's Sharing Economy, we set the parameters shown in Table 1. The fixed cost of the recycling center and the income from meeting different types of demand are generated based on a given distribution, while the fixed cost and capacity of the truck are determined values.

[0172] Table 1 Parameter setting table

[0173]

[0174]

[0175] 1. Verification test of EIUEDA algorithm

[0176] It should be noted that this implementation is used to verify the effectiveness and efficiency of the EIUEDA algorithm. The specific analysis and verification process is as follows:

[0177] Due to the complexity of the problem, the computation time increases dramatically with the problem size, and commercial software like CPLEX still cannot provide an optimal solution within 3600 seconds at a very small scale. To illustrate the effectiveness of the EIUEDA algorithm, we compared the EIUEDA algorithm with CPLEX when CPLEX can still obtain the optimal solution. Then, for larger instances, we examined the EIUEDA algorithm in terms of solution stability and computational efficiency. The comparison between CPLEX and the EIUEDA algorithm is shown in Table 2. In the table, the first column on the left is the instance size, where "4, 2, 2" means that there are 4 demand points, 2 potential recycling centers, and 2 available trucks. The objective function and computation time refer to the target value and computational CPU time of the relevant solution method, and the gap shows the difference between the objectives.

[0178] Table 2 Comparison between CPLEX and EIUEDA algorithms

[0179]

[0180]

[0181] As can be seen from Table 2, for small instances with only 4 demand locations, CPLEX can solve the problem faster than the EIUEDA algorithm with a gap of no more than 3%. As the instance size increases, the computation time of CPLEX increases dramatically, and the EIUEDA algorithm can still give the final solution within an acceptable time (less than 35 seconds) while still maintaining a small gap. With only 6 demand locations, 2 potential recycling centers, and 2 available trucks, CPLEX cannot find the optimal solution within 8000 seconds. In contrast, the EIUEDA algorithm obtains the final solution in 28 seconds with a gap of 0.31%. It is worth noting that the difference between the EIUEDA algorithm and the optimal solution does not increase with the increase in the instance ratio. The results verify the effectiveness and efficiency of the EIUEDA algorithm.

[0182] For the larger instance, 21 sets of experimental verification were conducted, of which the same scale experiments were conducted 5 times. The results of the target value and the calculation CPU time are summarized in Table 3 (Performance evaluation table of EIUEDA algorithm) and Table 4 (Efficiency evaluation table of EIUEDA algorithm). In these tables, "maximum value", "minimum value", "average value", "standard deviation", "coefficient of variation", "range" and "range ratio" represent the maximum value, minimum value, average value, standard deviation, coefficient of variation, range and range ratio of the target function value or calculation time of the 5 experiments.

[0183] In Table 3, the coefficient of variation of all groups is less than 0.02, and the range ratio is greater than 0.96. These results prove that the EIUEDA algorithm provides a stable solution in terms of the objective function value. Therefore, we prove that the performance of the EIUEDA algorithm is robust and stable.

[0184] Table 3 Performance evaluation of the EIUEDA algorithm

[0185]

[0186]

[0187] In Table 4, the coefficient of variation for all groups is less than 0.05, and the average value of the range ratio is less than 1.1. These results demonstrate that the EIUEDA algorithm provides solutions with similar parameter settings within a stable range of computational CPU time. In addition, the computation time does increase with the increase in scale, but the rate is not significant. Even in the largest case of 80 demand points, the average computation time is 4255.89 seconds, while CPLEX cannot terminate with the best solution within 8000 seconds with only 6 demand points. The results demonstrate the stability and effectiveness of the EIUEDA algorithm.

[0188] Table 4 Efficiency evaluation of the EIUEDA algorithm

[0189]

[0190]

[0191] 2. Convergence analysis of EIUEDA algorithm

[0192] It should be noted that this implementation is used to verify the convergence of the EIUEDA algorithm by analyzing three key parameters: the number of solutions per generation, the evolution rate, and the termination gap. The specific analysis and verification process is as follows:

[0193] 2.1. Quantitative analysis of solutions

[0194] The number of solutions per iteration is a key parameter in the EIUEDA algorithm. A small number of solutions will slow down the convergence speed and even lead to premature convergence, but a large number of solutions will waste computing resources. In Table 5, as the number of solutions increases, the computation time first decreases and then increases, while the objective function value decreases and then remains stable. The first time decrease is caused by a small number of iterations, while the second time increase is the result of too many iterations. The experiment not only provides us with an appropriate number of solutions in each iteration, but also proves that when the number of solutions in each iteration is greater than 100, the objective function value remains stable and avoids premature convergence.

[0195] Table 5: Analysis of the convergence of solution quantity

[0196]

[0197] 2.2 Evolution rate analysis

[0198] The evolution rate is another key parameter of the EIUEDA algorithm. A smaller evolution rate will lead to premature convergence, but a larger evolution rate will slow down the convergence and waste computing resources. In Table 6, as the evolution rate increases, the objective function value decreases and then remains stable, while the computing time always increases. The improvement in the objective function value shows that the EIUEDA algorithm will not experience premature convergence when the evolution rate is greater than 15%. In addition, a larger evolution rate will slow down the convergence and increase the number of iterations. The experiment not only provides us with a suitable evolution rate, but also proves that the EIUEDA algorithm can converge to the optimal solution when the evolution rate is greater than 15%.

[0199] Table 6: Convergence analysis of evolutionary rates

[0200]

[0201]

[0202] 2.3. Termination gap analysis

[0203] The termination gap is a key parameter for controlling the number of iterations in the EIUEDA algorithm. A small gap delays convergence, while a large gap weakens the quality of the solution and threatens the effectiveness of the EIUEDA algorithm. In Table 7, as the objective function value decreases and then remains stable as the termination gap decreases, the computational time first increases and then remains stable. When the terminal gap is less than 5%, the stable objective function value indicates that the EIUEDA algorithm provides us with the best solution. In addition, the stable computational time proves the convergence of the EIUEDA algorithm from the side, because the optimal solution cannot be improved by more iterations. The experiment not only provides us with a suitable termination gap, but also proves that the EIUEDA algorithm can converge to the optimal solution when the termination gap is less than 5%.

[0204] Table 7: Convergence analysis of termination gap

[0205]

[0206] 3. Case simulation

[0207] It should be noted that this implementation method is a case study based on actual data from a Chinese city, taking into account demand density, road network capacity and rebalancing priority, to analyze the implications of the EIUEDA algorithm in actual application. The specific analysis and verification process is as follows:

[0208] In this case, there are 80 demand points, 5 recycling centers, and 20 trucks in the city, among which the routes and load distributions of Truck 15 and Truck 19 are shown in Table 8.

[0209] Table 8 Routes and loads of trucks 15 and 19

[0210]

[0211]

[0212] It should be noted that the first two columns of Table 8 are the exact coordinates of the demand locations. "Truck" indicates the truck number; "Order" indicates the order of the trucks; the numbers in "Load" indicate (the damaged bike load, missing normal bike load, redundant normal bike load, and total capacity of each truck at the demand point). For truck 15, it travels to 8 demand points, first loading 70 redundant normal bikes at 3 different demand locations, and then dropping off 2 normal demand locations at other locations. For truck 19, at the demand point on the 4th route, it first drops off 119 redundant normal bikes to meet the demand while providing more available capacity, and then loads 167 damaged bikes. These actions demonstrate the superiority of the coordinated network of the shared bike system, in which capacity is used for recycling and rebalancing many times.

[0213] 3.1 Demand density analysis

[0214] For different cities, the demand density varies greatly. We conducted experiments based on 5 different demand density levels. The results are as follows: Figure 3 As shown, a performance analysis diagram of different demand levels is shown.

[0215] Reference Figure 3 , where the horizontal axis is the demand level, and as the demand density increases, the three types of demand increase in the same proportion, while the vertical axis is the cost. The five broken lines represent recycling income, rebalancing income, total income, total cost, and objective function value. When the demand density is less than level 2, the recycling and rebalancing income increase together with the objective function value because the more the demand is satisfied, the higher the income. However, the total cost increases slightly due to the need to rent more trucks. When the demand density is greater than level 3, as the recycling and rebalancing capacity is fully occupied, the income remains stable, but the total cost and the objective function value continue to decrease. This is because trucks do not need to travel long distances to meet demand, and the reduction in transportation costs leads to a reduction in total costs and objective function values. Therefore, bike-sharing platform operators can benefit from increased demand. If more trucks are rented, demand is better met and income is higher. If the recycling and rebalancing capacity is maintained, the travel distance is shorter, reducing costs and improving system performance.

[0216] 3.2 Road network capacity analysis

[0217] Traffic levels vary widely between cities and even between regions of a city. Low levels of transportation costs may be caused by poor traffic and lack of infrastructure, which directly leads to high levels of transportation costs. We conducted experiments based on 4 different road network capacity levels, and the results are as follows: Figure 4 As shown, a performance analysis diagram of different road network capacity levels is shown.

[0218] Reference Figure 4 , where the horizontal axis represents the transportation level, and the transportation cost decreases as the road network capacity level increases, while the vertical axis represents the transportation cost. As the transportation level increases, the objective function value increases together with the two types of benefits. However, the total cost peaks at level 2 because when the transportation level is low, trucks tend to only recover or rebalance bicycles near the distribution center. As the transportation level increases, trucks begin to actively travel to recover or rebalance bicycles. This change leads to a significant drop in the total cost, which in turn leads to a significant drop in the objective function value. This phenomenon tells us that bike-sharing platform operators should not only pay attention to market demand, but also to the road network capacity level. When the transportation level is poor, fewer trucks and shorter truck routes should be arranged.

[0219] 3.3 Rebalancing Priority Analysis

[0220] The rebalancing priority is different for different bike-sharing platforms or government policies. A company focused on sustainability will pay more attention to recycling due to strict policies, resulting in a lower rebalancing priority. We conducted experiments based on 7 different rebalancing priorities, and the results are as follows: Figure 5 , which is a schematic diagram showing performance analysis of different rebalancing priorities.

[0221] Reference Figure 5 , where the horizontal axis is the rebalancing priority, as it increases, the unit income of rebalancing will increase and the unit income of recycling will decrease, while the vertical axis is the cost. It should be noted that in priority 4, the unit income of rebalancing is equal to the income of recycling. The total income remains stable, which means that the unit income is controlled at an acceptable level. As the rebalancing priority increases, the total cost and the objective function value increase slightly, which shows that recycling is more beneficial to the shared bicycle system than rebalancing bicycles. This phenomenon tells us that bicycle sharing platform operators should pay more attention to recycling bicycles, which is not only good for our environment, but also can reduce the total cost of the shared bicycle system and increase the total income.

[0222] Through the above numerical experiments and case analysis, the stability and effectiveness of the EIUEDA algorithm are verified. At the same time, compared with traditional commercial solvers such as CPLEX, it can provide the best solution faster. For CPLEX, it cannot generate the best solution when there are only 6 demand positions within 8000 seconds, while the EIUEDA algorithm only provides 0.31% of the solutions in 27.64 seconds. In the actual application with 80 demand positions, the EIUEDA algorithm can still generate solutions in an average of less than 3600 seconds. At the same time, the coefficient of variation and range ratio of each group of experiments show the stability of the EIUEDA algorithm in terms of objective function value and calculation time. In addition, the convergence of the EIUEDA algorithm is proved by the convergence analysis of the EIUEDA algorithm, and the appropriate algorithm parameter settings are given, such as the number of solutions per iteration, evolution rate and termination interval. Finally, based on real-world applications, the EIUEDA algorithm can not only provide solutions quickly, but also provide management enlightenment with practical significance, which can effectively promote the sustainable development of the shared bicycle industry.

[0223] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0224] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0225] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0226] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0227] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0228] The above is a detailed introduction to a shared bicycle recycling and distribution method provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A shared bicycle recycling and distribution method, characterized in that: The method comprises: Step S1: Establishing a shared bicycle recycling allocation model, the shared bicycle recycling allocation model includes decision factors, an objective function of the total cost and constraints; wherein the decision factors include the location distribution of the recycling center, the number distribution of trucks, the truck route planning and the load demand distribution of the demand point; the total cost includes the fixed cost of the recycling center, the fixed cost of the truck, the transportation cost and the income brought by meeting the load demand distribution of the demand point; Step S2: Based on the estimated distribution algorithm, an exponentially initialized uniform estimated distribution algorithm is constructed to solve the shared bicycle recycling allocation model and obtain a feasible solution. Among them, the feasible solution It includes the number of recycling centers, the locations of recycling centers, the number of trucks, the path information of each truck, and the load allocated to each demand point, wherein the load allocated to the demand point includes the number of damaged bicycles at the demand point, the number of normal bicycles missing at the demand point, and the number of normal bicycles in excess at the demand point; Step S3: Repeat step S2 to obtain a feasible solution set Among them, m represents the iteration number, Represents all feasible solutions obtained after the mth iteration; Step S4: When the number of feasible solutions reaches a threshold, the feasible solution set is sorted in ascending order according to the objective function. Among all feasible solutions in , the top s% of feasible solutions are selected to form the set of surviving solutions. And the surviving solutions are The feasible solution that minimizes the objective function is taken as the optimal solution. Step S5: According to the surviving solution set The average number of trucks in all feasible solutions to generate a distribution of trucks According to the surviving solution set The frequency with which trucks are assigned to recycling centers generates the distribution of recycling centers According to the surviving solution set The frequency of each demand point being visited in the distribution of route planning is generated According to the surviving solution set The load information at each demand point in the , generates the load distribution corresponding to each demand point Step S6: Output the optimal solution of the shared bicycle recycling allocation model Among them, the optimal solution Including distribution of trucks Location of recycling centers Distribution of bicycle positions And the load distribution corresponding to the demand point Wherein, between step S5 and step S6, the method further includes: Step SE: Determine whether the iteration number m is less than or equal to the maximum iteration number L and whether the fluctuation of five consecutive iterations is greater than or equal to the preset minimum fluctuation rate. If so, proceed to step S3; otherwise, proceed to step S6.

2. The method according to claim 1, characterized in that The specific steps of step S1 include: Step S1-1: The truck starts from the recycling center k, and when passing through each demand point, it takes away the damaged bicycles at the demand point and / or takes away the excess normal bicycles and / or brings the missing normal bicycles, until the total load of the damaged bicycles on the truck reaches the threshold or the overall rebalancing demand reaches the upper limit, and then returns to the recycling center k; wherein the overall rebalancing demand reaches the upper limit means that the excess normal bicycles of all demand points are transferred to the demand points where the normal bicycles are missing; Step S1-2: Define the following variables: Step S1-3: Construct the objective function of the total cost: The goal is to minimize the total cost; where f k represents the fixed cost of the recycling center; g n represents the fixed cost of truck n; t ij Represents the distance between demand point i and demand point j; express or in, represents the profit from recycling a damaged bicycle at demand point i, represents the profit brought by bringing a missing normal bicycle to the demand point i, represents the profit brought by taking away a redundant normal bicycle at demand point i; express or in, represents the number of damaged bicycles that need to be recovered by truck n at demand point i, represents the number of normal bicycles that need to be brought to the demand point i by truck n, Indicates the number of normal bicycles that need to be taken away by truck n at demand point i; Step S1-4: Establish the following constraints: N,t1∈{1,2,3},2∈{1,2} In the formula, I represents the set of demand points, N represents the set of trucks, K represents the set of recycling centers, and c n represents the capacity of truck n, M is a preset constant, represents the number of damaged bicycles at demand point i, represents the number of normal bicycles missing at demand point i, represents the normal number of excess bicycles at demand point i, represents the load of the damaged single truck of truck n at demand point i, represents the load of the rebalanced single vehicle of truck n at demand point i, represents the load change of the damaged single truck of truck n from demand point i to demand point j, represents the load change of truck n when rebalancing from demand point i to demand point j, u in Eliminate sub-loop redundant variables; where, formula (4) limits the allocation of trucks to only those recycling centers that have been opened; formula (5) ensures that each truck can only be allocated to the corresponding recycling center once at most; formula (6) limits the designated trucks to travel between demand points; formula (7) controls that trucks cannot take away bicycles that exceed the demand; formulas (8) and (9) limit trucks from recovering damaged bicycles or rebalancing normal bicycles if they do not pass through the demand points; formulas (10) and (11) eliminate sub-loops, prevent trucks from passing through the same demand point multiple times, and ensure that the paths of all trucks include their corresponding recycling centers; Formulas (12) and (13) restrict the starting and ending points of trucks to designated recycling centers; Formula (14) is a flow balance constraint, which is used to associate the paths of trucks to ensure that a truck entering a demand point will exit from that demand point; Formula (15) indicates that a truck cannot return to the recycling center before completing the rebalancing of normal single vehicles; Formula (16) restricts the load of a truck to not exceed its capacity; Formulas (17) and (18) define the load variation of trucks at demand points; Formulas (19) and (20) define the load variation of each truck between demand points; Formula (21) restricts the value range of the variables in the formula.

3. The method according to claim 1, characterized in that The method further comprises: Follow these steps to generate an initial solution to the shared bike recycling allocation model: Step SA-1: determining the joint cost of each recycling center according to the fixed cost of the recycling center, the positional relationship between the recycling center and the demand point, and the load demand of the demand point, and arranging the recycling centers in ascending order according to the size of the joint cost; Step SA-2: Set the probability of assigning the truck to recycling center k to be Generate an initial solution for the recycling center Among them, k represents the kth recycling center in ascending order; Step SA-3: Determine the number of trucks based on the total demand of the demand point and generate an initial solution for the trucks. The total demand refers to the total number of damaged bicycles that need to be recycled at all demand points; Step SA-4: Based on the initial solution of the recycling center Assign trucks to the recycling center and set out from the recycling center. Set each truck to have a 50% probability of passing through each demand point. Each time the truck passes through a demand point, a uniformly distributed load is allocated to the truck. After the truck returns to the recycling center, delete the demand points that the truck has passed but has not been allocated a load. Step SA-5: Generate an initial solution for route planning based on the route planning of all trucks Generate an initial solution of the load demand at the demand point based on the loads of all trucks 4. The method according to claim 1, characterized in that: In step S2, the path information of each truck is obtained by the following steps: Step S2-1: Determine whether the total load of the damaged bicycles on the truck has reached a threshold or whether the overall rebalancing demand has reached an upper limit. If so, proceed to step S2-5; otherwise, proceed to step S2-2; Step S2-2: Determine whether there is a redundant normal bicycle on the truck, if so, proceed to step S2-3; otherwise, proceed to step S2-4; Step S2-3: Generate a first random number between [0, 1) by using a first random function; when the first random number belongs to [0, 0.7), go to the demand point where normal bicycles are missing, and distribute the bicycles according to the route planning. Generate the next path; when the first random number belongs to [0.7, 1), go to the demand point with normal bicycle surplus, and according to the distribution of route planning Generate the next path; Step S2-4: The truck randomly goes to a demand point and distributes according to the route planning. Generate the next path; Step S2-5: Determine whether there is a redundant normal bicycle on the truck; If so, go to the demand point where normal bicycles are missing and distribute them according to the route planning. Generate the next path; otherwise, return the corresponding recycling center.

5. The method according to claim 4, characterized in that In step S2, the load allocated to each truck at each demand point is determined by the following steps: Step S2-6: Generate a second random number between [0, 1) by a second random function, and determine whether the second random number belongs to [0, 0.7). If so, proceed to step S2-7; otherwise, proceed to step S2-8; Step S2-7: Load distribution corresponding to the current demand point Allocate the damaged bicycle corresponding to the current demand point to the truck; Step S2-8: Determine whether the current demand point is a demand point with normal bicycle surplus. If so, proceed to step S2-9; otherwise, proceed to step S2-10: Step S2-9: Load distribution of excess normal bicycles corresponding to the current demand point Allocate a normal single vehicle that is in excess of the current demand point to the truck; Step S2-10: Load distribution of missing normal bicycles corresponding to the current demand point Allocate the normal single vehicle missing from the current demand point to the truck; Step S2-11: Load distribution of damaged bicycles corresponding to the current demand point Assign the damaged bicycle corresponding to the current demand point to the truck.

Images