A logistics distribution path generation method based on a fishing algorithm

Abstract

The present invention discloses a method for generating a logistics distribution path based on a fishing algorithm, belonging to the technical field of logistics distribution, which includes the following steps: S1, establishing a mathematical model of the logistics distribution path problem; S2, determining the cost index value Ti and the weight coefficient Wi; S3, calculating the distribution cost matrix QM according to the determined cost index value Ti and the weight coefficient Wi; S4, randomly generating a group of feasible distribution paths as the initial positions of the fisherman individuals in the fishing algorithm and setting the corresponding parameter values; S5, performing iterations according to the principle of the fishing algorithm, calculating the fitness values of all fisherman individuals, and updating the positions and states of the fishermen; S6, when the algorithm runs to an end, outputting the found optimal distribution path sequence and the corresponding total cost value according to the position of the optimal fisherman individual. The present invention can find the optimal logistics distribution path in a relatively short time through the fishing algorithm, reduce the cost inputs such as the logistics distribution distance, time, and energy consumption in practical applications, and improve the logistics distribution efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of logistics distribution, and particularly relates to a method for generating a logistics distribution path based on a fishing algorithm. Background Art

[0002] The logistics distribution path refers to the route passed during the process of starting from a distribution center (or warehouse), going through a series of transportation activities, and finally delivering goods to customers to meet the needs of customer orders. It is one of the key links in the logistics distribution process. Through reasonable path planning and optimization methods and the application of advanced technologies, the logistics efficiency can be significantly improved, the logistics cost can be reduced, and the customer satisfaction can be enhanced. At the same time, with the continuous progress of technology and the enhancement of environmental awareness, the optimization of the logistics distribution path will also pay more attention to green environmental protection and sustainable development.

[0003] In the prior art, intelligent algorithms commonly used for generating logistics distribution paths include genetic algorithms, ant colony algorithms, particle swarm algorithms, etc. In the process of searching for the optimal solution, genetic algorithms and ant colony algorithms often require a large number of iterative calculations, and the convergence speed is slow, which means that it takes a long calculation time to obtain a relatively satisfactory result, and it is not suitable for logistics distribution scenarios with high real-time requirements. The particle swarm algorithm is prone to falling into a local optimal solution during the search process, resulting in the inability to find the globally optimal distribution path. This is because the intelligent algorithm may be affected by the initial solution during the search process, or may fall into a local optimal area during the search process and be unable to jump out. Therefore, it is necessary to improve the prior art. Summary of the Invention

[0004] The present invention proposes a method for generating a logistics distribution path based on a fishing algorithm, which can find the optimal logistics distribution path in a short time to reduce the cost inputs such as logistics distribution distance, time, and energy consumption in actual applications, and improve the logistics distribution efficiency.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A method for generating a logistics distribution path based on a fishing algorithm, comprising the following steps:

[0007] S1: Establish a mathematical model for the logistics distribution path problem, specifically as follows:

[0008] Assume that the logistics distribution path problem has M distribution points, and a feasible distribution path is represented as a vector θ=(p1p2…p k …p M ); where the integer p k ∈[1,M], indicating that the p k -th distribution point is located at the k-th position in this distribution path;

[0009] Let T i (p M , p M+1 ) = T i (p M , p1), then the objective function of the logistics distribution path problem is:

[0010]

[0011] Among them, T i (p k , p k+1 ) represents the value of the i-th cost index required to reach the distribution point p k from the distribution point p k+1 ; W i represents the weight coefficient of the i-th cost index and needs to satisfy N represents the number of cost indices to be considered;

[0012] Therefore, the optimal solution of the logistics distribution path problem is min(C(θ));

[0013] S2: Determine the cost index value T i and the weight coefficient W i ;

[0014] S3: Calculate the distribution cost matrix Q i according to the determined cost index value T i and the weight coefficient W M ;

[0015]

[0016] S4: Randomly generate a set of feasible distribution paths as the initial positions of the fisherman individuals in the fishing algorithm, and set the corresponding parameter values of the fishing algorithm according to the scale, characteristics and constraint conditions of the problem;

[0017] S5: According to the principle of the fishing algorithm, at the beginning of the iteration, each fisherman casts a net at the current position; then, the fitness values of all fisherman individuals are calculated using the distribution cost matrix Q M or the objective function; then, the positions and states of the fishermen are updated, and the fisherman individuals that do not meet the update conditions are randomly initialized; finally, the position of the optimal fisherman individual is obtained;

[0018] S6: When the algorithm ends, according to the position of the optimal fisherman individual, output the found optimal distribution path sequence and the corresponding total cost value.

[0019] Furthermore, in step S1, the cost indices include distribution distance, distribution time, and distribution energy consumption.

[0020] Further, in step S2, the cost indicators can be selected according to the actual situation or selected from the cost items of the existing historical data, so as to determine the corresponding cost indicator value T i ; then, according to the actual situation including specific requirements or priorities, weigh and sequentially determine the weights W corresponding to the selected cost indicators respectively i .

[0021] Further, in step S2, for the N selected cost indicators, if there are different measurement units or a large difference in the order of magnitude, it is necessary to normalize their values T i ; if only a single cost indicator is considered, no processing is required.

[0022] Further, in step S3, in practical applications, the values of T i (k,k + 1) and T i (k + 1,k) can be the same or different; in addition, if the delay factors including the loading, unloading or handover time of the distribution points are not considered, T i (k,k) can be set to 0, then we can get:

[0023]

[0024] Further, in step S5, the position and state of the fisherman are updated by using the mobile search or the shrinking search.

[0025] Due to the above technical solutions, the present invention has the following beneficial effects:

[0026] 1. The present invention establishes a mathematical model for the logistics distribution path problem. This model can simultaneously consider multiple cost-related index values that need to be optimized, namely cost indicators, such as distribution distance, distribution time or distribution energy consumption, etc., and can adjust the relative importance or significance between different cost indicators by setting weight coefficients, which is beneficial to generating the optimal distribution path subsequently.

[0027] 2. The present invention uses the fishing algorithm to optimize the logistics distribution path, which has the characteristics of fast convergence speed, good robustness, and not easy to fall into local optimum, etc. It can generate a globally optimal feasible path sequence in a short time, and effectively realize the rapid generation of the logistics distribution path.

[0028] 3. The present invention takes the level of the total cost as the evaluation criterion for the quality of the logistics distribution path. At the end of the optimization, in addition to obtaining the optimal logistics distribution path, it can also obtain the corresponding cost value that can be compared in size. Description of the Drawings

[0029] Figure 1 is the flowchart of the logistics distribution path generation method provided by the present invention;

[0030] Figure 2 is the convergence curve graph of the fishing algorithm in the present invention; Specific embodiments

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] As Figure 1 shown, a method for generating a logistics distribution path based on a fishing algorithm includes steps S1-S6.

[0033] S1: Establish a mathematical model for the logistics distribution path problem, specifically as follows:

[0034] Assume that the logistics distribution path problem has M distribution points, and a feasible distribution path is represented as a vector θ = (p1p2…p[[ID=ON]] k …p[[ID=ON]] M );where the integer p[[ID=ON]] k ∈ [1, M], indicating that the p[[ID=ON]] k th distribution point is located at the kth position in the distribution path.

[0035] Let T[[ID=ON]] i (p[[ID=ON]] M , p[[ID=ON]] M+1 ) = T[[ID=ON]] i (p[[ID=ON]] M , p1), then the objective function of the logistics distribution path problem is:

[0036]

[0037] where T[[ID=ON]] i (p[[ID=ON]] k , p[[ID=ON]] k+1 ) represents the value of the ith cost index required to reach the distribution point p[[ID=ON]] k from the distribution point p[[ID=ON]] k+1 ; W[[ID=ON]] i represents the weight coefficient of the ith cost index, and needs to satisfy N represents the number of cost indexes to be considered.

[0038] The cost index is a cost factor to be considered in the distribution. The cost index includes distribution distance, distribution time, and distribution energy consumption. In practical applications, it is generally necessary to perform normalization processing on different cost indexes. The weight coefficient of the cost index is a weight value corresponding to the cost index, which represents the relative importance of the index.

[0039] Therefore, the optimal solution to the logistics distribution path problem is min(C(θ)), that is, find the one with the lowest total cost among all feasible distribution paths as the optimal solution.

[0040] S2: Determine the cost index value T i and the weight coefficient W i . The cost index can be selected according to the actual situation, or selected from the cost items of the existing historical data, so as to determine the corresponding cost index value T i . For the N selected cost indexes, if there are different measurement units or a large difference in the order of magnitude, the value T i needs to be normalized. If only a single cost index is considered, no processing is required. Then, according to the actual situation including specific requirements or priorities, weigh and determine the respective weights W corresponding to the selected cost indexes in turn i .

[0041] S3: Calculate the distribution cost matrix Q according to the determined cost index value T i and the weight coefficient W i . The calculation method of the cost matrix Q M is as follows: M In practical applications, the values of T

[0042]

[0043] (k,k + 1) and T i (k + 1,k) can be the same or different; in addition, if the delay factor including the loading, unloading or handover time of the distribution points is not considered, T i (k,k) can be set to 0, then we can get: i

[0044]

[0045] S4: Randomly generate a group of feasible distribution paths as the initial positions of the fishermen individuals in the fishing algorithm, and set the corresponding parameter values of the fishing algorithm according to the scale, characteristics and constraint conditions of the problem. Randomly generating a group of feasible distribution paths is to generate a random permutation of M distribution points through a random function, and the number of generated random permutations is the same as the number of fishermen individuals. That is, a random permutation corresponds to the initial position of a fisherman.

[0046] S5: According to the principle of the fishing algorithm, at the beginning of the iteration, each fisherman casts a net at the current position; then, use the distribution cost matrix Q MOr the objective function calculates the fitness values of all fisherman individuals; then, the positions and states of the fishermen are updated, and the fisherman individuals that do not meet the update conditions are randomly initialized; finally, the position of the optimal fisherman individual is obtained. Among them, specifically, the positions and states of the fishermen are updated by using mobile search or contraction search.

[0047] S6: The algorithm runs to completion. According to the position of the optimal fisherman individual, the found optimal delivery path sequence and the corresponding total cost value are output.

[0048] Embodiment

[0049] The present invention will be further described below through specific embodiments.

[0050] S1: Establish a mathematical model for the logistics delivery path problem.

[0051] S2: Determine the cost index value T i and the weight coefficient W i . In this example, only the delivery time is selected as the cost index, so the corresponding weight is equal to 1.

[0052] S3: Calculate the cost matrix Q M , since the delivery time is selected as the cost index in S2 and the weight is 1, then Q M = QW = Q. Assuming that under the condition of not considering the loading and unloading point delay factors, the delivery cost matrix containing 8 delivery points is calculated according to the historical delivery time data as:

[0053]

[0054] S4: Randomly generate a group of feasible delivery paths, and set the corresponding parameter values of the fishing algorithm according to the scale, characteristics and constraint conditions of the problem. In this example, the logistics delivery path contains 8 delivery points, so the position of the jth fisherman can be represented by the vector . Initialize the algorithm bulletin board, and set the basic parameters of the fishing algorithm as: the number of fishermen is 3, the casting radius is 5, the casting times is 6, the contraction coefficient is 0.8, the invalid contraction threshold is 3, and the number of iterations is 100 times. Thus, the initial positions of the randomly generated fishermen can be represented by θ1, θ2 and θ3.

[0055] S5: Obtain the position of the optimal fisherman individual based on the fishing algorithm. According to the cost data and algorithm parameters in steps S3 and S4, find the optimal logistics delivery path through the fishing algorithm. The specific method is as follows:

[0056] (1) All fishermen need to cast the net 6 times at the current position θ j and obtain 6 different net-casting points. Specifically, the net-casting point set of the jth fisherman is represented as where, denotes the current position of the \(j\)-th fisherman, i.e., Each fishing point is obtained by randomly swapping any \(\kappa\) number pairs in j where \(\kappa\) j is the current fishing radius of the \(j\)-th fisherman.

[0057] (2) Calculate the fitness values of all fishermen individuals using the distribution cost matrix \(Q\) M or the objective function, including the fitness values corresponding to the positions of the fishermen and the fishing points. If the current group optimal value is better than the bulletin board, update the bulletin board.

[0058] The specific calculation method of the fitness value is as follows. Let the point Substitute it into the objective function of the aforementioned logistics distribution path problem to get

[0059]

[0060] So there is

[0061]

[0062] That is

[0063]

[0064] As described in step S2, in this example, only the distribution time is selected as the cost index, i.e., \(N = 1\), so the above formula can be simplified to

[0065]

[0066] Substitute the value of the distribution cost matrix \(Q\) in step S3 M to get

[0067]

[0068] Therefore, the fitness value corresponding to the point is 51.6149.

[0069] (3) For each fisherman individual, it is necessary to make a judgment according to the following conditions and perform the corresponding update operation. Specifically, if is satisfied, then the \(j\)-th fisherman performs a mobile search, i.e., updates the current position of the \(j\)-th fisherman to \(\theta\) j ', and resets its invalid shrinkage count to zero. If And if the number of invalid contractions is less than the invalid contraction threshold, then the j-th fisherman performs a contraction search, that is, increments the number of invalid contractions of the j-th fisherman by 1, then multiplies its current casting radius by the contraction coefficient and rounds down to obtain a number as the current casting radius value of the fisherman, that is, let If the above two conditions are not met, then the j-th fisherman is randomly initialized.

[0070] (4) If the number of iterations has not reached 100, then go to (1), otherwise go to step S6.

[0071] S6: According to the position of the optimal fisherman individual, output the optimal logistics distribution path sequence recorded in the bulletin board and the corresponding total cost value. Based on the data and parameter settings of the above embodiments, running the fishing algorithm once, the obtained optimal logistics distribution sequence is (1 6 7 5 8 4 2 3), and the corresponding cost is 29.88491.

[0072] In addition, to further verify the reliability and feasibility of the present invention, the algorithm is continuously and independently run 50 times. The optimal sequence obtained each time is consistent with the result of the above single run, and the average cost obtained from 50 runs is also 29.88491; the average fitness convergence curve corresponding to 50 runs is as Figure 2 shown.

[0073] The present invention establishes a mathematical model for the logistics distribution path problem. This model can simultaneously consider multiple cost-related index values that need to be optimized, namely cost indicators such as distribution distance, distribution time, or distribution energy consumption, etc., and can adjust the relative importance or significance between different cost indicators by setting weight coefficients, which is beneficial for subsequent generation of the optimal distribution path. The present invention uses the fishing algorithm to optimize the logistics distribution path, which has the characteristics of fast convergence speed, good robustness, and not easily falling into local optima. It can generate a globally optimal feasible path sequence in a short time and effectively realize the rapid generation of the logistics distribution path. The present invention uses the level of the total cost as the criterion for evaluating the quality of the logistics distribution path. At the end of the optimization, in addition to obtaining the optimal logistics distribution path, a cost value that can be used to compare sizes can also be obtained.

[0074] The above description is a detailed description of the preferred feasible embodiments of the present invention, but the embodiments are not used to limit the patent application scope of the present invention. Any equivalent changes or modifications completed under the technical spirit disclosed by the present invention should fall within the patent scope covered by the present invention.

Claims

1. A method for generating a logistics distribution path based on a fishing algorithm, characterized in that It includes the following steps: S1: Establish a mathematical model for the logistics distribution path problem as follows: Suppose the logistics distribution path problem has M distribution points, and a feasible distribution path is represented as a vector θ = (p1 p2…p k …p M ); where the integer p k ∈ [1, M], indicating that the p k -th distribution point is located at the k-th position in this distribution path; Let T i (p M , p M+1 ) = T i (p M , p1), then the objective function of the logistics distribution path problem is: Among them, T i (p k , p k+1 ) represents the value of the i-th cost index required to reach the delivery point p k from the delivery point p k+1 to the delivery point p i ; W represents the weight coefficient of the i-th cost index and needs to satisfy N represents the number of cost indices to be considered; Therefore, the optimal solution to the logistics distribution path problem is min(C(θ)); S2: Determine the cost index value T i and the weight coefficient W i ; S3: According to the determined cost index value T i and the weight coefficient W i calculate the distribution cost matrix Q M ; S4: Randomly generate a set of feasible distribution paths as the initial positions of the fisherman individuals in the fishing algorithm, and set the corresponding parameter values of the fishing algorithm according to the scale, characteristics, and constraint conditions of the problem; S5: According to the principle of the fishing algorithm, at the beginning of each iteration, each fisherman casts a fishing net at the current position; then, the fitness values of all fisherman individuals are calculated using the distribution cost matrix Q M or the objective function; next, the positions and states of the fishermen are updated, and the fisherman individuals that do not meet the update conditions are randomly initialized; finally, the position of the optimal fisherman individual is obtained. S6: When the algorithm finishes running, output the found optimal distribution path sequence and the corresponding total cost value according to the position of the optimal fisherman individual.

2. The logistics distribution path generation method based on the fishing algorithm according to claim 1, wherein In step S1, the cost metrics include distribution distance, distribution time, and distribution energy consumption.

3. A method for generating a logistics distribution path based on a fishing algorithm according to claim 1, wherein, In step S2, the cost index can be selected according to the actual situation or determined from the cost items of the existing historical data, so as to determine the corresponding cost index value T i ; then, according to the actual situation including specific requirements or priorities, weigh and determine the weights W corresponding to the selected cost indices respectively in turn i .

4. A method for generating a logistics distribution path based on a fishing algorithm according to claim 3, wherein, In step S2, for the selected N cost indicators, if they involve different measurement units or have a large difference in magnitude, their values T i need to be normalized; If only a single cost metric is considered, no processing is performed.

5. A method for generating a logistics distribution path based on a fishing algorithm according to claim 1, characterized in that In step S3, in practical applications, T i (k, k + 1) and T i (k + 1, k) can have the same or different values; in addition, if the delay factor including the loading, unloading, or handover time of the distribution point is not considered, let T i (k, k) = 0, then we get:

6. A method for generating a logistics distribution path based on a fishing algorithm according to claim 1, characterized in that, In step S5, update the position and state of the fisherman using moving search or shrinking search.

Images