Supply chain optimization method, device and readable medium based on IJS-SVM model

Through the supply chain optimization method based on the IJS-SVM model, the problem of insufficient model accuracy and interpretability in the simulation technology is solved, and efficient management and cost optimization of the supply chain are achieved, which is suitable for complex supply chain environments.

CN118709859BActive Publication Date: 2024-11-29HUAQIAO UNIVERSITY
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411180414.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-11-29
Estimated Expiration
2044-08-27

AI Technical Summary

Technical Problem

In the supply chain management, existing simulation technologies have problems such as model accuracy and complexity limited by data quality and modeling assumptions, high computational costs, high professional knowledge requirements, and insufficient interpretability of simulation results.

Method used

Using the supply chain optimization method based on the IJS-SVM model, by building a three-level supply chain network, optimizing the support vector machine model using the improved jellyfish search algorithm, and explaining the characteristics of uncertain factors in combination with the SHAP algorithm, calculating the weight value and uncertainty probability , solve the optimal solution of the objective function to obtain the optimal supply chain solution.

Benefits of technology

It improves the accuracy and interpretability of the model, reduces calculation costs, simplifies operating processes, enhances the popularity of technology, helps enterprises effectively manage supply chains, and achieves cost minimization and profit maximization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118709859B_ABST
    Figure CN118709859B_ABST
Patent Text Reader

Abstract

The present invention discloses a supply chain optimization method, device and readable medium based on an IJS-SVM model, which relates to the field of data processing, including: constructing a three-level supply chain network, and establishing an objective function of the three-level supply chain network and its corresponding constraints; constructing a support vector machine model, optimizing the support vector machine model by an improved jellyfish search algorithm to obtain an optimized support vector machine model, training the optimized support vector machine model to obtain a trained support vector machine model; using a SHAP algorithm to interpret the output result of the trained support vector machine model, calculating the SHAP value of each uncertain factor feature and the weight value of each uncertain factor feature, calculating the uncertainty probability according to the weight value of each uncertain factor feature, solving the optimal solution of the objective function under the constraint condition based on the uncertainty probability, obtaining the optimal supply chain solution, and solving the uncertainty problem in the supply chain process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a supply chain optimization method, device and readable medium based on an IJS-SVM model. Background Art

[0002] In today's globalized and competitive business environment, supply chain management has become one of the key factors for enterprises to achieve sustainable development. However, the increasing complexity and uncertainty of the supply chain has become one of the main challenges that restrict enterprises from effective management. The uncertainty of the supply chain includes many factors, such as market demand fluctuations, disruptions in raw material supply, transportation delays, political reasons and natural disasters, which will have a profound impact on the stability and efficiency of the supply chain. In this context, simulation technology is widely used to study supply chain management to help enterprises better understand and cope with the challenges brought by uncertainty. Simulation technology can model complex supply chain systems and evaluate the effects of different decision-making schemes through experiments and analysis, providing managers with decision support and optimization suggestions. Simulation technology is a very cost-effective method to evaluate the parameters of specific links, usually using a constructed mathematical model combined with relevant algorithms for studying uncertainty. Using the simulation results, managers can optimize the management strategy of each link in the supply chain to minimize costs and maximize profits.

[0003] However, current simulation technology also has some limitations. For example, the accuracy and complexity of the model are limited by the quality of the data used and the rationality of the modeling assumptions. In addition, the interpretability and operability of simulation results may sometimes be insufficient, resulting in deviations in actual application. High computing costs and the requirement for professional knowledge also make it difficult for some companies to implement simulation technology. Summary of the invention

[0004] The purpose of this application is to propose a supply chain optimization method, device and readable medium based on the IJS-SVM model to address the above-mentioned technical problems.

[0005] In a first aspect, the present invention provides a supply chain optimization method based on an IJS-SVM model, comprising the following steps:

[0006] Construct a three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses, and multiple assembly sites, and establish the objective function of the three-level supply chain network and its corresponding constraints;

[0007] A support vector machine model is constructed, and the support vector machine model is optimized by an improved jellyfish search algorithm to obtain an optimized support vector machine model, and the optimized support vector machine model is trained to obtain a trained support vector machine model, wherein the input data of the trained support vector machine model is the uncertainty factor characteristics of the three-level supply chain network, and the output result of the trained support vector machine model is the classification result of whether the corresponding product is completed assembly late or the product is completed assembly on time;

[0008] The SHAP algorithm is used to interpret the output of the trained support vector machine model and calculate the SHAP value of each uncertainty factor feature;

[0009] The weight value of each uncertain factor feature is calculated according to the SHAP value of each uncertain factor feature, and the uncertainty probability is calculated according to the weight value of each uncertain factor feature. Based on the uncertainty probability, the optimal solution of the objective function under the constraint conditions is solved to obtain the optimal supply chain solution.

[0010] As a preferred option, the objective function of the three-level supply chain network is:

[0011] ;

[0012] Where Z represents the cost of the three-level supply chain network, W represents the total number of warehouses, , T represents the total number of time periods in a cycle, ; N represents the total number of manufacturing plants, ; S represents the total number of assembly sites, ; E represents the total number of products, ; represents the fixed cost of building the w-th warehouse, is 1 or 0. If the wth warehouse is established, then is 1, otherwise is 0; represents the unit production cost of the e-th product produced by the n-th manufacturing plant in the t-th period, represents the number of products in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period; represents the unit transportation cost of the e-th product from the n-th manufacturing plant to the w-th warehouse, represents the transportation volume of the e-th product from the n-th manufacturing plant to the w-th warehouse in the t-th period under the influence of uncertain factors, represents the unit transportation cost of shipping the e-th product from the w-th warehouse to the s-th assembly site in the t-th period, represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period under the influence of uncertain factors, represents the unit inventory cost of the remaining products of the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, represents the unit inventory cost of the remaining products in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, represents the number of products that were not delivered on time at the s-th assembly site in the t-th period, represents the cost caused by the untimely supply of the e-th product at the s-th assembly site during the t-th period, I represents the unit cost of the construction delay, Indicates the number of days the construction period is delayed;

[0013] The constraints are:

[0014] Production process:

[0015] ;

[0016] ;

[0017] ;

[0018] ;

[0019] Logistics warehouse link:

[0020] ;

[0021] ;

[0022] ;

[0023] ;

[0024] ;

[0025] Assembly site links:

[0026] ;

[0027] ;

[0028] ;

[0029] ;

[0030] ;

[0031] ;

[0032] ;

[0033] in, represents the quantity of the e-th product in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period, represents the unit volume of the e-th product, is 1 or 0. If the e-th product is produced by the n-th manufacturing plant in the t-th period, then is 1, otherwise is 0; represents the total available capacity of the nth manufacturing plant in the tth period, represents the setup time for the production of the e-th product in the n-th manufacturing plant, and its average value is , the standard deviation is ; represents the transportation volume of the e-th product shipped from the n-th manufacturing plant to the w-th warehouse in the t-th period without considering the uncertainty factors. represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period without considering the uncertainty factors, represents the number of e-th products that the n-th manufacturing plant can produce in parallel, It represents the production cycle of the e-th product produced by the n-th manufacturing plant, and its average value is , the standard deviation is ; f represents the uncertainty probability, represents the total available time of the nth manufacturing plant in the tth period, and the average value is , the standard deviation is ; represents the total amount of products delivered by the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the t-1th period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, is 1 or 0. In the tth period, if the vehicle has the ability to successfully transport the eth product from the nth manufacturing plant to the wth warehouse, then is 1, otherwise is 0; represents the shipping vehicle capacity from the nth manufacturing plant to the wth warehouse, represents the unit area of ​​the e-th product, represents the remaining storage area of ​​the w-th warehouse in the t-th period, represents the weight of the e-th product, represents the number of shelves remaining in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-1th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, is 1 or 0. If the wth warehouse provides the necessary product equipment for the sth assembly site, then is 1, otherwise is 0; represents the number of products assembled at the s-th assembly site during the t-th period. represents the standard deviation of the assembly quantity of the e-th product at the s-th assembly site in the t-th period, represents the average number of assembly quantities of the e-th product at the s-th assembly site in the t-th period, represents the capacity of the shipping vehicle from the w-th warehouse to the s-th assembly site, represents the number of products that were not delivered on time at the s-th assembly site in the t-1th period, represents the number of products that were not delivered on time at the s-th assembly site in the t-th period, Represents the basic planning cycle of an engineering project. Indicates the time required for engineering simulation to complete all tasks.

[0034] As a preferred embodiment, the support vector machine model is optimized by an improved jellyfish search algorithm to obtain an optimized support vector machine model, which specifically includes:

[0035] Collect training data. The input data in the training data are the features of the uncertainty factors, and the output labels are 0 and 1, where 0 means that the product is assembled late, and 1 means that the product is assembled on time.

[0036] Initialize the penalty coefficient c of the support vector machine model, the kernel function parameter g of the support vector machine model, the population size Q of jellyfish, the maximum number of iterations M, and the upper limit U of the independent variable b and the lower limit of the independent variable L b ;

[0037] The initialized penalty coefficient c and kernel function parameter g are used as the parameters of the support vector machine model for the kth iteration to start the following iterative process:

[0038] Jellyfish use a timing control mechanism to switch between following the ocean current and swimming autonomously;

[0039] In the phase of following the ocean current, the position of the jellyfish of the current iteration is updated using the following formula:

[0040] ;

[0041] in, is the position of the i-th jellyfish after the k-th iteration, is the position of the i-th jellyfish after the k+1th iteration, is the optimal position of the jellyfish in the population after the kth iteration, is the distribution coefficient, Represents a random number between 0 and 1. represents the average position of all jellyfish in the population;

[0042] In the autonomous swimming stage, if the jellyfish performs passive movement, the position of the jellyfish of the current iteration is updated using the following formula:

[0043] ;

[0044] in, is the coefficient of motion;

[0045] If the jellyfish is actively moving, the movement direction and position update formula are as follows:

[0046] ;

[0047] ;

[0048] in, Indicates the direction of movement, is a jellyfish randomly selected from the population after the kth iteration, represents the fitness function;

[0049] The movement of jellyfish is regulated by the time control mechanism, and the control function formula is:

[0050] ;

[0051] in, is a random value in [0,1] that changes with the number of iterations, M is the maximum number of iterations, is the threshold constant, if , jellyfish follow the ocean currents; otherwise, jellyfish perform active and passive movements;

[0052] The Logistic chaotic map is introduced to initialize the jellyfish population, as shown in the following formula:

[0053] ;

[0054] in, is the value of the chaotic map of the position of the i-th jellyfish, is the initial value of the chaotic map, ranging from [0,1], and ;

[0055] Further introducing the Cauchy mutation, the formula for the position of the jellyfish is updated to:

[0056] ;

[0057] ;

[0058] ;

[0059] in, is the scale parameter of the Cauchy distribution, The vector representing the direction in which the jellyfish moves in the kth iteration;

[0060] Each time the jellyfish updates its position, a reverse learning strategy is introduced to select a better solution by comparing the fitness function values. The formula for updating the jellyfish position is as follows:

[0061] ;

[0062] ;

[0063] in, and Respectively represent the minimum and maximum values ​​of the position after the kth iteration, represents the reverse solution of the position of the i-th jellyfish after the k-th iteration, and They represent the position of the i-th jellyfish after the k-th iteration and the fitness function value corresponding to its reverse solution;

[0064] Input the input data in the training data into the support vector machine model of the k+1th iteration, predict the output result, calculate the classification error rate of the support vector machine model of the k+1th iteration based on the output result and the classification label in the training data, and use the classification error rate of the support vector machine model of the k+1th iteration as the fitness function value of the jellyfish after the k+1th iteration. The formula for the classification error rate is as follows:

[0065] ;

[0066] in, Indicates the number of samples in the output that are consistent with the classification label. Represents the total number of samples in the training data;

[0067] If the fitness function value of the jellyfish after the k+1th iteration is less than the fitness function value of the jellyfish after the kth iteration, the fitness function value after the k+1th iteration replaces the fitness function value of the jellyfish after the kth iteration, otherwise, the fitness function value of the jellyfish after the kth iteration continues to be used;

[0068] Determine whether k+1 satisfies the maximum number of iterations M. If so, the iteration ends, and the fitness function value of the jellyfish of the last iteration is used as the optimal fitness function value of the jellyfish, and the penalty coefficient c of the support vector machine model and the kernel function parameter g of the support vector machine model corresponding to the optimal fitness function value of the jellyfish are determined. Otherwise, the number of iterations is increased by 1, and the above iterative process is repeated;

[0069] The optimized support vector machine model is constructed by the penalty coefficient c of the support vector machine model corresponding to the optimal jellyfish fitness function value and the kernel function parameter g of the support vector machine model.

[0070] Preferably, the SHAP algorithm is used to interpret the output results of the trained support vector machine model and calculate the SHAP value of each uncertainty factor feature, specifically including:

[0071] The calculation formula of the SHAP value of the uncertainty factor characteristic p of the three-level supply chain network is as follows:

[0072] ;

[0073] in, represents the feature subset that does not contain the uncertainty factor feature p of the three-level supply chain network, O represents the set of all feature subsets, represents the total number of uncertainties in the three-stage supply chain network, and “!” represents factorial. represents the index of the uncertainty factor feature p of the three-stage supply chain network, Indicates adding the uncertainty factor feature p of the three-level supply chain network to the feature subset The value predicted by the SHAP model, Representing feature subsets The corresponding SHAP model predicted value.

[0074] Preferably, the weight value of each uncertainty factor feature is calculated according to the SHAP value of each uncertainty factor feature, and the uncertainty probability is calculated according to the weight value of each uncertainty factor feature, specifically including:

[0075] The following formula is used to calculate the weight value of the uncertainty factor characteristic p of the three-level supply chain network:

[0076] ;

[0077] in, is the normalized weight value of the uncertainty factor feature p;

[0078] The uncertainty probability f is calculated using the following formula:

[0079] ;

[0080] in, It represents the random probability value of the uncertainty factor characteristic p occurring in the tth time period.

[0081] As a preferred method, the optimal solution of the objective function under the constraints is solved based on uncertainty probability to obtain the optimal supply chain solution, which specifically includes:

[0082] Substitute the uncertainty probability into the constraint conditions to obtain the constraint conditions under the known uncertainty probability, and solve the optimal solution of the objective function under the constraint conditions under the known uncertainty probability, and obtain the optimal supply chain solution through the optimal solution.

[0083] In a second aspect, the present invention provides a supply chain optimization device based on an IJS-SVM model, characterized in that it includes:

[0084] A supply chain network construction module is configured to construct a three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses, and multiple assembly sites, and establish an objective function of the three-level supply chain network and its corresponding constraints;

[0085] The model building module is configured to build a support vector machine model, optimize the support vector machine model by using an improved jellyfish search algorithm to obtain an optimized support vector machine model, train the optimized support vector machine model to obtain a trained support vector machine model, the input data of the trained support vector machine model is the uncertainty factor characteristics of the three-level supply chain network, and the output result of the trained support vector machine model is a classification result of whether the corresponding product is delayed in completing assembly or the product is completed on time;

[0086] An explanation module is configured to use a SHAP algorithm to explain the output results of the trained support vector machine model and calculate a SHAP value of each uncertainty factor feature;

[0087] The solution module is configured to calculate the weight value of each uncertain factor feature according to the SHAP value of each uncertain factor feature, calculate the uncertainty probability according to the weight value of each uncertain factor feature, and solve the optimal solution of the objective function under the constraint conditions based on the uncertainty probability to obtain the optimal supply chain solution.

[0088] In a third aspect, the present invention provides an electronic device comprising one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0089] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0090] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0091] Compared with the prior art, the present invention has the following beneficial effects:

[0092] (1) The supply chain optimization method based on the IJS-SVM model proposed in this paper combines the Logistic chaos mapping strategy, Cauchy mutation strategy and reverse learning strategy on the basis of the classic jellyfish search algorithm to obtain an improved jellyfish search algorithm to improve the solution speed and avoid local convergence.

[0093] (2) The supply chain optimization method based on the IJS-SVM model proposed in the present invention optimizes the support vector machine model through an improved jellyfish search algorithm, determines the optimal penalty coefficient c and kernel function parameter g, and trains the optimized support vector machine model to obtain a trained support vector machine model. The trained support vector machine model can classify the characteristics of uncertainty factors and obtain the classification results of whether the corresponding product is delayed in completing assembly or the product is completed on time.

[0094] (3) The supply chain optimization method based on the IJS-SVM model proposed in this invention analyzes the characteristics of uncertainty factors and their impacts in each link of the three-level supply chain through the SHAP interpretation method, and obtains the weight values ​​of these uncertainty factor characteristics. The uncertainty probability is calculated according to the weight value, and it is substituted into the constraint conditions to solve the optimal solution of the objective function under the constraint conditions, thereby obtaining the optimal supply chain solution, which can effectively solve the uncertainty problem in the supply chain process and has high practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments 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.

[0096] Figure 1 A schematic diagram of a supply chain optimization method based on an IJS-SVM model according to an embodiment of the present application;

[0097] Figure 2 A schematic diagram of a three-level supply chain network of a supply chain optimization method based on an IJS-SVM model according to an embodiment of the present application;

[0098] Figure 3 A schematic diagram of the flow of an IJS-SVM model of a supply chain optimization method based on an IJS-SVM model according to an embodiment of the present application;

[0099] Figure 4 A schematic diagram of a supply chain optimization device based on an IJS-SVM model according to an embodiment of the present application;

[0100] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0101] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0102] Figure 1 A supply chain optimization method based on the IJS-SVM model provided in an embodiment of the present application is shown, comprising the following steps:

[0103] S1, construct a three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses and multiple assembly sites, and establish the objective function of the three-level supply chain network and its corresponding constraints.

[0104] In a specific embodiment, the objective function of the three-level supply chain network is:

[0105] ;

[0106] Where Z represents the cost of the three-level supply chain network, W represents the total number of warehouses, , T represents the total number of time periods in a cycle, ; N represents the total number of manufacturing plants, ; S represents the total number of assembly sites, ; E represents the total number of products, ; represents the fixed cost of building the w-th warehouse, is 1 or 0. If the wth warehouse is established, then is 1, otherwise is 0; represents the unit production cost of the e-th product produced by the n-th manufacturing plant in the t-th period, represents the number of products in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period; represents the unit transportation cost of the e-th product from the n-th manufacturing plant to the w-th warehouse, represents the transportation volume of the e-th product from the n-th manufacturing plant to the w-th warehouse in the t-th period under the influence of uncertain factors, represents the unit transportation cost of shipping the e-th product from the w-th warehouse to the s-th assembly site in the t-th period, represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period under the influence of uncertain factors, represents the unit inventory cost of the remaining products of the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, represents the unit inventory cost of the remaining products in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, represents the number of products that were not delivered on time at the s-th assembly site in the t-th period, represents the cost caused by the untimely supply of the e-th product at the s-th assembly site during the t-th period, I represents the unit cost of the construction delay, Indicates the number of days the construction period is delayed;

[0107] The constraints are:

[0108] Production process:

[0109] ;

[0110] ;

[0111] ;

[0112] ;

[0113] Logistics warehouse link:

[0114] ;

[0115] ;

[0116] ;

[0117] ;

[0118] ;

[0119] Assembly site links:

[0120] ;

[0121] ;

[0122] ;

[0123] ;

[0124] ;

[0125] ;

[0126] ;

[0127] in, represents the quantity of the e-th product in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period, represents the unit volume of the e-th product, is 1 or 0. If the e-th product is produced by the n-th manufacturing plant in the t-th period, then is 1, otherwise is 0; represents the total available capacity of the nth manufacturing plant in the tth period, represents the setup time for the production of the e-th product in the n-th manufacturing plant, and its average value is , the standard deviation is ; represents the transportation volume of the e-th product shipped from the n-th manufacturing plant to the w-th warehouse in the t-th period without considering the uncertainty factors. represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period without considering the uncertainty factors, represents the number of e-th products that the n-th manufacturing plant can produce in parallel, It represents the production cycle of the e-th product produced by the n-th manufacturing plant, and its average value is , the standard deviation is ; f represents the uncertainty probability, represents the total available time of the nth manufacturing plant in the tth period, and the average value is , the standard deviation is ; represents the total amount of products delivered by the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the t-1th period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, is 1 or 0. In the tth period, if the vehicle has the ability to successfully transport the eth product from the nth manufacturing plant to the wth warehouse, then is 1, otherwise is 0; represents the shipping vehicle capacity from the nth manufacturing plant to the wth warehouse, represents the unit area of ​​the e-th product, represents the remaining storage area of ​​the w-th warehouse in the t-th period, represents the weight of the e-th product, represents the number of shelves remaining in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-1th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, is 1 or 0. If the wth warehouse provides the necessary product equipment for the sth assembly site, then is 1, otherwise is 0; represents the number of products assembled at the s-th assembly site during the t-th period. represents the standard deviation of the assembly quantity of the e-th product at the s-th assembly site in the t-th period, represents the average number of assembly quantities of the e-th product at the s-th assembly site in the t-th period, represents the capacity of the shipping vehicle from the w-th warehouse to the s-th assembly site, represents the number of products that were not delivered on time at the s-th assembly site in the t-1th period, represents the number of products that were not delivered on time at the s-th assembly site in the t-th period, Represents the basic planning cycle of an engineering project. Indicates the time required for engineering simulation to complete all tasks.

[0128] Specifically, refer to Figure 2, construct a multi-product, multi-stage three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses and multiple assembly sites, and establish the objective function and constraints of the three-level supply chain network. The objective function of the embodiment of the present application is to minimize the cost of the entire supply chain. The cost comprehensively considers multiple cost factors, including the fixed cost of warehouse construction, production cost, transportation cost, inventory holding cost of manufacturing plants and warehouses, and out-of-stock costs and other operating costs caused by failure to meet assembly site needs in a timely manner. In order to take these costs into account, it is necessary to establish a system simulation model, that is, the three-level supply chain network model, which ensures the feasibility and efficiency of the supply chain through a series of constraints.

[0129] The constraints include: limiting the production capacity and production time of the manufacturing plant; specifying the delivery quantity and inventory level of the manufacturing plant; ensuring that the transportation volume from the manufacturing plant to the warehouse does not exceed the capacity limit of the vehicle; limiting the number of products transported to the assembly site; specifying the inventory level of the warehouse in different time periods; and ensuring that more than one warehouse can serve an assembly site. In addition, ensure that the equipment demand of the assembly site in any time period does not exceed the total capacity of the warehouse serving the site; limit the number of products allocated to each assembly site in a specific time period to not exceed its demand. Describes the actual number of products delivered to the warehouse and assembly site in different time periods, and its relationship with the equipment shortage. Finally, ensure that the entire project is completed within the basic planning cycle to avoid additional costs due to extended construction period.

[0130] S2, construct a support vector machine model, optimize the support vector machine model through the improved jellyfish search algorithm to obtain an optimized support vector machine model, train the optimized support vector machine model to obtain a trained support vector machine model, the input data of the trained support vector machine model is the uncertainty factor characteristics of the three-level supply chain network, and the output result of the trained support vector machine model is the classification result of whether the corresponding product is delayed in completing assembly or the product is completed on time.

[0131] In a specific embodiment, the support vector machine model is optimized by using an improved jellyfish search algorithm to obtain an optimized support vector machine model, which specifically includes:

[0132] Collect training data. The input data in the training data are the features of the uncertainty factors, and the output labels are 0 and 1, where 0 means that the product is assembled late, and 1 means that the product is assembled on time.

[0133] Initialize the penalty coefficient c of the support vector machine model, the kernel function parameter g of the support vector machine model, the population size Q of jellyfish, the maximum number of iterations M, and the upper limit U of the independent variable b and the lower limit of the independent variable L b ;

[0134] The initialized penalty coefficient c and kernel function parameter g are used as the parameters of the support vector machine model for the kth iteration to start the following iterative process:

[0135] Jellyfish use a timing control mechanism to switch between following the ocean current and swimming autonomously;

[0136] In the phase of following the ocean current, the position of the jellyfish of the current iteration is updated using the following formula:

[0137] ;

[0138] in, is the position of the i-th jellyfish after the k-th iteration, is the position of the i-th jellyfish after the k+1th iteration, is the optimal position of the jellyfish in the population after the kth iteration, is the distribution coefficient, Represents a random number between 0 and 1. represents the average position of all jellyfish in the population;

[0139] In the autonomous swimming stage, if the jellyfish performs passive movement, the position of the jellyfish of the current iteration is updated using the following formula:

[0140] ;

[0141] in, is the coefficient of motion;

[0142] If the jellyfish is actively moving, the movement direction and position update formula are as follows:

[0143] ;

[0144] ;

[0145] in, Indicates the direction of movement, is a jellyfish randomly selected from the population after the kth iteration, represents the fitness function;

[0146] The movement of jellyfish is regulated by the time control mechanism, and the control function formula is:

[0147] ;

[0148] in, is a random value in [0,1] that changes with the number of iterations, M is the maximum number of iterations, is the threshold constant, if , jellyfish follow the ocean currents; otherwise, jellyfish perform active and passive movements;

[0149] The Logistic chaotic map is introduced to initialize the jellyfish population, as shown in the following formula:

[0150] ;

[0151] in, is the value of the chaotic map of the position of the i-th jellyfish, is the initial value of the chaotic map, ranging from [0,1], and ;

[0152] Further introducing the Cauchy mutation, the formula for the position of the jellyfish is updated to:

[0153] ;

[0154] ;

[0155] ;

[0156] in, is the scale parameter of the Cauchy distribution, The vector representing the direction in which the jellyfish moves in the kth iteration;

[0157] Each time the jellyfish updates its position, a reverse learning strategy is introduced to select a better solution by comparing the fitness function values. The formula for updating the jellyfish position is as follows:

[0158] ;

[0159] ;

[0160] in, and Respectively represent the minimum and maximum values ​​of the position after the kth iteration, represents the reverse solution of the position of the i-th jellyfish after the k-th iteration, and They represent the position of the i-th jellyfish after the k-th iteration and the fitness function value corresponding to its reverse solution;

[0161] Input the input data in the training data into the support vector machine model of the k+1th iteration, predict the output result, calculate the classification error rate of the support vector machine model of the k+1th iteration based on the output result and the classification label in the training data, and use the classification error rate of the support vector machine model of the k+1th iteration as the fitness function value of the jellyfish after the k+1th iteration. The formula for the classification error rate is as follows:

[0162] ;

[0163] in, Indicates the number of samples in the output that are consistent with the classification label. Represents the total number of samples in the training data;

[0164] If the fitness function value of the jellyfish after the k+1th iteration is less than the fitness function value of the jellyfish after the kth iteration, the fitness function value after the k+1th iteration replaces the fitness function value of the jellyfish after the kth iteration, otherwise, the fitness function value of the jellyfish after the kth iteration continues to be used;

[0165] Determine whether k+1 satisfies the maximum number of iterations M. If so, the iteration ends, and the fitness function value of the jellyfish of the last iteration is used as the optimal fitness function value of the jellyfish, and the penalty coefficient c of the support vector machine model and the kernel function parameter g of the support vector machine model corresponding to the optimal fitness function value of the jellyfish are determined. Otherwise, the number of iterations is increased by 1, and the above iterative process is repeated;

[0166] The optimized support vector machine model is constructed by the penalty coefficient c of the support vector machine model corresponding to the optimal jellyfish fitness function value and the kernel function parameter g of the support vector machine model.

[0167] Specifically, the embodiment of the present application uses a support vector machine optimized based on an improved jellyfish search algorithm to quantify uncertainty factors, thereby optimizing supply chain costs. The jellyfish search algorithm (JS algorithm) determines the optimal location by simulating the movement of jellyfish in the ocean and the phenomenon of jellyfish tides formed by ocean currents. The algorithm assumes the following:

[0168] 1. Jellyfish use a timing control mechanism to switch between following the ocean current and swimming independently.

[0169] 2. Jellyfish are attracted to locations where food is plentiful.

[0170] 3. The amount of food a jellyfish finds is determined by its position and the value of the fitness function.

[0171] In the phase of following the current, the jellyfish are attracted by the nutrients in the current, and the direction of the current is determined by the average vector from each jellyfish to the current best position. After a period of time, a population of jellyfish gradually forms, and individual jellyfish swim around their own positions (passive movement). The position update formula in It is related to the movement length, and is obtained through experimental analysis. At the same time, the jellyfish will move towards their companions who have more food (active movement).

[0172] As the jellyfish move toward the ocean current, a population of jellyfish gradually forms. The jellyfish first move passively, and then over time, they move toward where there is more food (active movement). The movement of the jellyfish is regulated by the time control mechanism, and the control function formula is .

[0173] In the jellyfish search algorithm, Logistic chaos mapping is introduced to initialize the jellyfish population. In order to prevent the jellyfish search algorithm from falling into the local optimal solution, Cauchy mutation and reverse learning strategy are introduced. Among them, the role of Cauchy mutation is to expand the search space and enhance the ability of the population to jump out of the local optimal solution; the reverse learning strategy is introduced every time the jellyfish updates its position, and a better solution is selected by comparing the fitness function value.

[0174] The support vector machine (SVM) model is a commonly used classification model. Its main idea is to find a decision boundary (usually a hyperplane) that maximizes the interval between two categories and makes the sample point closest to the plane have the farthest distance to the plane.

[0175] Based on the IJS principle process, it can be found that the penalty coefficient c and kernel function parameter g have a great influence on the classification results of sample data. In view of the fast convergence speed and accurate search for the optimal solution of the improved jellyfish search algorithm, it is used to optimize the penalty coefficient c and kernel function parameter g of the support vector machine model. The SVM model is constructed by these two optimal parameters optimized by the IJS algorithm to obtain the IJS-SVM model. The importance of supply chain uncertainty factors is quantitatively measured by the SHAP value, and the results of these weight normalization are input into the three-level supply chain network to approach the real simulation environment. When the initialized penalty coefficient c and kernel function parameter g are used as the parameters of the support vector machine model for the kth iteration, the corresponding fitness function value after the kth iteration is 1.

[0176] S3, the SHAP algorithm is used to interpret the output results of the trained support vector machine model and calculate the SHAP value of each uncertainty factor feature.

[0177] In a specific embodiment, step S3 specifically includes:

[0178] The calculation formula of the SHAP value of the uncertainty factor characteristic p of the three-level supply chain network is as follows:

[0179] ;

[0180] in, represents the feature subset that does not contain the uncertainty factor feature p of the three-level supply chain network, O represents the set of all feature subsets, represents the total number of uncertainties in the three-stage supply chain network, and “!” represents factorial. represents the index of the uncertainty factor feature p of the three-stage supply chain network, Indicates adding the uncertainty factor feature p of the three-level supply chain network to the feature subset The value predicted by the SHAP model, Representing feature subsets The corresponding SHAP model predicted value.

[0181] Specifically, the SHAP (Shapley Additive Explanations) algorithm is a method used to explain machine learning models, which can be used to measure the contribution of each feature.

[0182] S4, calculate the weight value of each uncertain factor feature according to the SHAP value of each uncertain factor feature, calculate the uncertainty probability according to the weight value of each uncertain factor feature, and solve the optimal solution of the objective function under the constraint conditions based on the uncertainty probability to obtain the optimal supply chain solution.

[0183] In a specific embodiment, the weight value of each uncertainty factor feature is calculated according to the SHAP value of each uncertainty factor feature, and the uncertainty probability is calculated according to the weight value of each uncertainty factor feature, specifically including:

[0184] The following formula is used to calculate the weight value of the uncertainty factor characteristic p of the three-level supply chain network:

[0185] ;

[0186] in, is the normalized weight value of the uncertainty factor feature p;

[0187] The uncertainty probability f is calculated using the following formula:

[0188] ;

[0189] in, It represents the random probability value of the uncertainty factor characteristic p occurring in the tth time period.

[0190] Specifically, by bringing the weight values ​​of the uncertainty factor characteristics into the calculation formula of the uncertainty probability, we can obtain the probability of occurrence of supply chain uncertainty factors, that is, the uncertainty probability, which in turn affects the entire supply chain process.

[0191] In a specific embodiment, solving the optimal solution of the objective function under the constraint conditions based on uncertainty probability to obtain the optimal supply chain solution specifically includes:

[0192] Substitute the uncertainty probability into the constraint conditions to obtain the constraint conditions under the known uncertainty probability, and solve the optimal solution of the objective function under the constraint conditions under the known uncertainty probability, and obtain the optimal supply chain solution through the optimal solution.

[0193] Specifically, refer to Figure 3 ,The steps of optimizing support vector machine by improved jellyfish search algorithm are as follows:

[0194] Step 1: Obtain demand changes, weather conditions, production links, transportation distances, traffic conditions and other factors as uncertainty factor features in the training data, and the progress in the training data as classification labels, where the classification labels are 0 (delayed assembly completion) and 1 (on-time assembly completion) to distinguish different assembly results.

[0195] Step 2: Select a training set and a test set from the training data. The embodiment of the present application uses 10-fold cross validation to process the training data.

[0196] Step 3: Initialize the parameters of the SVM model (c, g), the jellyfish population size Q, the maximum number of iterations M, and the upper limit of the independent variable U b and the lower limit of the independent variable L b .

[0197] Step 4: Use the above method to update the position of the jellyfish.

[0198] Step 5: Calculate the fitness function value and use the classification error rate as the fitness function value. If the current jellyfish's fitness function value is less than the saved fitness function value, replace the previously saved fitness function value with the current jellyfish's fitness function value. Otherwise, use the original fitness function value and record the parameters (c, g) corresponding to the optimal fitness function value at this time. This step is executed until the iteration ends or the end condition is met.

[0199] Step 6: Use the optimal parameters (c, g) to build the optimized SVM model.

[0200] Step 7: Train the training data with the optimized SVM model to obtain the trained SVM model. Use the SHAP algorithm to interpret the trained SVM model to obtain the SHAP value of each uncertainty factor feature, normalize it, and obtain the weight value of the uncertainty factor feature of the supply chain.

[0201] Step 8: Calculate the uncertainty probability according to the weight value of the uncertainty factor characteristics, and substitute it into the constraint conditions to solve the optimal solution of the objective function formula under the constraint conditions, so as to obtain the optimal supply chain solution.

[0202] The supply chain optimization method based on the IJS-SVM model proposed in the embodiment of the present application has significant advantages in the field of simulation technology and successfully overcomes the above-mentioned problems. First, the embodiment of the present application significantly improves the accuracy and reliability of the model. Secondly, the interpretability and practicality of the simulation results have been greatly improved, allowing managers to understand and apply these results more intuitively. Furthermore, the embodiment of the present application optimizes computing efficiency, reduces implementation costs, and enables more companies to afford and apply this technology. In addition, the embodiment of the present application also simplifies the operating process, so that even users who lack professional knowledge can easily get started, which comprehensively improves the popularity of the technology.

[0203] Using the simulation results, managers can optimize the management strategy of each link in the supply chain to minimize costs and maximize profits. Through the simulation technology of the embodiments of the present application, enterprises can more effectively cope with the challenges in supply chain management, improve overall operational efficiency, and ensure that they maintain a leading position in the fiercely competitive market.

[0204] Further references Figure 4 As an implementation of the methods shown in the above figures, the present application provides an embodiment of a supply chain optimization device based on an IJS-SVM model. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0205] The embodiment of the present application provides a supply chain optimization device based on the IJS-SVM model, including:

[0206] The supply chain network construction module 1 is configured to construct a three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses, and multiple assembly sites, and establish an objective function of the three-level supply chain network and its corresponding constraints;

[0207] Model building module 2 is configured to build a support vector machine model, optimize the support vector machine model by using an improved jellyfish search algorithm to obtain an optimized support vector machine model, train the optimized support vector machine model to obtain a trained support vector machine model, the input data of the trained support vector machine model is the uncertainty factor characteristics of the three-level supply chain network, and the output result of the trained support vector machine model is the classification result of whether the corresponding product is delayed in completing assembly or the product is completed on time;

[0208] An interpretation module 3 is configured to interpret the output results of the trained support vector machine model using a SHAP algorithm and calculate a SHAP value of each uncertainty factor feature;

[0209] Solving module 4 is configured to calculate the weight value of each uncertain factor feature according to the SHAP value of each uncertain factor feature, calculate the uncertainty probability according to the weight value of each uncertain factor feature, and solve the optimal solution of the objective function under the constraint conditions based on the uncertainty probability to obtain the optimal supply chain solution.

[0210] Figure 5 Schematic diagram of the hardware structure of the electronic device provided by the embodiment of the present invention. Figure 5 As shown, the electronic device of this embodiment includes: a processor 501 and a memory 502; wherein the memory 502 is used to store computer-executable instructions; the processor 501 is used to execute the computer-executable instructions stored in the memory to implement the various steps performed by the electronic device in the above embodiment. For details, please refer to the relevant description in the above method embodiment.

[0211] Optionally, the memory 502 may be independent or integrated with the processor 501 .

[0212] When the memory 502 is independently provided, the electronic device further includes a bus 503 for connecting the memory 502 and the processor 501 .

[0213] The embodiment of the present invention further provides a computer storage medium, in which computer execution instructions are stored. When the processor 501 executes the computer execution instructions, the above method is implemented.

[0214] The embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by the processor 501, the above method is implemented.

[0215] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0216] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to implement the solution of this embodiment.

[0217] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each module may exist physically separately, or two or more modules may be integrated into one unit. The above-mentioned module-composed unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0218] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor 501 to perform some steps of the methods of various embodiments of the present application.

[0219] It should be understood that the processor 501 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or the processor 501 may be any conventional processor 501, etc. The steps of the method disclosed in the invention may be directly embodied in the execution of the hardware processor 501, or may be executed by a combination of hardware and software modules in the processor 501.

[0220] The memory 502 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk, or an optical disk.

[0221] The bus 503 may be an Industry Standard Architecture (ISA), a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus 503 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus 503 in the drawings of the present application is not limited to only one bus 503 or one type of bus 503.

[0222] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0223] An exemplary storage medium is coupled to the processor 501, so that the processor 501 can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor 501. The processor 501 and the storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor 501 and the storage medium can also exist as discrete components in an electronic device or a main control device.

[0224] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0225] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A supply chain optimization method based on the IJS-SVM model, characterized in that: The following steps are involved: A three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses, and multiple assembly sites is constructed, and the objective function of the three-level supply chain network and its corresponding constraints are established. The objective function of the three-level supply chain network is: Wherein, Z represents the cost of the three-level supply chain network, W represents the total number of warehouses, w=1, 2, ..., W, T represents the total number of time periods in a cycle, t=1, 2, ..., T; N represents the total number of manufacturing plants, n=1, 2, ..., N; S represents the total number of assembly sites, s=1, 2, ..., S; E represents the total number of products, e=1, 2, ..., E; F w represents the fixed cost of building the wth warehouse, A w is 1 or 0. If the wth warehouse is established, then A w is 1, otherwise A w is 0; represents the unit production cost of the e-th product produced by the n-th manufacturing plant in the t-th period, represents the quantity of the e-th product in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period; represents the unit transportation cost of shipping the e-th product from the n-th manufacturing plant to the w-th warehouse in the t-th period, represents the transportation volume of the e-th product from the n-th manufacturing plant to the w-th warehouse in the t-th period under the influence of uncertain factors, represents the unit transportation cost of shipping the e-th product from the w-th warehouse to the s-th assembly site in the t-th period, represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period under the influence of uncertain factors, represents the unit inventory cost of the remaining products of the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, represents the unit inventory cost of the remaining products in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, represents the number of products that were not delivered on time at the s-th assembly site in the t-th period, represents the cost caused by the untimely supply of the e-th product at the s-th assembly site during the t-th period, I represents the unit cost of the construction delay, and t′ represents the number of days of the construction delay; The constraints are: Production process: Logistics warehouse link: Assembly site links: T bp ≤T b ; in, Indicates any V represents the quantity of the e-th product in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period. e represents the unit volume of the e-th product, is 1 or 0. If the e-th product is produced by the n-th manufacturing plant in the t-th period, then is 1, otherwise =0; V nt represents the total available capacity of the nth manufacturing plant in the tth period, represents the setup time for the production of the e-th product in the n-th manufacturing plant, and its average value is represents the transportation volume of the e-th product shipped from the n-th manufacturing plant to the w-th warehouse in the t-th period without considering the uncertainty factors. represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period without considering the uncertainty factors, represents the number of e-th products that the n-th manufacturing plant can produce in parallel, It represents the production cycle of the e-th product produced by the n-th manufacturing plant, and its average value is f represents the uncertainty probability, t nt It represents the total available time of the nth manufacturing plant in the tth period, with an average value of μt nt ; represents the total amount of products delivered by the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the t-1th period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, is 1 or 0. In the tth period, if the vehicle has the ability to successfully transport the eth product from the nth manufacturing plant to the wth warehouse, then is 1, otherwise =0; G nw represents the shipping vehicle capacity from the nth manufacturing plant to the wth warehouse, S e represents the unit area of ​​the e-th product, S′ wt W represents the remaining storage area of ​​the w-th warehouse in the t-th period, e represents the weight of the e-th product, N′ wt represents the number of shelves remaining in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-1th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, Y sw is 1 or 0. If the wth warehouse provides the necessary product equipment for the sth assembly site, then Y sw is 1, otherwise Y sw is 0; represents the number of products assembled at the s-th assembly site during the t-th period. represents the standard deviation of the assembly quantity of the e-th product at the s-th assembly site in the t-th period, G represents the average number of assembly quantities of the e-th product at the s-th assembly site in the t-th period, ws represents the capacity of the shipping vehicle from the w-th warehouse to the s-th assembly site, represents the number of products that were not delivered on time at the s-th assembly site in the t-1th period, T represents the number of products that were not delivered on time at the s-th assembly site in the t-th period. b Represents the basic planning cycle of the project, T bp Indicates the time required for engineering simulation to complete all tasks; Construct a support vector machine model, and optimize the support vector machine model through the improved jellyfish search algorithm to obtain the optimized support vector machine model, which specifically includes: Collect training data, wherein the input data in the training data are features of uncertainty factors, and the output labels are 0 and 1, where 0 indicates that the product is assembled late, and 1 indicates that the product is assembled on time; Initialize the penalty coefficient c of the support vector machine model, the kernel function parameter g of the support vector machine model, the population size Q of jellyfish, the maximum number of iterations M, and the upper limit U of the independent variable b and the lower limit of the independent variable L b ; The initialized penalty coefficient c and kernel function parameter g are used as the parameters of the support vector machine model for the kth iteration to start the following iterative process: Jellyfish use a timing control mechanism to switch between following the ocean current and swimming autonomously; In the phase of following the ocean current, the position of the jellyfish of the current iteration is updated using the following formula: in, is the position of the i-th jellyfish after the k-th iteration, is the position of the i-th jellyfish after the k+1th iteration, is the optimal position of the jellyfish in the population after the kth iteration, β is the distribution coefficient, rand(0,1) represents a random number between 0 and 1, and τ represents the average position of all jellyfish in the population; In the autonomous swimming stage, if the jellyfish performs passive movement, the position of the jellyfish of the current iteration is updated using the following formula: Among them, γ is the motion coefficient; If the jellyfish is actively moving, the movement direction and position update formula are as follows: in, Indicates the direction of movement, is a jellyfish randomly selected from the population after the kth iteration, and fin(·) represents the fitness function; The movement of jellyfish is regulated by the time control mechanism, and the control function formula is: Where C(k) is a random value between [0,1] that changes with the number of iterations, M is the maximum number of iterations, C0 is a threshold constant, if C(k) ≥ C0, the jellyfish moves with the ocean current; otherwise, the jellyfish performs active and passive movements; The Logistic chaotic map is introduced to initialize the jellyfish population, as shown in the following formula: X i+1 =ηX i (1-X i ),0≤X0≤1; Among them, X i is the value of the chaotic map of the position of the i-th jellyfish, X0 is the initial value of the chaotic map, the range is [0,1], and η=4; Further introducing the Cauchy mutation, the formula for the position of the jellyfish is updated to: Where λ is the scale parameter of the Cauchy distribution, The vector representing the direction in which the jellyfish moves in the kth iteration; Each time the jellyfish updates its position, a reverse learning strategy is introduced to select a better solution by comparing the fitness function values. The formula for updating the jellyfish position is as follows: in, and Respectively represent the minimum and maximum values ​​of the position after the kth iteration, X oppo represents the reverse solution of the position of the i-th jellyfish after the k-th iteration, and fin(X oppo ) represent the position of the i-th jellyfish after the k-th iteration and the fitness function value corresponding to its reverse solution; Input data in the training data is input into the support vector machine model of the k+1th iteration, and an output result is obtained by prediction. The classification error rate of the support vector machine model of the k+1th iteration is calculated based on the output result and the classification label in the training data. The classification error rate of the support vector machine model of the k+1th iteration is used as the fitness function value of the jellyfish after the k+1th iteration. The formula of the classification error rate is as follows: Among them, n acc Indicates the number of samples in the output that are consistent with the classification label, n total Represents the total number of samples in the training data; If the fitness function value of the jellyfish after the k+1th iteration is less than the fitness function value of the jellyfish after the kth iteration, the fitness function value after the k+1th iteration replaces the fitness function value of the jellyfish after the kth iteration, otherwise, the fitness function value of the jellyfish after the kth iteration continues to be used; Determine whether k+1 satisfies the maximum number of iterations M. If so, the iteration ends, and the fitness function value of the jellyfish of the last iteration is used as the optimal fitness function value of the jellyfish, and the penalty coefficient c of the support vector machine model and the kernel function parameter g of the support vector machine model corresponding to the optimal fitness function value of the jellyfish are determined. Otherwise, the number of iterations is increased by 1, and the above iterative process is repeated; An optimized support vector machine model is constructed by using the penalty coefficient c of the support vector machine model corresponding to the fitness function value of the optimal jellyfish and the kernel function parameter g of the support vector machine model; the optimized support vector machine model is trained to obtain a trained support vector machine model, the input data of the trained support vector machine model is the uncertainty factor characteristics of the three-level supply chain network, and the output result of the trained support vector machine model is the classification result of whether the corresponding product is assembled late or the product is assembled on time; The output of the trained support vector machine model is interpreted using the SHAP algorithm to calculate the SHAP value of each uncertainty factor feature; The weight value of each uncertain factor feature is calculated according to the SHAP value of each uncertain factor feature, the uncertainty probability is calculated according to the weight value of each uncertain factor feature, and the optimal solution of the objective function under the constraints is solved based on the uncertainty probability to obtain the optimal supply chain solution.

2. The supply chain optimization method based on the IJS-SVM model according to claim 1 is characterized in that: The SHAP algorithm is used to interpret the output results of the trained support vector machine model and calculate the SHAP value of each uncertainty factor feature, specifically including: The calculation formula of the SHAP value of the uncertainty factor characteristic p of the three-level supply chain network is as follows: Among them, Ω represents the feature subset that does not include the uncertainty factor feature p of the three-level supply chain network, O represents the set of all feature subsets, ε represents the total number of uncertainty factor features of the three-level supply chain network, "!" represents factorial, ρ represents the index of the uncertainty factor feature p of the three-level supply chain network, SHAP(Ω∪{ρ}) represents the value predicted by the SHAP model after adding the uncertainty factor feature p of the three-level supply chain network to the feature subset Ω, and SHAP(Ω) represents the value predicted by the SHAP model corresponding to the feature subset Ω.

3. The supply chain optimization method based on the IJS-SVM model according to claim 2 is characterized in that: Calculating the weight value of each uncertainty factor feature according to the SHAP value of each uncertainty factor feature, and calculating the uncertainty probability according to the weight value of each uncertainty factor feature, specifically including: The weight value of the uncertainty factor characteristic p of the three-level supply chain network is calculated using the following formula: Among them, P′ ρ is the normalized weight value of the uncertainty factor feature p; The uncertainty probability f is calculated using the following formula: Among them, r ρt It represents the random probability value of the uncertainty factor characteristic p occurring in the tth time period.

4. The supply chain optimization method based on the IJS-SVM model according to claim 1 is characterized in that: Solving the optimal solution of the objective function under the constraints based on the uncertainty probability to obtain the optimal supply chain solution specifically includes: Substitute the uncertainty probability into the constraint condition to obtain the constraint condition under the condition of known uncertainty probability, and solve the optimal solution of the objective function under the constraint condition under the condition of known uncertainty probability, and obtain the optimal supply chain solution through the optimal solution.

5. A supply chain optimization device based on the IJS-SVM model, characterized in that: include: The supply chain network construction module is configured to construct a three-level supply chain network consisting of multiple manufacturing plants, multiple warehouses, and multiple assembly sites, and establish the objective function of the three-level supply chain network and its corresponding constraints. The objective function of the three-level supply chain network is: Wherein, Z represents the cost of the three-level supply chain network, W represents the total number of warehouses, w=1, 2, ..., W, T represents the total number of time periods in a cycle, t=1, 2, ..., T; N represents the total number of manufacturing plants, n=1, 2, ..., N; S represents the total number of assembly sites, s=1, 2, ..., S; E represents the total number of products, e=1, 2, ..., E; F w represents the fixed cost of building the wth warehouse, A w is 1 or 0. If the wth warehouse is established, then A w is 1, otherwise A w is 0; represents the unit production cost of the e-th product produced by the n-th manufacturing plant in the t-th period, represents the quantity of the e-th product in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period; represents the unit transportation cost of shipping the e-th product from the n-th manufacturing plant to the w-th warehouse in the t-th period, represents the transportation volume of the e-th product from the n-th manufacturing plant to the w-th warehouse in the t-th period under the influence of uncertain factors, represents the unit transportation cost of shipping the e-th product from the w-th warehouse to the s-th assembly site in the t-th period, represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period under the influence of uncertain factors, represents the unit inventory cost of the remaining products of the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, represents the unit inventory cost of the remaining products in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, represents the number of products that were not delivered on time at the s-th assembly site in the t-th period, represents the cost caused by the untimely supply of the e-th product at the s-th assembly site during the t-th period, I represents the unit cost of the construction delay, and t′ represents the number of days of the construction delay; The constraints are: Production process: Logistics warehouse link: Assembly site links: T bp ≤T b ; in, Indicates any V represents the quantity of the e-th product in the order placed by the s-th assembly site to the n-th manufacturing plant in the t-th period. e represents the unit volume of the e-th product, is 1 or 0. If the e-th product is produced by the n-th manufacturing plant in the t-th period, then is 1, otherwise =0; V nt represents the total available capacity of the nth manufacturing plant in the tth period, represents the setup time for the production of the e-th product in the n-th manufacturing plant, and its average value is represents the transportation volume of the e-th product shipped from the n-th manufacturing plant to the w-th warehouse in the t-th period without considering the uncertainty factors. represents the transportation volume of the e-th product from the w-th warehouse to the s-th assembly site in the t-th period without considering the uncertainty factors, represents the number of e-th products that the n-th manufacturing plant can produce in parallel, It represents the production cycle of the e-th product produced by the n-th manufacturing plant, and its average value is f represents the uncertainty probability, t nt It represents the total available time of the nth manufacturing plant in the tth period, with an average value of μt nt ; represents the total amount of products delivered by the nth manufacturing plant in the tth period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the t-1th period, represents the remaining inventory of products that the nth manufacturing plant does not need to ship to the warehouse at the end of the tth period, is 1 or 0. In the tth period, if the vehicle has the ability to successfully transport the eth product from the nth manufacturing plant to the wth warehouse, then is 1, otherwise =0; G nw represents the shipping vehicle capacity from the nth manufacturing plant to the wth warehouse, S e represents the unit area of ​​the e-th product, S′ wt W represents the remaining storage area of ​​the w-th warehouse in the t-th period, e represents the weight of the e-th product, N′ wt represents the number of shelves remaining in the w-th warehouse in the t-th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-1th period, represents the remaining inventory of products that do not need to be delivered by the w-th warehouse in the t-th period, Y sw is 1 or 0. If the wth warehouse provides the necessary product equipment for the sth assembly site, then Y sw is 1, otherwise Y sw is 0; represents the number of products assembled at the s-th assembly site during the t-th period. represents the standard deviation of the assembly quantity of the e-th product at the s-th assembly site in the t-th period, G represents the average number of assembly quantities of the e-th product at the s-th assembly site in the t-th period, ws represents the capacity of the shipping vehicle from the w-th warehouse to the s-th assembly site, represents the number of products that were not delivered on time at the s-th assembly site in the t-1th period, T represents the number of products that were not delivered on time at the s-th assembly site in the t-th period. b Represents the basic planning cycle of the project, T bp Indicates the time required for engineering simulation to complete all tasks; The model building module is configured to build a support vector machine model, optimize the support vector machine model through an improved jellyfish search algorithm, and obtain an optimized support vector machine model, which specifically includes: Collect training data, wherein the input data in the training data are features of uncertainty factors, and the output labels are 0 and 1, where 0 indicates that the product is assembled late, and 1 indicates that the product is assembled on time; Initialize the penalty coefficient c of the support vector machine model, the kernel function parameter g of the support vector machine model, the population size Q of jellyfish, the maximum number of iterations M, and the upper limit U of the independent variable b and the lower limit of the independent variable L b ; The initialized penalty coefficient c and kernel function parameter g are used as the parameters of the support vector machine model for the kth iteration to start the following iterative process: Jellyfish use a timing control mechanism to switch between following the ocean current and swimming autonomously; In the phase of following the ocean current, the position of the jellyfish of the current iteration is updated using the following formula: in, is the position of the i-th jellyfish after the k-th iteration, is the position of the i-th jellyfish after the k+1th iteration, is the optimal position of the jellyfish in the population after the kth iteration, β is the distribution coefficient, rand(0,1) represents a random number between 0 and 1, and τ represents the average position of all jellyfish in the population; In the autonomous swimming stage, if the jellyfish performs passive movement, the position of the jellyfish of the current iteration is updated using the following formula: Among them, γ is the motion coefficient; If the jellyfish is actively moving, the movement direction and position update formula are as follows: in, Indicates the direction of movement, is a jellyfish randomly selected from the population after the kth iteration, and fin(·) represents the fitness function; The movement of jellyfish is regulated by the time control mechanism, and the control function formula is: Where C(k) is a random value between [0,1] that changes with the number of iterations, M is the maximum number of iterations, C0 is a threshold constant, if C(k) ≥ C0, the jellyfish moves with the ocean current; otherwise, the jellyfish performs active and passive movements; The Logistic chaotic map is introduced to initialize the jellyfish population, as shown in the following formula: X i+1 =ηX i (1-X i ),0≤X0≤1; Among them, X i is the value of the chaotic map of the position of the i-th jellyfish, X0 is the initial value of the chaotic map, the range is [0,1], and η=4; Further introducing the Cauchy mutation, the formula for the position of the jellyfish is updated to: Where λ is the scale parameter of the Cauchy distribution, The vector representing the direction in which the jellyfish moves in the kth iteration; Each time the jellyfish updates its position, a reverse learning strategy is introduced to select a better solution by comparing the fitness function values. The formula for updating the jellyfish position is as follows: in, and Respectively represent the minimum and maximum values ​​of the position after the kth iteration, X oppo represents the reverse solution of the position of the i-th jellyfish after the k-th iteration, and fin(X oppo ) represent the position of the i-th jellyfish after the k-th iteration and the fitness function value corresponding to its reverse solution; Input data in the training data is input into the support vector machine model of the k+1th iteration, and an output result is obtained by prediction. The classification error rate of the support vector machine model of the k+1th iteration is calculated based on the output result and the classification label in the training data. The classification error rate of the support vector machine model of the k+1th iteration is used as the fitness function value of the jellyfish after the k+1th iteration. The formula of the classification error rate is as follows: Among them, n acc Indicates the number of samples in the output that are consistent with the classification label, n total Represents the total number of samples in the training data; If the fitness function value of the jellyfish after the k+1th iteration is less than the fitness function value of the jellyfish after the kth iteration, the fitness function value after the k+1th iteration replaces the fitness function value of the jellyfish after the kth iteration, otherwise, the fitness function value of the jellyfish after the kth iteration continues to be used; Determine whether k+1 satisfies the maximum number of iterations M. If so, the iteration ends, and the fitness function value of the jellyfish of the last iteration is used as the optimal fitness function value of the jellyfish, and the penalty coefficient c of the support vector machine model and the kernel function parameter g of the support vector machine model corresponding to the optimal fitness function value of the jellyfish are determined. Otherwise, the number of iterations is increased by 1, and the above iterative process is repeated; An optimized support vector machine model is constructed by using the penalty coefficient c of the support vector machine model corresponding to the fitness function value of the optimal jellyfish and the kernel function parameter g of the support vector machine model; the optimized support vector machine model is trained to obtain a trained support vector machine model, the input data of the trained support vector machine model is the uncertainty factor characteristics of the three-level supply chain network, and the output result of the trained support vector machine model is the classification result of whether the corresponding product is assembled late or the product is assembled on time; An explanation module is configured to use a SHAP algorithm to explain the output results of the trained support vector machine model and calculate the SHAP value of each uncertainty factor feature; The solution module is configured to calculate the weight value of each uncertain factor feature according to the SHAP value of each uncertain factor feature, calculate the uncertainty probability according to the weight value of each uncertain factor feature, and solve the optimal solution of the objective function under the constraints based on the uncertainty probability to obtain the optimal supply chain solution.

6. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Citation Information

Patent Citations

  • Information pushing method and device, computer equipment and storage medium

    CN112256975A

  • Systems and methods for inventory management and optimization

    CN113228068A