Supply chain demand distribution solving method and device, equipment, storage medium and program product

By combining deep learning-based particle swarm algorithm and deep Q learning, the problem of low supply chain demand allocation efficiency in the existing technology is solved, more efficient supply-side demand allocation is achieved, and supply chain cost and flexibility are optimized.

CN118940987APending Publication Date: 2024-11-12GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410795831.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-19
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing supply chain demand allocation solution methods are inefficient, making it difficult to effectively manage product or service requirements, optimize inventory management and improve supply chain flexibility.

Method used

Using the deep learning-based particle swarm algorithm, by obtaining the supply demand allocation model, the supplier equipment is corresponding to the particles, and the position constraint range of the particles is determined according to the upper and lower limits of production efficiency, and using deep Q learning to optimize the particle swarm algorithm to realize the supply side Requirement allocation.

Benefits of technology

It significantly improves the solution efficiency of supply chain demand allocation, can more efficiently approach the global optimal position, optimize supply chain costs, and improve algorithm performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940987A_ABST
    Figure CN118940987A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of computer processing, provides a supply chain demand distribution solving method and device, equipment, a storage medium and a program product, and can efficiently solve a supply chain demand distribution result. Determining a particle swarm algorithm based on deep learning; obtaining a supply demand distribution model; and inputting production data of a supply side into the supply demand distribution model, and then solving by using a particle swarm algorithm based on deep learning to obtain a demand distribution result of the supply side.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer processing technology, and in particular to a method, apparatus, computer equipment, storage medium and computer program product for solving supply chain demand allocation. Background Art

[0002] Supply chain demand allocation is the process of reasonably allocating the demand for products or services in the supply chain according to market demand and resource supply, achieving effective management and allocation of product or service demand, helping to optimize inventory management, improve production efficiency, enhance supply chain flexibility, improve customer satisfaction, and reduce supply chain risks, thereby achieving efficient operation and sustainable development of the supply chain. However, the existing supply chain demand allocation solution methods still have certain limitations, and more efficient and feasible solution methods need to be further explored. Summary of the invention

[0003] Based on this, it is necessary to provide a method, device, computer equipment, storage medium and computer program product for solving supply chain demand allocation in response to the above technical problems.

[0004] The present application provides a method for solving supply chain demand allocation, the method comprising:

[0005] Determine the particle swarm algorithm based on deep learning;

[0006] Obtain supply demand allocation model;

[0007] After the production data on the supply side is input into the supply and demand allocation model, it is solved using a particle swarm algorithm based on deep learning to obtain the demand allocation result on the supply side.

[0008] In one embodiment, after the production data of the supply side is input into the supply demand allocation model, a particle swarm algorithm based on deep learning is used to solve the problem, including:

[0009] Correspond the supplier equipment in the supply chain system to the particles in the particle swarm algorithm;

[0010] Determine the position constraint range of each particle based on the upper and lower limits of the supplier's equipment production efficiency;

[0011] After the production data of the supply side is input into the supply demand allocation model, the supply chain demand allocation model is solved according to the position constraint range of each particle, the neural network corresponding to each particle and the particle swarm algorithm.

[0012] In one embodiment, the position constraint range of each particle is determined according to the upper and lower limits of the supplier's equipment production efficiency, including:

[0013] Obtain the upper and lower limits of the supplier's equipment production efficiency;

[0014] According to the upper limit value and the lower limit value, the position constraint range of each particle is obtained.

[0015] In one embodiment, mapping supplier devices in the supply chain system to particles in a particle swarm algorithm includes:

[0016] The number of supplier devices in the supply chain system is taken as the number of particles;

[0017] Each vendor device is mapped to a particle.

[0018] In one embodiment, the neural network is a deep Q-network.

[0019] The present application provides a supply chain demand allocation solution device, the device comprising:

[0020] An algorithm determination module, used to determine a particle swarm algorithm based on deep learning;

[0021] A model acquisition module is used to acquire a supply and demand allocation model;

[0022] The model solving module is used to input the production data of the supply side into the supply and demand allocation model, and then solve it using the particle swarm algorithm based on deep learning to obtain the demand allocation result of the supply side.

[0023] The present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the above method.

[0024] The present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to execute the above method.

[0025] The present application provides a computer program product, on which a computer program is stored, and the computer program is executed by a processor to perform the above method.

[0026] The above-mentioned supply chain demand allocation solution method, device, computer equipment, storage medium and computer program product determine the particle swarm algorithm based on deep learning; obtain the supply demand allocation model; after inputting the production data of the supply side into the supply demand allocation model, use the particle swarm algorithm based on deep learning to solve it, and obtain the demand allocation result of the supply side, which can solve the problem more efficiently and improve the solution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 An application environment diagram of a method for solving supply chain demand allocation in one embodiment;

[0029] Figure 2 A flowchart of a method for solving supply chain demand allocation in one embodiment;

[0030] Figure 3 A schematic diagram of the correspondence between the elements of a particle swarm algorithm and deep Q learning in one embodiment;

[0031] Figure 4 It is a structural block diagram of a solution device for supply chain demand allocation in one embodiment;

[0032] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0034] The solution method for supply chain demand allocation provided in this application involves a particle swarm algorithm based on deep learning and a supply demand allocation model, such as Figure 1 The method can be executed by a computer device, including Figure 2 The steps shown are:

[0035] Step S201, determining a particle swarm algorithm based on deep learning;

[0036] Among them, deep learning can be deep Q learning, and deep Q learning is used to optimize the particle swarm algorithm. Specifically, the optimization of the particle swarm algorithm mainly includes adjusting the inertia weight And the acceleration factor and ; Deep Q learning can be used to treat particles as intelligent agents and make them interact with the population environment. By learning the relationship between the state and action of particles and designing a reward mechanism, the particle swarm algorithm can dynamically adjust the parameter values ​​according to the optimal strategy obtained by reinforcement learning during operation. In this way, particles can maintain a balance between global and local searches and efficiently approach the global optimal position, thereby significantly improving the performance of the particle swarm algorithm. Accordingly, the particle swarm algorithm based on deep Q learning can be called a parameter-improved particle swarm algorithm based on deep Q learning (the English name is DQNPSO, the full name is Deep Q-Network Particle Swarm Optimization).

[0037] Combining deep Q learning with the particle swarm optimization algorithm, first discretize the search iteration process of the particle swarm optimization algorithm according to the idea of ​​reinforcement learning. On this basis, formulate elements such as state, action, and reward, as defined in detail as follows: Figure 3 shown.

[0038] (1) Discrete time

[0039] During training, the number of iterations of the particle swarm algorithm is used as the decision time node, that is, the deep Q learning algorithm runs with the iterations of the particle swarm algorithm.

[0040] (2) Status

[0041] The Euclidean distance between the current position of the particle and the global optimal position is used as the decision state. According to the different degrees of distance, it can be divided into four states: nearest, near, far and farthest. The criteria for distinguishing the decision space states are shown in Table 1:

[0042]

[0043] Table 1

[0044] (3) Action

[0045] The parameter that has the most significant impact on the performance of the particle swarm algorithm is as well as and , so the action space can be set as three-dimensional , the action value is between -1 and 1. According to the reasonable value range of the parameter, the action is mapped into the range using linear transformation. The range of is set to [0,2], and The range of is set to [0,4]. The linear transformation formula is described as:

[0046]

[0047] (4) Rewards

[0048] In the state Next action Get the next state When , the reward value of the environment feedback, the reward mechanism is set as:

[0049]

[0050] The reward criterion is whether the global optimal value changes, and positive rewards are given once a change occurs. This reward mechanism can quickly improve the optimization speed of the algorithm.

[0051] (5) Discount factor :

[0052] The discount factor is used to discount the rewards at other times. When the discount factor is larger, the algorithm pays more attention to the long-term impact; conversely, it pays more attention to the immediate reward.

[0053] In the process of combining deep Q learning with particle swarm algorithm, the number of particles in the population is quite large relative to the neural network. If so many particles are allowed to update a deep Q network together, the change of any particle will cause a large fluctuation in this unique network, and it is not possible to express the state difference between each particle well. Therefore, an independent neural network can be established for each particle as its own exploration experience for dynamic parameter selection. Doing so can enable each particle to truly represent an intelligent agent to explore the environment, which is conducive to improving the local search ability of the algorithm.

[0054] At the same time, this application proposes an experience sharing mechanism, which is mainly used to share the experience of particles in the global optimal position. This mechanism can promote mutual learning of neural networks between particles and make information exchange in the population more efficient. By quickly learning the optimal strategy, this experience sharing measure can enhance the global search ability of the algorithm.

[0055] The DQNPSO algorithm is mainly divided into two parts, the specific process is as follows:

[0056] Part 1: The process of training neural networks with deep Q learning.

[0057] 1. Initialize the experience pool RB and related training parameters, such as the current training times iterate=1, and the maximum training times is TrainMax;

[0058] 2. Randomly initialize the weights of the neural network , state action value function , initialize the target network weights ;

[0059] 3. Initialize particle swarm , calculate the fitness of each particle , and compare to get the global optimal position gbest, the current number of iterations , set the maximum number of iterations .

[0060] Enter the continuous training process until the number of training times reaches TrainMax. The training process is as follows:

[0061] 1. Initialize the state sequence;

[0062] 2. Based on probability Randomly select an action, calculate the reward value of the action, and move to the next state;

[0063] 3. The form of is stored in the experience pool RB;

[0064] 4. Randomly sample n quadruplets from the experience pool RB for neural network training;

[0065] 5. Calculate the reward value under the current state:

[0066] ;

[0067] 6. Minimize the loss function: , and use back propagation to update the neural network parameters;

[0068] 7. Update the target network by evaluating the network parameters at regular intervals.

[0069] Save the trained neural network weights for particle swarm algorithm optimization.

[0070] Part 2: Optimizing the Particle Swarm Algorithm Process Based on a Well-Trained Neural Network

[0071] Initialize particle swarm , calculate the fitness of each particle separately, get the initial global optimal position gbest, the current number of iterations , set the maximum number of iterations .

[0072] Enter the continuous iteration process until the number of iterations reaches , the iteration process is as follows:

[0073] 1. According to the Euclidean distance, the current state is obtained, the global optimal position is found, and the neural network is determined to be optimal;

[0074] 2. Under the current state, the action is output through the optimal neural network and converted into the corresponding parameters of the particle swarm optimization algorithm according to the linear transformation formula;

[0075] 3. For each particle, apply the obtained parameters to the particle swarm velocity and position update formula, and update the neural network corresponding to the particle.

[0076] Step S202, obtaining a supply demand allocation model.

[0077] The supply demand allocation model can be:

[0078]

[0079] is the total supply chain cost; Indicates Equipment from suppliers in Production efficiency at all times; For the Equipment from suppliers in Production cost at the moment; Respectively The production cost coefficients of the equipment of each supplier are constants; is the total production time; is the total number of equipment provided by the supplier; For the The lower limit of the production efficiency of the equipment of each supplier. It can be seen from the supply demand allocation model that the total cost of the supply chain consists of two parts, one is the production cost, and the other is the loss caused by the valve point effect in production. In the case of long-term and large-scale, the loss cannot be ignored, otherwise it will cause deviations in the calculation results.

[0080] If the supply demand allocation model is to operate normally, the constraints of the supply chain and supplier equipment can be taken into account, so that the optimal solution can meet the conditions. This supply demand allocation model mainly considers two constraints, namely supply balance constraints and supplier equipment production efficiency constraints.

[0081] (1) Supply balance constraints

[0082] At any given moment, the total number of products produced by all supplier equipment is equal to the sum of the number of products lost during transportation and the product demand.

[0083]

[0084] In the above formula, Indicated in The product demand at that moment is Equipment from suppliers in The product demand at this moment is:

[0085] Indicated in The number of products lost during transportation at any time, The loss coefficient matrix function is described as:

[0086]

[0087] (2) Supplier equipment production efficiency constraints

[0088]

[0089] In the above formula, Respectively represent The upper and lower limits of the production efficiency of the equipment of each supplier.

[0090] Step S203, after inputting the production data of the supply side into the supply demand allocation model, the particle swarm algorithm based on deep learning is used to solve it to obtain the demand allocation result of the supply side.

[0091] Since the supply-demand allocation model exhibits non-convex and non-differentiable characteristics, it is difficult to solve it efficiently using general methods. The DQNPSO algorithm can be used to solve it efficiently and obtain the demand allocation results on the supply side. In addition, the solution performance of the DQNPSO algorithm can also be verified through simulation.

[0092] In the above-mentioned solution method for supply chain demand allocation, a particle swarm algorithm based on deep learning is determined; a supply demand allocation model is obtained; after the production data of the supply side is input into the supply demand allocation model, the particle swarm algorithm based on deep learning is used to solve the problem, and the demand allocation result of the supply side is obtained, which can solve the problem more efficiently and improve the solution efficiency.

[0093] In one embodiment, after the production data of the supply side is input into the supply and demand allocation model, a particle swarm algorithm based on deep learning is used to solve it, including: corresponding the supplier equipment in the supply chain system to the particles in the particle swarm algorithm; determining the position constraint range of each particle according to the upper and lower limits of the production efficiency of the supplier equipment; after the production data of the supply side is input into the supply and demand allocation model, the supply chain demand allocation model is solved according to the position constraint range of each particle, the neural network corresponding to each particle, and the particle swarm algorithm.

[0094] Exemplarily, before solving the problem using the DQNPSO algorithm, the supplier equipment is matched with the particles, such as the number of supplier equipment is matched with the number of particles, and the production efficiency of the supplier equipment corresponds to the dimension of the particles.

[0095] After the correspondence, the DQNPSO algorithm is initialized. The initialization process includes: initializing the neural network parameters, the experience pool, and determining the position constraint range of each particle according to the upper and lower limits of the production efficiency of each supplier's equipment, thereby constraining the position of each particle so that the particle must move within the corresponding position constraint range, which can be expressed as:

[0096] In the above formula, Indicates The particle in The value of the dimension, Respectively represent The particle in The production efficiency of the dimension is the upper and lower limits; Rand is a randomly generated number in the range [0,1].

[0097] After the initialization is completed, the constraint processing strategy is determined. Specifically, in the supply demand allocation model, in order to enable the supply demand allocation model to operate normally, equality and inequality constraints can be set. When the DQNPSO algorithm is running, it is inevitable that particles will exceed the position constraint range. It needs to be processed as follows. Particles that exceed the limit will be reset to the boundary.

[0098] ; .

[0099] Among them, v ij represents the particle velocity, V min Represents the lower limit of particle velocity, V max represents the upper limit of particle speed, x ij represents the particle position, X min represents the lower bound of the feasible region of particle position, X max Represents the upper bound of the particle feasible region.

[0100] After determining the constraint processing strategy, iterative optimization can be performed. In the iterative optimization process, not only the neural network is continuously optimized and updated, but also the optimal action strategy obtained by the neural network is applied to the value of the particle swarm algorithm parameters, and the particle position and speed calculation formula is updated accordingly. After the optimization iteration, the global optimal position of the particle is used as the production efficiency of each supplier's equipment at different times, and the fitness value of the particle is the optimal target value, that is, the optimal supply chain cost of the entire supply chain demand allocation model; according to the production efficiency and optimal supply chain cost of the supplier's equipment at different times, the demand allocation result on the supply side is formed.

[0101] In one embodiment, the position constraint range of each particle is determined according to the upper and lower limits of the supplier's equipment production efficiency, including: obtaining the upper and lower limits of the supplier's equipment production efficiency; and obtaining the position constraint range of each particle according to the upper and lower limits.

[0102] In this embodiment, the upper limit and lower limit of the production efficiency of the supplier's equipment can be obtained, and the upper limit of the production efficiency is used as the upper limit of the position of the particles corresponding to the supplier's equipment, and the lower limit of the production efficiency is used as the lower limit of the position of the particles corresponding to the supplier's equipment. According to the upper limit and lower limit of the position of the particle, a position constraint range of the particle is formed.

[0103] In one embodiment, the supplier devices in the supply chain system are mapped to particles in the particle swarm algorithm, including: using the number of supplier devices in the supply chain system as the number of particles; and mapping each supplier device to a particle.

[0104] For example, the supply chain system includes 5 supplier devices, the number of particles can be set to 5, and each supplier device corresponds to a particle, which can be expressed as: .

[0105] In one embodiment, when the deep learning is deep Q learning, the obtained neural network is a deep Q network.

[0106] In order to intuitively compare the performance of DQNPSO and PSO algorithms in the supply and demand allocation problem, the number of population particles of the two algorithms is , maximum number of iterations ; In PSO algorithm ; No parameter setting is required in the DQNPSO algorithm, and the inertia weight search range is ; The search range of the acceleration factor is .

[0107] The characteristics of each supplier's equipment and the hourly demand of the supply demand allocation model are shown in Table 2 and Table 3 respectively:

[0108]

[0109] Table 2

[0110]

[0111] Table 3

[0112] The transmission loss coefficient matrix is ​​shown below:

[0113]

[0114] The data and transmission loss coefficient matrix are substituted into the supply demand allocation model, and the DQNPSO algorithm is used to simulate using simulation tools. Table 4 lists the production efficiency and cost of each supplier's equipment in the two algorithms.

[0115]

[0116] Table 4

[0117] The DQNPSO algorithm converges very quickly after 50 iterations, while the PSO algorithm converges after 90 iterations. From the optimization effect point of view, the cost of the DQNPSO algorithm is lower. The DQNPSO algorithm is superior to the PSO algorithm in terms of convergence speed and optimization degree, which greatly saves resource consumption.

[0118] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0119] Based on the same inventive concept, the embodiment of the present application also provides a supply chain demand allocation solution device for implementing the supply chain demand allocation solution method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more supply chain demand allocation solution device embodiments provided below can refer to the limitations of the supply chain demand allocation solution method above, and will not be repeated here.

[0120] In one embodiment, Figure 4 As shown, a solution device for supply chain demand allocation is provided, comprising:

[0121] An algorithm determination module 401 is used to determine a particle swarm algorithm based on deep learning;

[0122] Model acquisition module 402, used to acquire a supply demand allocation model;

[0123] The model solving module 403 is used to input the production data of the supply side into the supply demand allocation model, and then solve it using a particle swarm algorithm based on deep learning to obtain a demand allocation result of the supply side.

[0124] In one embodiment, the model solving module 403 is also used to: correspond supplier equipment in the supply chain system to particles in the particle swarm algorithm; determine the position constraint range of each particle according to the upper and lower limits of the supplier equipment production efficiency; after inputting the production data of the supply side into the supply demand allocation model, solve the supply chain demand allocation model according to the position constraint range of each particle, the neural network corresponding to each particle, and the particle swarm algorithm.

[0125] In one embodiment, the model solving module 403 is further used to: obtain the upper limit value and the lower limit value of the production efficiency of the supplier's equipment; and obtain the position constraint range of each particle according to the upper limit value and the lower limit value.

[0126] In one embodiment, the model solving module 403 is further used to: use the number of supplier devices in the supply chain system as the number of particles; and correspond each supplier device to a particle.

[0127] In one embodiment, the neural network is a deep Q-network.

[0128] Each module in the above-mentioned supply chain demand allocation solution device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0129] In an exemplary embodiment, a computer device is provided, the internal structure of which can be shown as follows: Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data involved in the above method. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for solving supply chain demand allocation is implemented.

[0130] Those skilled in the art will understand that Figure 5The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0131] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above-mentioned various method embodiments when executing the computer program.

[0132] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0133] In one embodiment, a computer program product is provided, on which a computer program is stored, and the computer program is used by a processor to execute the steps in the above-mentioned various method embodiments.

[0134] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0135] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0136] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0137] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for solving supply chain demand allocation, characterized in that: The method comprises: Determine the particle swarm algorithm based on deep learning; Obtain supply demand allocation model; After the production data on the supply side is input into the supply and demand allocation model, it is solved using a particle swarm algorithm based on deep learning to obtain the demand allocation result on the supply side.

2. The method according to claim 1, characterized in that After the production data of the supply side is input into the supply demand allocation model, the particle swarm algorithm based on deep learning is used to solve it, including: Correspond the supplier equipment in the supply chain system to the particles in the particle swarm algorithm; Determine the position constraint range of each particle based on the upper and lower limits of the supplier's equipment production efficiency; After the production data of the supply side is input into the supply demand allocation model, the supply chain demand allocation model is solved according to the position constraint range of each particle, the neural network corresponding to each particle and the particle swarm algorithm.

3. The method according to claim 2, characterized in that According to the upper and lower limits of the supplier's equipment production efficiency, determine the position constraint range of each particle, including: Obtain the upper and lower limits of the supplier's equipment production efficiency; According to the upper limit value and the lower limit value, the position constraint range of each particle is obtained.

4. The method according to claim 2, characterized in that: The supplier equipment in the supply chain system is mapped to the particles in the particle swarm algorithm, including: The number of supplier devices in the supply chain system is taken as the number of particles; Each vendor device is mapped to a particle.

5. The method according to claim 2, characterized in that: The neural network is a deep Q network.

6. A supply chain demand allocation solution device, characterized in that: The device comprises: An algorithm determination module, used to determine a particle swarm algorithm based on deep learning; A model acquisition module is used to acquire a supply and demand allocation model; The model solving module is used to input the production data of the supply side into the supply and demand allocation model, and then solve it using the particle swarm algorithm based on deep learning to obtain the demand allocation result of the supply side.

7. The device according to claim 6, characterized in that Model solving module, also used for: Correspond the supplier equipment in the supply chain system to the particles in the particle swarm algorithm; Determine the position constraint range of each particle based on the upper and lower limits of the supplier's equipment production efficiency; After the production data of the supply side is input into the supply demand allocation model, the supply chain demand allocation model is solved according to the position constraint range of each particle, the neural network corresponding to each particle and the particle swarm algorithm.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

10. 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 5 is implemented.