Material ordering and replenishment decision-making method and service platform based on GA-PSO

Abstract

The invention relates to a GA-PSO-based material ordering and replenishment decision-making method and a service platform. The method comprises the steps of obtaining material ordering amounts corresponding to various target materials in each ordering period; coding the number of delivery cycles in each ordering cycle and the material ordering quantity corresponding to the multiple target materials to generate multiple first matrix type coding individuals; after carrying out at least one genetic operation by utilizing a genetic algorithm based on the plurality of first matrix type coding individuals, determining the fitness of a second sub-coding body corresponding to each second matrix type coding individual generated by current iteration, and carrying out genetic operation iteration based on the fitness and the second matrix type coding individuals to obtain a second sub-coding body corresponding to each second matrix type coding individual; until a plurality of candidate matrix type coding individuals are generated; after candidate sub-coding bodies in the candidate matrix type coding individuals serve as candidate particles, iteration processing is conducted on the multiple candidate particles on the basis of a particle swarm optimization algorithm, and target particles including target material replenishment decision information are selected from the multiple generated candidate particles.

Description

Technical Field

[0001] The present application relates to the field of intelligent manufacturing technology, and in particular to a decision-making method and service platform for material ordering and replenishment based on GA-PSO. Background Art

[0002] In the relevant technologies, coatings manufacturers not only need to continuously innovate products and improve service quality, but also must effectively control material ordering costs, thus making the optimization of procurement management crucial.

[0003] In related technologies, material procurement involves multiple complex decision points and constraints. In the bulk ordering stage, manufacturers need to decide the order quantity of each material and which supplier to choose. This requires not only considering the price and demand of the material, but also weighing the price discounts offered by the supplier. At the same time, in the batch delivery stage, manufacturers also need to accurately plan the delivery quantity and distribution vehicles of each material based on multiple factors such as the order quantity of the current cycle, vehicle type, warehouse capacity, etc., to ensure the timeliness and efficiency of supply.

[0004] In the related technology, the existing optimization algorithms for material inventory replenishment decisions have insufficient solution speed and cannot meet the needs of rapid decision-making when dealing with such complex problems. At the same time, since the problem of material inventory replenishment decisions involves a large amount of data and calculations, traditional optimization algorithms often have defects such as low computational efficiency and poor solution quality.

[0005] At present, no effective solution has been proposed for the problems in related technologies of material inventory replenishment decision-making, such as the inability to meet the demand for quick decision-making, low decision-making efficiency and poor material inventory replenishment effect. Summary of the invention

[0006] The embodiment of the present application provides a material ordering and replenishment decision-making method and service platform based on GA-PSO, so as to at least solve the problems that the related technology cannot meet the demand for rapid decision-making, has low decision-making efficiency and poor material inventory replenishment effect for material inventory replenishment decision-making.

[0007] In a first aspect, an embodiment of the present application provides a GA-PSO-based material ordering and replenishment decision-making method, comprising: determining, in a plurality of preset ordering cycles, a plurality of delivery cycles agreed upon in each of the ordering cycles, and obtaining the material ordering quantities corresponding to the plurality of target materials to be replenished in each of the ordering cycles, wherein the material ordering quantities are used to characterize the total amount of materials required to be delivered for each target material in the corresponding plurality of delivery cycles; encoding the number of the delivery cycles and the material ordering quantities corresponding to the plurality of target materials, to generate a plurality of first matrix-type coding individuals, wherein the first matrix-type coding individuals include a plurality of first matrix-type sub-coding bodies having a plurality of first sub-coding bodies, and the first sub-coding bodies are used to characterize the material ordering quantities of a target material in the plurality of delivery cycles corresponding to one of the ordering cycles. The material replenishment decision information includes a replenishment supply object and a replenishment supply quantity; after performing at least one genetic operation based on the plurality of the first matrix-type coding individuals using a preset genetic algorithm GA, determining the fitness of the second sub-coding body corresponding to each second matrix-type coding individual generated in the iteration, and performing genetic operation iterations based on the fitness and the second matrix-type coding individuals until a plurality of candidate matrix-type coding individuals are generated, wherein the genetic evolution operation includes: individual selection based on a roulette wheel method and an elite retention strategy, adaptive crossover based on gene position, and dynamic mutation, and the fitness is determined according to the ordering cost, transportation cost, storage cost, and emergency transfer cost generated by the replenishment supply object corresponding to the second sub-coding body in replenishing the target material of the replenishment supply quantity; After taking the candidate sub-coding bodies in the candidate matrix-type coding individuals as candidate particles, based on the preset particle swarm optimization algorithm PSO, a particle swarm search iterative process is performed on the multiple candidate particles, and a target particle is selected from the generated multiple candidate particles to obtain a decision result, wherein the decision result includes the target material replenishment decision information corresponding to the target particle, and the particle swarm optimization algorithm PSO adopts a dynamic adaptive inertia factor and a dynamically adjusted acceleration coefficient.

[0008] In a second aspect, an embodiment of the present application provides a service platform, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the material ordering and replenishment decision-making method based on GA-PSO as described in the first aspect are implemented.

[0009] Compared with the related art, the GA-PSO-based material ordering and replenishment decision-making method and service platform provided in the embodiment of the present application adopts, in a plurality of preset ordering cycles, to determine a plurality of delivery cycles agreed upon in each of the ordering cycles, and obtain the material ordering quantities corresponding to the plurality of target materials to be replenished in each of the ordering cycles, wherein the material ordering quantities are used to characterize the total amount of materials required to be delivered for each target material in the corresponding plurality of delivery cycles; the number of the delivery cycles and the material ordering quantities corresponding to the plurality of target materials are encoded to generate a plurality of first matrix-type encoded individuals, wherein the first matrix-type encoded individuals include a plurality of A first matrix-type sub-encoding body having a plurality of first sub-encoding bodies, wherein the first sub-encoding body is used to represent material replenishment decision information of a target material in a plurality of the delivery cycles corresponding to a said ordering cycle, wherein the material replenishment decision information includes a replenishment supply object and a replenishment supply quantity; after performing at least one genetic operation using a preset genetic algorithm GA based on the plurality of the first matrix-type encoding individuals, determining the fitness of the second sub-encoding body corresponding to each second matrix-type encoding individual generated in the current iteration, and performing genetic operation iterations based on the fitness and the second matrix-type encoding individuals until a plurality of candidate matrix-type encoding individuals are generated; After taking the candidate sub-coding bodies in the candidate matrix-type coding individuals as candidate particles, based on the preset particle swarm optimization algorithm PSO, a particle swarm search iterative process is performed on the multiple candidate particles, and a target particle is selected from the generated multiple candidate particles to obtain a decision result, wherein the decision result includes the target material replenishment decision information corresponding to the target particle, and solves the problem that the material inventory replenishment decision cannot meet the demand for fast decision-making, the decision efficiency is low, and the material inventory replenishment effect is not good. By combining the global search capability of the genetic algorithm and the group learning characteristics of the particle swarm optimization algorithm PSO, a more effective exploration of the solution space is achieved, and a solution closer to the global optimal solution is found, thereby improving the beneficial effect of the solution efficiency.

[0010] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 It is a hardware structure block diagram of a terminal of a GA-PSO-based material ordering and replenishment decision-making method according to an embodiment of the present application; Figure 2 is a flow chart of a material ordering and replenishment decision-making method based on GA-PSO according to an embodiment of the present application; Figure 3 is a schematic diagram of a matrix-type coding individual provided in an embodiment of the present application; Figure 4 is a schematic diagram of a genetic operation based on position crossover according to an embodiment of the present application; Figure 5 is a schematic diagram of a randomly changing dynamic mutation operation performed in an embodiment of the present application; Figure 6 is a schematic diagram of a dynamic variation operation of exchanging adjacent cycles in an embodiment of the present application; Figure 7 is a schematic diagram of a dynamic variation operation of random sharing in an embodiment of the present application; Figure 8 is a schematic diagram of a dynamic mutation operation of random merging performed in an embodiment of the present application; Fig. 9 is a schematic diagram of particle search based on a particle swarm optimization algorithm in an embodiment of the present application; Fig.10 A structural block diagram of a material ordering and replenishment decision-making device based on GA-PSO according to an embodiment of the present application. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated 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 intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application. In addition, it can also be understood that although the efforts made in this development process may be complex and lengthy, for ordinary technicians in the field related to the contents disclosed in the present application, some changes such as design, manufacturing or production based on the technical contents disclosed in the present application are only conventional technical means, and should not be understood as insufficient contents disclosed in the present application.

[0013] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0014] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantitative limitation and may represent the singular or plural. The terms "including", "comprising", "having" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include unlisted steps or units, or may also include other steps or units inherent to these processes, methods, products or devices. The "multiple links" involved in this application refer to links greater than or equal to two. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" may represent: A exists alone, A and B exist at the same time, and B exists alone. The terms "first", "second", "third" and the like involved in this application are only used to distinguish similar objects and do not represent a specific ordering of objects.

[0015] The method embodiment provided in this embodiment can be executed in a terminal, a computer or a similar computing device. Taking running on a terminal as an example, Figure 1 : is a hardware structure block diagram of a terminal of a GA-PSO-based material ordering and replenishment decision method according to an embodiment of the present application. Figure 1 As shown, the terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. Optionally, the terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is for illustration only and does not limit the structure of the above terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0016] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the decision-making method for material ordering and replenishment based on GA-PSO in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0017] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by the communication provider of the terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0018] This embodiment provides a material ordering and replenishment decision-making method based on GA-PSO running on the above terminal. Figure 2 is a flow chart of a material ordering and replenishment decision-making method based on GA-PSO according to an embodiment of the present application, such as Figure 2 As shown, the process includes the following steps:

[0019] Step S201, in the preset multiple ordering cycles, determine the multiple delivery cycles agreed upon in each ordering cycle, and obtain the material order quantities corresponding to the multiple target materials to be replenished in each ordering cycle, wherein the material order quantity is used to represent the total amount of materials that need to be delivered for each target material in the corresponding multiple delivery cycles.

[0020] The execution subject of the embodiment of the present application corresponds to a manufacturer (for example, a manufacturer in the coatings industry) that processes materials for inventory replenishment, and the material inventory replenishment decision method described in the embodiment of the present application corresponds to the process of batch ordering and batch delivery of the manufacturer. It should be understood that there are multiple ordering cycles during the decision-making period of the embodiment of the present application, and each ordering cycle includes multiple delivery cycles, and the allocated delivery refers to the supplier delivering the corresponding target materials in each negotiated delivery cycle; at the same time, it should also be understood that the manufacturer determines the material ordering quantity in the ordering cycle, that is, the decision point of each ordering cycle is the ordering quantity of each target material and the selected supplier, which needs to comprehensively consider the material inventory, material price, material demand, and supplier price discount; each ordering cycle contains multiple delivery cycles, and the manufacturer determines the delivery quantity of the target material that the supplier needs to deliver in the delivery cycle, that is, the decision point of each delivery cycle is the supplier's delivery quantity of each target material and the selected current cycle order quantity, vehicle type, supplier distance, warehouse capacity, storage cost, and out-of-stock cost; in this embodiment, the target materials include but are not limited to emulsions, resins, and titanium dioxide required to generate coatings.

[0021] In this embodiment, the manufacturer is expected to deliver multiple target materials in each ordering cycle according to each delivery cycle. The minimum data unit of the result of the decision in this embodiment of the application is the quantity of each target material delivered in each delivery cycle and who (supplier) delivers it, and the decision result includes the delivery data of the target materials in all delivery cycles of all ordering cycles within the preset decision period (including the replenishment supply object and the replenishment supply quantity).

[0022] Step S202, encode the number of delivery cycles and the material order quantities corresponding to the multiple target materials, and generate multiple first matrix-type coding individuals, wherein the first matrix-type coding individuals include multiple first matrix-type sub-coding bodies with multiple first sub-coding bodies, and the first sub-coding bodies are used to represent material replenishment decision information of a target material in multiple delivery cycles corresponding to one ordering cycle, and the material replenishment decision information includes replenishment supply objects and replenishment supply quantities.

[0023] In this embodiment, after determining the delivery cycle and material order quantity corresponding to an order cycle, the material order quantity is allocated to each delivery cycle in a manner that meets the production demand (material replenishment demand determined based on the initial inventory demand). Figure 3, 170 units of resin are allocated to delivery cycle 1, delivery cycle 2 and delivery cycle 3, and all 170 units of resin are allocated to delivery cycle 1, and the replenishment supply of delivery cycle 2 and delivery cycle 3 is 0, thereby completing the initialization of one dimension (replenishment supply, delivery quantity) in the population initialization process. Then, an exhaustive method is used to select the target material with the smallest cost of supplying the corresponding replenishment supply from multiple alternative replenishment supply objects as the corresponding replenishment supply object, thereby associating the corresponding replenishment supply object (supplier), and then realizing the encoding generation of the first matrix-type encoding individual; reference Figure 3 In this embodiment, a first matrix-type sub-code body represents the delivery status of each target material in all delivery cycles (e.g., delivery cycle 1 to delivery cycle 3) within an ordering cycle (e.g., ordering cycle 1); at the same time, a first sub-code body represents the delivery status of a target material within an ordering cycle. For example, taking the target material titanium dioxide as an example, a first sub-code body corresponds to: delivery cycle 1: 0 units of titanium dioxide, delivery cycle 2: 200 units of titanium dioxide, delivery cycle 3: 170 units of titanium dioxide, and at the same time, the corresponding replenishment supply quantity will be associated with the corresponding replenishment supply object ( Figure 3 not shown).

[0024] Step S203, after performing at least one genetic operation using a preset genetic algorithm GA based on multiple first matrix-type coding individuals, determine the fitness of the second sub-coding body corresponding to each second matrix-type coding individual generated in the current iteration, and perform genetic operation iterations based on the fitness and the second matrix-type coding individuals until multiple candidate matrix-type coding individuals are generated, wherein the genetic evolution operation includes: individual selection based on the roulette method and the elite retention strategy, adaptive crossover based on gene position, and dynamic mutation, and the fitness is determined based on the ordering cost, transportation cost, storage cost, and emergency transfer cost generated by the replenishment supply object corresponding to the second sub-coding body in replenishing the target material of the replenishment supply quantity.

[0025] In this embodiment, the fitness of each corresponding character coding body is calculated, and then the fitness of multiple target materials in the same ordering cycle is summed up, the individual fitness corresponding to the first matrix-type sub-coding individual of the country is calculated, and then the individual fitness in multiple ordering cycles is summed up to obtain the total fitness corresponding to a corresponding matrix-type coding individual.

[0026] It is understandable that performing genetic operations is a means understood and known by those skilled in the art, but the present application adopts genetic operations that are different from existing genetic operations, such as individual selection based on roulette method and elite retention strategy, adaptive crossover based on gene position, and dynamic mutation. This part of genetic operations maintains the validity and diversity of the solution, while accelerating the convergence process of the algorithm.

[0027] Step S204, after taking the candidate sub-coding bodies in the candidate matrix-type coding individuals as candidate particles, based on the preset particle swarm optimization algorithm PSO, a particle swarm search iterative process is performed on multiple candidate particles, and a target particle is selected from the generated multiple candidate particles to obtain a decision result, wherein the decision result includes the target material replenishment decision information corresponding to the target particle, and the particle swarm optimization algorithm PSO adopts a dynamic adaptive inertia factor and a dynamically adjusted acceleration coefficient.

[0028] In this embodiment, the implementation method of the particle swarm optimization algorithm is understandable and known to those skilled in the art, but the present application adopts a processing method that is different from the conventional particle swarm optimization algorithm, that is, a dynamic adaptive inertia factor and a dynamically adjusted acceleration coefficient are used. The dynamic adaptive inertia factor and the dynamically adjusted acceleration coefficient are configured and adjusted during the search process so that they can adaptively change according to the fitness value of the chromosome and the number of iterations. Compared with the fixed inertia factor, the dynamic adaptive inertia factor can better balance the global and local search capabilities. By dynamically adjusting the acceleration factor, the diversity of the particle swarm in the early search stage and the convergence of the algorithm in the later search stage are achieved.

[0029] Through the above steps S201 to S204, multiple delivery cycles agreed upon in each ordering cycle are determined in the preset multiple ordering cycles, and the material ordering quantities corresponding to the multiple target materials to be replenished in each ordering cycle are obtained, and the material ordering quantities are used to represent the total amount of materials required to be delivered for each target material in the corresponding multiple delivery cycles; the number of delivery cycles and the material ordering quantities corresponding to the multiple target materials are encoded to generate multiple first matrix-type encoding individuals, the first matrix-type encoding individuals include multiple first matrix-type sub-encoding bodies with multiple first sub-encoding bodies, the first sub-encoding bodies are used to represent the material replenishment decision information of a target material in multiple delivery cycles corresponding to an ordering cycle, and the material replenishment decision information includes the replenishment supply object and the replenishment supply quantity; after performing at least one genetic operation using the preset genetic algorithm GA based on the multiple first matrix-type encoding individuals, it is determined that The fitness of the second sub-coding body corresponding to each second matrix-type coding individual generated by the iteration is calculated, and based on the fitness and the second matrix-type coding individual, a genetic operation iteration is performed until a plurality of candidate matrix-type coding individuals are generated; after the candidate sub-coding bodies in the candidate matrix-type coding individuals are taken as candidate particles, a particle swarm search iteration process is performed on the plurality of candidate particles based on a preset particle swarm optimization algorithm PSO, and a target particle is selected from the plurality of candidate particles generated to obtain a decision result, which includes the target material replenishment decision information corresponding to the target particle, and solves the problem that the material inventory replenishment decision cannot meet the demand for fast decision-making, the decision efficiency is low, and the material inventory replenishment effect is not good. By combining the global search capability of the genetic algorithm and the group learning characteristics of the particle swarm optimization algorithm PSO, a more effective exploration of the solution space is achieved, a solution closer to the global optimal solution is found, and the beneficial effect of improving the solution efficiency is achieved.

[0030] It should be noted that the decision-making method of the embodiment of the present application can obtain the optimal solution for the "bulk ordering-batch delivery" process in the paint manufacturing enterprise and achieve rapid convergence; the decision-making method of the embodiment of the present application is highly flexible and adaptable, and the parameters in the GA-PSO algorithm (such as crossover probability, mutation probability, inertia weight, etc.) are all adaptively adjusted and dynamically adjusted to adjust according to the complexity and scale of the specific problem, so that the algorithm can remain efficient and stable in various scenarios; the embodiment of the present application optimizes and adjusts key steps such as encoding method, chromosome crossover and mutation; in terms of encoding method, matrix encoding is adopted to more effectively represent the decision variables of inventory replenishment. In terms of chromosome crossover and mutation, individual selection based on roulette method and elite retention strategy, adaptive crossover based on gene position, and dynamic mutation are adopted to maintain the effectiveness and diversity of the solution, while accelerating the convergence process of the algorithm.

[0031] It should be further explained that this method aims to solve the problems of low solution efficiency and low convergence quality faced by traditional metaheuristic algorithms when dealing with complex operations research models, especially the dilemma of difficulty in achieving the optimal solution for the "bulk ordering-batch delivery" process optimization in paint manufacturing companies. By combining the global search capability of GA and the group learning characteristics of PSO, the method of the embodiment of the present invention can more effectively explore the solution space and find a solution that is closer to the global optimum. This method not only improves the solution efficiency, but also significantly improves the convergence quality.

[0032] In some embodiments, the number of delivery cycles and the material order quantities corresponding to the various target materials are encoded to generate a plurality of first matrix-type encoding individuals, which is implemented by the following steps:

[0033] Step 21, determine the target information of multiple alternative supply objects, and randomly distribute the material order quantity corresponding to each target material within multiple delivery cycles of an ordering cycle to obtain the first delivery quantity of each target material within a delivery cycle, wherein the target information includes material discount rate and transportation cost parameters.

[0034] In this embodiment, after determining the delivery cycle and material order quantity corresponding to an order cycle, the material order quantity is allocated to each delivery cycle in a manner that meets the production demand (material replenishment demand determined based on the initial inventory demand). Figure 3 , distribute 170 units of resin to delivery cycle 1, delivery cycle 2 and delivery cycle 3, and distribute all 170 units of resin to delivery cycle 1, and the replenishment supply of delivery cycle 2 and delivery cycle 3 is 0, so as to obtain the first delivery quantity of resin in each delivery cycle

[0035] Step 22, based on the material discount rate and the transportation cost parameters, using the exhaustive method, determine the alternative supply object for delivering the target material of the first delivery quantity from multiple alternative supply objects, use the determined alternative supply object and the first delivery quantity as the replenishment supply object and the replenishment supply quantity respectively, and generate a first sub-code body corresponding to the corresponding target material, wherein the replenishment supply object is the alternative supply object with the smallest first cost generated by supplying the target material of the first delivery quantity, and the first cost is the sum of the ordering cost and the transportation cost determined respectively according to the material discount rate and the transportation cost parameters.

[0036] In this embodiment, the optimal supplier exhaustive method is used to determine the replenishment supply object, which specifically includes the following steps:

[0037] Step 1: Calculate the ordering cost based on the gene value of the first sub-coding body, for example: Figure 3For emulsion, first calculate the sum of the delivery quantities of the three delivery cycles under the ordering cycle Q1, that is, 150+200=350, and then calculate the ordering cost of each alternative supplier s based on the discount rate of each alternative supplier s corresponding to this delivery quantity Then, the model is determined according to the delivery volume of each delivery cycle, and the corresponding transportation cost of each candidate supply object is calculated according to the distance of each candidate supply object. , add up the two costs, and select the alternative supply object with the smallest total cost as the option.

[0038] Step 2, repeat step 1, and calculate all the purchase cycles Q of the emulsion. Similarly, calculate the ordering cost of target raw materials such as resin, and determine the alternative suppliers for different target materials with different ordering cycles. Finally, add up all the costs separately to get the ordering cost and .

[0039] Step 3: Storage cost calculation

[0040] Since materials have lead times, the target materials delivered in the previous delivery cycle can only arrive in the next cycle. Therefore, the end-of-period holdings = beginning-of-period inventory + the order quantity of the previous delivery cycle - the demand quantity. For example, the end-of-period holdings of the previous production cycle of the resin are the beginning-of-period holdings of 50 minus the production demand of 10, and the end-of-period holdings of its production cycle 2 are the beginning-of-period holdings of 40 plus the order quantity of one delivery cycle of 170 minus the production demand of 200, which equals the end-of-period holdings of 10. Similarly, the end-of-period holdings of all materials in all production cycles are calculated, and the storage cost is calculated based on the end-of-period holdings. In addition, when the end-of-period holdings are greater than the maximum storage capacity of the warehouse, it is necessary to calculate the warehouse rental cost generated by the additional storage of materials. Finally, the storage cost is obtained.

[0041] Step 4: When the beginning inventory plus the order quantity for one delivery cycle is less than the production demand, an emergency transfer will occur, and the cost of the emergency transfer will be determined based on the shortage quantity of the material.

[0042] Step 23, encode all first sub-coding bodies corresponding to multiple target materials in the same ordering cycle into a first matrix-type sub-coding body, and encode the first matrix-type sub-coding bodies corresponding to multiple ordering cycles into a first matrix-type coding individual, and generate multiple first matrix-type coding individuals through population initialization.

[0043] In this embodiment, a plurality of first matrix-type coded individuals form an initial population, each of which represents a potential material inventory replenishment decision strategy. Figure 3As shown, the values ​​in the table represent the delivery quantity of each target raw material in each delivery cycle, and the material order quantity in one ordering cycle will be equal to the sum of the delivery quantities of each corresponding delivery cycle; Figure 3 The delivery quantity of the emulsion in delivery cycle 1 is 0, the delivery quantity in delivery cycle 2 is 150, and the delivery quantity in delivery cycle 3 is 200. Therefore, the delivery quantity in ordering cycle 1 is the sum of the three, that is, 350. Similarly, the delivery quantities of the resin in delivery cycles 1, 2, and 3 are 170, 0, and 0 respectively, and the delivery quantity in ordering cycle 1 is 170.

[0044] Through the above steps 21 to 23, the initial population is initialized.

[0045] In some of the embodiments, after generating a plurality of first matrix-type coding individuals, the following steps are performed:

[0046] Step 31, determine the current matrix coding individual for individual selection based on the roulette method and the elite retention strategy, and based on the fitness of all current sub-coding bodies corresponding to each current matrix coding individual, calculate the total fitness corresponding to each current matrix coding individual according to a preset objective function, wherein the current matrix coding individual includes one of the following: a first matrix coding individual, a second matrix coding individual.

[0047] In this embodiment, the preset objective function is: Min(PC+SC+TC+EC), and the calculated total fitness is: F=Min(PC+SC+TC+EC), where: Among them, PC represents the total material ordering cost corresponding to a corresponding matrix-coded individual; SC represents the total target material storage cost corresponding to a corresponding matrix-coded individual; TC represents the total material transportation cost corresponding to a corresponding matrix-coded individual; EC represents the total material emergency replenishment cost corresponding to a corresponding matrix-coded individual; i represents the i-th target material, I represents the set of target materials, i∈I; j represents the j-th replenishment supply object, J represents the set of replenishment supply objects, j∈J; q represents the q-th ordering cycle, Q represents the set of ordering cycles, q∈Q; k represents the k-th discount plan of the target material, K represents the set of discount plans, k∈K; m represents the m-th delivery cycle, M represents the set of delivery cycles, m∈M; v represents the v-th vehicle, V represents the set of vehicles, v∈V; is the discount rate corresponding to the kth discount scheme k for the i-th target material by the j-th replenishment supply object in the q-th ordering period; is the base price of the i-th target material for the j-th replenishment supply object in the q-th ordering cycle; It is Unit storage cost of the target material; It is Order cycle Delivery cycle The actual storage quantity of the target material; is the unit rental warehouse storage cost of the i-th target material; It is Order cycle Delivery cycle Storage capacity of rented warehouses for various materials; Replenishment supply object The corresponding distance cost; is the rental cost corresponding to the vth vehicle; It is Unit emergency transfer cost of the target material; It is Order cycle Delivery cycle The out-of-stock quantity of the target material; is a decision variable, indicating whether the vth vehicle is selected when the i-th target material in the m-th delivery cycle under the q-th ordering cycle is delivered to the j-th replenishment supply object; is a decision variable, which indicates the quantity of the i-th target material selected to be ordered from the j-th replenishment supply object in the q-th ordering cycle; is a decision variable, indicating whether the i-th target material in the m-th delivery cycle under the q-th ordering cycle chooses to be replenished from the j-th replenishment supply object; is a decision variable, indicating whether the vth vehicle is selected when replenishing the i-th target material in the m-th delivery cycle under the q-th ordering cycle from the j-th replenishment supply object.

[0048] Step 32, based on the total fitness, select the current matrix-type coding individuals whose total fitness is less than the preset fitness threshold from multiple current matrix-type coding individuals to obtain the first candidate coding individuals, wherein the first candidate coding individuals are used to represent the best individuals in the current group.

[0049] In this embodiment, the roulette method can retain excellent individuals to a certain extent, but in complex optimization problems, the best individual of the current group may still be lost in the next generation, resulting in the genetic algorithm being unable to converge to the global optimal solution. In this embodiment, an elite retention strategy is introduced, that is, the best first candidate coding individuals that appear in the evolution process of multiple current matrix coding individuals are directly copied to the next generation without genetic operations, that is, the remaining offspring will be directly retained from the best individuals in the previous generation, thereby ensuring that the algorithm can continue to approach the global optimal solution.

[0050] Step 33, after calculating the first selection probability based on the transposed fitness generated by transposing each total fitness, use the first selection probability to perform individual selection based on the roulette method on all current matrix-type coding individuals except the first candidate coding individual to obtain a preset number of second candidate coding individuals, and use all first candidate coding individuals and all second candidate coding individuals as the first intention coding individuals corresponding to the selection result of completing the individual selection based on the roulette method and the elite retention strategy.

[0051] In this embodiment, the total fitness is transposed according to the following formula: Where gen represents the number of iterations, n represents the code of the nth matrix-coded individual in the corresponding population, N represents the set of individual codes, n∈N, represents the nth current matrix-coded individual, It indicates the transposed fitness of the nth current matrix-coded individual after gen iterations. It represents the total fitness of the nth current matrix-coded individual after completing gen iterations.

[0052] In this embodiment, the first selection probability is calculated according to the following formula: in, Represents the nth current matrix-coded individual in the population after completing gen iterations The probability of the first choice being selected.

[0053] In this embodiment, selection is continuously performed according to the first selection probability until the number of second candidate code individuals is filled to at least 95% of individuals, that is, a preset number of second candidate code individuals is obtained.

[0054] Through the above steps 31 to 33, the genetic operation of individual selection based on the roulette method and the elite retention strategy is realized, so that the algorithm can continuously approach the global optimal solution.

[0055] In some of the embodiments, after generating the first intention coding individual, the following steps are further implemented:

[0056] Step 41, select the maximum total fitness, the minimum total fitness and the average total fitness from the total fitness corresponding to all first intention coding individuals, and calculate the current crossover probability corresponding to each first intention coding individual in the current iteration based on the total fitness corresponding to each first intention coding individual and the maximum total fitness, the minimum total fitness and the average total fitness.

[0057] In this embodiment, the current crossover probability corresponding to each first intention coding individual in the current iteration is adaptively updated according to the following formula: in, It is The current crossover probability of the nth first intention coded individual in the iterated population, , are the minimum and maximum crossover probabilities, respectively, It is The total fitness of the i-th first intention encoding individual in the iterated population, , , They are the maximum total fitness, the minimum total fitness, and the average total fitness in the population respectively; it can be understood that, in this embodiment, the influence of the evolutionary algebra and the individual fitness function value is considered to maintain the diversity of the population and alleviate the exhaustive search process in GA-PSO; at the same time, an adaptive update mechanism is introduced to promote the regeneration of the population. In the context of the adaptive crossover operation, the crossover probability is adaptively reduced during the iteration process to increase the possibility of retaining high-quality solutions.

[0058] In step 42, a preset number of intentional sub-coding bodies are randomly selected from all the intentional sub-coding bodies corresponding to the first intentional coding individual as the first cross-sub-coding body group corresponding to each first intentional coding individual, and the intentional sub-coding bodies other than the first cross-sub-coding body group are used as the second cross-sub-coding body group corresponding to the first intentional coding body.

[0059] Step 43, after selecting two first intended coding individuals from all first intended coding individuals based on the current crossover probability to obtain the first parent coding individual and the second parent coding individual, the first crossover sub-coding body group corresponding to the first parent coding individual and the second crossover sub-coding body group corresponding to the second parent coding individual are filled into the first child coding individual according to the corresponding coding positions in the first parent coding individual and the second parent coding individual, and the first crossover sub-coding body group corresponding to the second parent coding individual and the second crossover sub-coding body group corresponding to the first parent coding individual are filled into the second child coding individual according to the corresponding coding positions in the first parent coding individual and the second parent coding individual.

[0060] Step 44, taking the first offspring coding individual and the second offspring coding individual generated each time as the second intended coding individual, and taking the corresponding second intended coding individual as the result of completing the genetic operation of adaptive crossover based on gene position at that time.

[0061] In this embodiment, the crossover operation process based on position crossover (Order-Based Crossover, referred to as OBX) is adopted. Figure 4 , in the two parent chromosomes (the two first intention coding individuals, reference Figure 4 Randomly select several positions (corresponding to the first crossover sub-coding body group) from parent 1 and parent 2 in the parent 2. The positions can be discontinuous. First, find the position of the selected gene in parent 1 in parent 2, and then use the remaining genes in parent 2 to generate offspring, and ensure that the positions correspond. Put the selected genes in parent 1 into the remaining positions of the offspring in order. Another offspring is obtained in a similar way. Specifically, refer to Figure 4 For parent 1, it randomly selects genes at positions 1, 2, and 6. Naturally, parent 2 selects genes at positions 3, 4, and 5. For the first offspring (reference Figure 4 The offspring 1 in the parent generation 1 finds the position of the selected gene (1, 2, 6) in the parent generation 1, and finds the position of the selected gene (3, 4, 5) in the parent generation 2, and fills in the selected genes accordingly, thus generating the corresponding offspring 1; for the second offspring (reference Figure 4 Offspring 2 in the parent generation is the opposite of offspring 1. The positions of the selected genes (1, 2, 6) are found in parent 2, and the positions of the selected genes (3, 4, 5) are found in parent 1. The selected genes are filled in accordingly to generate the corresponding offspring 2.

[0062] Through the above steps 41 to 44, the genetic operation of adaptive crossover based on gene position is realized.

[0063] In some embodiments, after generating the second intention coding individual, the following steps are further implemented:

[0064] Step 51, according to a preset dynamic adjustment method, determine the current mutation probability corresponding to the current mutation operation.

[0065] Step 52, based on the current mutation probability, select the second intention coding individual that is currently undergoing dynamic mutation from all the second intention coding individuals, and randomly select a target mutation operation from a preset plurality of mutation operations, wherein the mutation operation includes one of the following: random change, adjacent cycle exchange, random sharing, and random merging.

[0066] Step 53, performing an edge operation corresponding to the target mutation operation on the selected second intention coding individual, to generate a current matrix coding individual that completes the genetic operation, wherein the current matrix coding individual includes multiple third sub-coding bodies, and the third sub-coding body includes the second sub-coding body and one of the candidate sub-coding bodies.

[0067] In some embodiments, when the mutation operation is a random change, reference Figure 5 In step 53, the edge operation corresponding to the target mutation operation is performed on the selected second intention coding individual to generate the matrix coding individual that completes the genetic operation, which is achieved through the following steps:

[0068] Step 531-1, randomly selecting a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intention coding bodies;

[0069] Step 531-2, randomly selecting at least one delivery cycle from the multiple delivery cycles corresponding to the selected fourth sub-code body, and randomly changing the first delivery quantity corresponding to the corresponding delivery cycle to obtain a new delivery quantity, and using an exhaustive method to determine an alternative supplier of the target material that delivers the new delivery quantity based on the material discount rate and the transportation cost parameter;

[0070] Step 531 - 3 : based on the new delivery quantity and the determined candidate supply object, the selected fourth sub-encoding body is updated to generate a third sub-encoding body located at the same encoding position.

[0071] In this embodiment, the delivery quantity 150 of delivery cycle 2 of order cycle 1 is randomly changed to 50 to generate a new delivery quantity, thereby realizing a dynamic mutation operation based on random changes.

[0072] In some embodiments, when the mutation operation is adjacent cycle swapping, reference Figure 6 In step 53, the edge operation corresponding to the target mutation operation is performed on the selected second intention coding individual to generate the matrix coding individual that completes the genetic operation, which is achieved through the following steps:

[0073] Step 532-1: randomly select a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intended coding bodies.

[0074] Step 532-2, randomly select two adjacent delivery cycles from the multiple delivery cycles corresponding to the selected fourth sub-code body, and exchange the replenishment supply objects and replenishment supply quantities corresponding to the two adjacent delivery cycles to generate a third sub-code body located at the same coding position.

[0075] In this embodiment, the delivery quantity 150 of delivery cycle 1 in ordering cycle 1 and the delivery quantity 0 of delivery cycle 2 in ordering cycle 1 are swapped. Then, in the new third sub-code body, the delivery quantity of delivery cycle 1 in ordering cycle 1 is 0, and the delivery quantity of delivery cycle 2 in ordering cycle 1 is 150, thereby realizing a dynamic variation operation based on the exchange of adjacent cycles.

[0076] In some embodiments, when the mutation operation is random sharing, refer to Figure 7 In step 53, the edge operation corresponding to the target mutation operation is performed on the selected second intention coding individual to generate the matrix coding individual that completes the genetic operation, which is achieved through the following steps:

[0077] Step 533 - 1 : randomly select a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intention coding bodies.

[0078] Step 533-2: randomly select a target delivery cycle and a shared delivery cycle from the multiple delivery cycles corresponding to the selected fourth sub-code body, and randomly allocate a first shared delivery amount from the first delivery amount corresponding to the target delivery cycle, and allocate the first shared delivery amount to the shared delivery cycle;

[0079] Step 533-3, based on the material discount rate and the transportation cost parameters, use the exhaustive method to determine the target material of the remaining delivery quantity of the target delivery period and the alternative supply objects of the first shared delivery quantity, and obtain the first supply object and the second supply object.

[0080] Step 533 - 4 , based on the remaining delivery quantity in the target delivery period and the first supply object, the first shared delivery quantity and the second supply object, the selected fourth sub-code body is updated to generate a third sub-code body located at the same code position.

[0081] In this embodiment, in order to realize the re-batching operation of the delivery quantity in the batch delivery process, a mutation operator is used to randomly distribute the delivery quantity of a certain delivery cycle to other delivery cycles. That is, the delivery quantity 150 of delivery cycle 2 in ordering cycle 1 is randomly distributed to the two delivery cycles, so that the delivery quantity of delivery cycle 1 under ordering cycle 1 is 100, and the delivery quantity of delivery cycle 2 under ordering cycle 1 is 50, thereby realizing a dynamic mutation operation based on random sharing.

[0082] In some embodiments, when the mutation operation is random merging, reference Figure 8 In step 53, the edge operation corresponding to the target mutation operation is performed on the selected second intention coding individual to generate the matrix coding individual that completes the genetic operation, which is achieved through the following steps:

[0083] Step 534 - 1 : randomly select a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intention coding bodies.

[0084] Step 534-2, randomly select two delivery cycles from the multiple delivery cycles corresponding to the selected fourth sub-code body, merge the first delivery quantity of one of the two delivery cycles into the other delivery cycle, and determine the merged delivery quantity and the remaining delivery quantity corresponding to the two delivery cycles after the random merger.

[0085] Step 534-3, based on the material discount rate and the transportation cost parameters, the exhaustive method is used to determine the alternative supply objects corresponding to the target materials for the combined delivery quantity and the remaining delivery quantity, respectively, to obtain the third supply object and the fourth supply object.

[0086] Step 534-4: update the selected fourth sub-encoding body according to the remaining delivery quantity and the third supply object, the remaining delivery quantity and the second supply object, so as to generate a third sub-encoding body located at the same encoding position.

[0087] In this embodiment, in order to realize the merging operation of delivery quantities in the batch delivery process, a mutation operator is used to randomly merge the delivery quantities of two delivery cycles into one delivery cycle. That is, the delivery quantity 50 of delivery cycle 1 under ordering cycle 1 and the delivery quantity 100 of delivery cycle 2 under ordering cycle 1 are merged and merged into one delivery cycle, so that the delivery quantity of delivery cycle 2 under ordering cycle 1 is 150, realizing a dynamic mutation operation based on random merging.

[0088] Through the above steps 51 to 53, the genetic operation of dynamic variation is realized.

[0089] In some of the embodiments, after taking the candidate sub-encoding bodies in the candidate matrix-type encoding individuals as candidate particles, a particle swarm search process is performed on the plurality of candidate particles based on a preset particle swarm optimization algorithm PSO, which is implemented by the following steps:

[0090] Step 61, obtaining an initial particle population including a plurality of candidate particles, and determining an initial particle position and a preset initial particle velocity of each candidate particle, wherein the initial particle position is used to characterize the material replenishment decision information corresponding to the candidate sub-code body.

[0091] Step 62, based on a preset dynamic adaptive change method, respectively determine the dynamic adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient corresponding to the current particle search, wherein the dynamic sub-adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient are generated by adaptively changing according to the fitness corresponding to each candidate sub-encoder and the number of search iterations.

[0092] Step 63, based on the initial particle position, the initial particle velocity, the dynamic adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient, the particle swarm optimization algorithm is used to perform at least one population search iteration on the initial particle population until the total fitness corresponding to the matrix-type coding individual corresponding to the generated particle population is no greater than the preset fitness threshold, and at least one target matrix-type coding individual including multiple candidate particles is obtained.

[0093] In this embodiment, the particle swarm parameters change adaptively. It can be understood that the particle swarm algorithm does not have mass and direction when initialized. In the optimal solution space, each independent particle moves at its own speed. The mass of each particle is determined by the adaptability function, and the direction and displacement of the particle are determined by its speed variable. In the process of moving, the particle continuously adjusts its own moving path and direction through its own moving experience and the moving experience of other particles. Usually, the particles can find their optimal position in the space and obtain the optimal solution.

[0094] In this embodiment, the corresponding sub-encoding body is treated as a particle in the particle swarm optimization algorithm, thereby realizing a population search based on the particle swarm optimization algorithm PSO; in this embodiment, the particle swarm search process needs to use a dynamic adaptive inertia factor and acceleration coefficients c1 and c2, Affects the global and local search capabilities of particles. c1 and c2 determine the degree of attention paid by particles to individual optimal solutions and global optimal solutions respectively. In this embodiment, the dynamic adaptive inertia factor The acceleration coefficients c1 and c2 are configured and adjusted during the search process so that they can adapt to the fitness value of the chromosome and the number of iterations. Compared with the fixed inertia factor, the dynamic adaptive inertia factor It can better balance the global and local search capabilities. The adaptive ability is reflected by the acceleration coefficients c1 and c2. By dynamically adjusting the acceleration factor, the diversity of the particle swarm in the early search stage and the convergence of the algorithm in the later search stage can be achieved. It can better balance the global and local search capabilities; in this embodiment, the dynamic adaptive inertia factor The acceleration coefficients c1 and c2 are calculated using the following formula: in, Represents the maximum dynamic adaptive inertia factor , Represents the minimum dynamic adaptive inertia factor , genmax indicates the maximum number of iterations, It is The total fitness of the nth candidate matrix-coded individual in the iterated population, , are the maximum total fitness and the minimum total fitness in the population respectively.

[0095] In this embodiment, the dynamic adaptive inertia factor After adaptively updating the acceleration coefficients c1 and c2, the candidate particles are iterated one by one according to the following formula: in, represents the initial velocity of the nth candidate matrix-coded individual and all its corresponding candidate particles in the population of the gen iteration, Represents the position of the nth candidate matrix-coded individual in the population of the gen iteration; represents the inertia weight; r1 and r2 represent random numbers in the range of (0,1), Wbest represents the optimal position of the candidate matrix coding individual, and GWbest represents the global optimal position.

[0096] In this embodiment, the position of the candidate particle is updated according to the speed. In this embodiment, the position can also be regarded as the structure of the matrix-type coding individual, and the position update refers to the change of the gene value on the corresponding sub-coding body, which is expressed by W n,gen+1 represents the position of the nth matrix-coded individual in the population of the gen+1th iteration, then: .

[0097] In this embodiment, reference Fig. 9 ,The particle position update operation in this embodiment includes: first, the inertial position of the particle, which is determined by the velocity of the particle itself, that is, ; followed by the best position of the particle in history, which can be expressed as Wbest; finally, the position of the particle in the current population with the global optimal solution GWbest. The position of the particle after update is jointly contributed by these three factors. The particle will be in accordance with the dynamic adaptive inertia factor , acceleration factor and Add them together and finally get a new particle.

[0098] In some of the embodiments, a target particle is selected from a plurality of generated candidate particles to obtain a decision result, which is achieved by the following steps:

[0099] Step 71, obtaining at least one target matrix-type coding individual corresponding to a candidate particle, and calculating the fitness of the candidate sub-coding body corresponding to each candidate particle;

[0100] Step 72, according to the fitness of the candidate sub-code body, select the candidate particle with the smallest fitness from multiple candidate particles to obtain the target particle, wherein the material replenishment decision information corresponding to the candidate sub-code body corresponding to the target particle is the target material replenishment decision information.

[0101] This embodiment also provides a material ordering and replenishment decision-making device based on GA-PSO, which is used to implement the above embodiments and preferred implementations, and will not be repeated for what has been described. As used below, the terms "module", "unit", "subunit", etc. can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0102] Fig.10 is a structural block diagram of a material ordering and replenishment decision-making device based on GA-PSO according to an embodiment of the present application, such as Fig.10 As shown, the device includes an acquisition module 101, an encoding module 102, an iteration module 103 and a processing module 104, wherein:

[0103] The acquisition module 101 is used to determine the multiple delivery cycles agreed upon in each ordering cycle among the preset multiple ordering cycles, and to obtain the material ordering quantities corresponding to the multiple target materials to be replenished in each ordering cycle, wherein the material ordering quantity is used to represent the total amount of materials required to be delivered for each target material in the corresponding multiple delivery cycles;

[0104] The encoding module 102 is coupled to the acquisition module 101 and is used to encode the number of delivery cycles and the material order quantities corresponding to the multiple target materials, and generate multiple first matrix-type encoding individuals, wherein the first matrix-type encoding individuals include multiple first matrix-type sub-encoding bodies having multiple first sub-encoding bodies, and the first sub-encoding bodies are used to represent the material replenishment decision information of a target material in multiple delivery cycles corresponding to an ordering cycle, and the material replenishment decision information includes a replenishment supply object and a replenishment supply quantity;

[0105] The iteration module 103 is coupled to the encoding module 102 and is used to determine the fitness of the second sub-encoding body corresponding to each second matrix-type encoding individual generated in the iteration after performing at least one genetic operation using a preset genetic algorithm GA based on multiple first matrix-type encoding individuals, and perform genetic operation iterations based on the fitness and the second matrix-type encoding individuals until multiple candidate matrix-type encoding individuals are generated, wherein the genetic evolution operation includes: individual selection based on the roulette method and the elite retention strategy, adaptive crossover based on gene position, and dynamic mutation, and the fitness is determined according to the ordering cost, transportation cost, storage cost, and emergency transfer cost generated by the replenishment supply object corresponding to the second sub-encoding body in replenishing the target material of the replenishment supply quantity;

[0106] The processing module 104 is coupled to the iteration module 103 and is used to perform particle swarm search iterative processing on multiple candidate particles based on a preset particle swarm optimization algorithm PSO after taking the candidate sub-coding bodies in the candidate matrix-type coding individuals as candidate particles, and select a target particle from the generated multiple candidate particles to obtain a decision result, wherein the decision result includes the target material replenishment decision information corresponding to the target particle, and the particle swarm optimization algorithm PSO adopts a dynamic adaptive inertia factor and a dynamically adjusted acceleration coefficient.

[0107] In some embodiments, the encoding module 102 further includes:

[0108] An allocation unit is used to determine target information of multiple alternative supply objects, and randomly allocate the material order quantity corresponding to each target material within multiple delivery cycles of an ordering cycle to obtain the first delivery quantity of each target material within a delivery cycle, wherein the target information includes material discount rate and transportation cost parameters.

[0109] The selection unit is coupled to the allocation unit and is used to determine, based on the material discount rate and the transportation cost parameter, an alternative supply object for delivering a target material of a first delivery quantity from a plurality of alternative supply objects by an exhaustive method, and use the determined alternative supply object and the first delivery quantity as a replenishment supply object and a replenishment supply quantity, respectively, to generate a first sub-code body corresponding to the corresponding target material, wherein the replenishment supply object is the alternative supply object with the smallest first cost generated by supplying the target material of the first delivery quantity, and the first cost is the sum of the ordering cost and the transportation cost determined respectively according to the material discount rate and the transportation cost parameter.

[0110] The encoding unit is coupled to the selection unit and is used to encode all first sub-encoding bodies corresponding to multiple target materials in the same ordering cycle into a first matrix-type sub-encoding body, and encode the first matrix-type sub-encoding bodies corresponding to multiple ordering cycles into a first matrix-type encoding individual, and generate multiple first matrix-type encoding individuals through population initialization.

[0111] In some of the embodiments, after generating a plurality of first matrix-type coding individuals, the iteration module 93 is further used to determine a current matrix-type coding individual for individual selection based on the roulette method and the elite retention strategy, and based on the fitness of all current sub-coding bodies corresponding to each current matrix-type coding individual, calculate the total fitness corresponding to each current matrix-type coding individual according to a preset objective function, wherein the current matrix-type coding individual includes one of the following: a first matrix-type coding individual, a second matrix-type coding individual; according to the total fitness, screen out from the plurality of current matrix-type coding individuals the individual whose total fitness is less than a preset fitness threshold The current matrix-type coded individual is used to obtain a first candidate coded individual, wherein the first candidate coded individual is used to characterize the best individual in the current group; after calculating the first selection probability based on the transposed fitness generated by transposing each total fitness, the first selection probability is used to perform individual selection based on the roulette method on all current matrix-type coded individuals except the first candidate coded individual to obtain a preset number of second candidate coded individuals, and all the first candidate coded individuals and all the second candidate coded individuals are used as the first intention coded individuals corresponding to the selection result of completing the individual selection based on the roulette method and the elite retention strategy.

[0112] In some embodiments, the iteration module 103 is further used to calculate the total fitness F according to the following formula: F=Min(PC+SC+TC+EC) Among them, PC represents the total material ordering cost corresponding to a corresponding matrix-coded individual; SC represents the total target material storage cost corresponding to a corresponding matrix-coded individual; TC represents the total material transportation cost corresponding to a corresponding matrix-coded individual; EC represents the total material emergency replenishment cost corresponding to a corresponding matrix-coded individual; i represents the i-th target material, I represents the set of target materials, i∈I; j represents the j-th replenishment supply object, J represents the set of replenishment supply objects, j∈J; q represents the q-th ordering cycle, Q represents the set of ordering cycles, q∈Q; k represents the k-th discount plan of the target material, K represents the set of discount plans, k∈K; m represents the m-th delivery cycle, M represents the set of delivery cycles, m∈M; v represents the v-th vehicle, V represents the set of vehicles, v∈V; is the discount rate corresponding to the kth discount scheme k for the i-th target material by the j-th replenishment supply object in the q-th ordering period; is the base price of the i-th target material for the j-th replenishment supply object in the q-th ordering cycle; It is Unit storage cost of the target material; It is Order cycle Delivery cycle The actual storage quantity of the target material; is the unit rental warehouse storage cost of the i-th target material; It is Order cycle Delivery cycle Storage capacity of rented warehouses for various materials; Replenishment supply object The corresponding distance cost; is the rental cost corresponding to the vth vehicle; It is Unit emergency transfer cost of the target material; It is Order cycle Delivery cycle The out-of-stock quantity of the target material; is a decision variable, indicating whether the vth vehicle is selected when the i-th target material in the m-th delivery cycle under the q-th ordering cycle is delivered to the j-th replenishment supply object; is a decision variable, which indicates the quantity of the i-th target material selected to be ordered from the j-th replenishment supply object in the q-th ordering cycle; is a decision variable, indicating whether the i-th target material in the m-th delivery cycle under the q-th ordering cycle chooses to be replenished from the j-th replenishment supply object; is a decision variable, indicating whether the vth vehicle is selected when replenishing the i-th target material in the m-th delivery cycle under the q-th ordering cycle from the j-th replenishment supply object.

[0113] In some of the embodiments, after generating the first intention coding individual, the iteration module 103 is also used to select the maximum total fitness, the minimum total fitness and the average total fitness from the total fitness corresponding to all the first intention coding individuals, and calculate the current crossover probability corresponding to each first intention coding individual in the current iteration based on the total fitness corresponding to each first intention coding individual and the maximum total fitness, the minimum total fitness and the average total fitness; randomly select a preset number of intention sub-coding bodies from all the intention sub-coding bodies corresponding to the first intention coding individual as the first crossover sub-coding body group corresponding to each first intention coding individual, and use the intention sub-coding bodies other than the first crossover sub-coding body group as the second crossover sub-coding body group corresponding to the first intention coding body; based on the current crossover probability, select a preset number of intention sub-coding bodies from all the first intention coding individuals After selecting two first intended coding individuals from the body in turn to obtain the first parent coding individual and the second parent coding individual, the first crossover sub-coding body group corresponding to the first parent coding individual and the second crossover sub-coding body group corresponding to the second parent coding individual are filled into the first child coding individual according to the corresponding coding positions in the first parent coding individual and the second parent coding individual, and the first crossover sub-coding body group corresponding to the second parent coding individual and the second crossover sub-coding body group corresponding to the first parent coding individual are filled into the second child coding individual according to the corresponding coding positions in the first parent coding individual and the second parent coding individual; the first child coding individual and the second child coding individual generated each time are used as the second intended coding individual, and the corresponding second intended coding individual is used as the result of completing the genetic operation of adaptive crossover based on gene position at that time.

[0114] In some of the embodiments, after generating the second intention coding individual, the iteration module 103 is further used to determine the current mutation probability corresponding to the current mutation operation according to a preset dynamic adjustment method; based on the current mutation probability, the second intention coding individual that is currently undergoing dynamic mutation is selected from all the second intention coding individuals, and a target mutation operation is randomly selected from a preset plurality of mutation operations, wherein the mutation operation includes one of the following: random change, adjacent cycle exchange, random sharing, and random merging; the edge operation corresponding to the target mutation operation is performed on the selected second intention coding individual to generate the current matrix coding individual that completes the genetic operation, wherein the current matrix coding individual includes multiple third sub-coding bodies, and the third sub-coding body includes a second sub-coding body and one of a candidate sub-coding body.

[0115] In some of the embodiments, when the mutation operation is a random change, the iteration module 103 is also used to randomly select a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intention coding individuals; randomly select at least one delivery cycle from the multiple delivery cycles corresponding to the selected fourth sub-coding body, and randomly change the first delivery quantity corresponding to the corresponding delivery cycle to obtain a new delivery quantity, and determine the alternative supply object for delivering the target material of the new delivery quantity by exhaustive method based on the material discount rate and transportation cost parameters; based on the new delivery quantity and the determined alternative supply object, update the selected fourth sub-coding body to generate a third sub-coding body located at the same coding position.

[0116] In some embodiments, when the mutation operation is adjacent period exchange, the iteration module 103 is further used to randomly select a fourth sub-encoding body corresponding to at least one target material from all fourth sub-encoding bodies corresponding to the selected second intention encoding body; randomly select two adjacent delivery cycles from the multiple delivery cycles corresponding to the selected fourth sub-encoding body, and exchange the replenishment supply objects and replenishment supply quantities corresponding to the two adjacent delivery cycles to generate a third sub-encoding body located at the same encoding position;

[0117] In some embodiments, when the mutation operation is random sharing, the iteration module 103 is also used to randomly select a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intention coding individual; randomly select a target delivery cycle and a shared delivery cycle from the multiple delivery cycles corresponding to the selected fourth sub-coding body, and randomly share a first shared delivery quantity from the first delivery quantity corresponding to the target delivery cycle, and share the first shared delivery quantity to the shared delivery cycle; based on the material discount rate and the transportation cost parameter, the target material of the remaining delivery quantity of the target delivery cycle and the candidate supply object of the first shared delivery quantity are determined by exhaustive method to obtain the first supply object and the second supply object; according to the remaining delivery quantity of the target delivery cycle and the first supply object, the first shared delivery quantity and the second supply object, the selected fourth sub-coding body is updated to generate a third sub-coding body located at the same coding position;

[0118] In some of the embodiments, when the mutation operation is a random merge, the iteration module 103 is also used to randomly select a fourth sub-coding body corresponding to at least one target material from all fourth sub-coding bodies corresponding to the selected second intention coding individuals; randomly select two delivery cycles from the multiple delivery cycles corresponding to the selected fourth sub-coding body, and merge the first delivery quantity of one of the two delivery cycles into the other delivery cycle, and determine the combined delivery quantity and the remaining delivery quantity corresponding to the two delivery cycles after the random merger; based on the material discount rate and the transportation cost parameters, use the exhaustive method to determine the alternative supply objects corresponding to the target materials that deliver the combined delivery quantity and the remaining delivery quantity, respectively, to obtain the third supply object and the fourth supply object; according to the remaining delivery quantity and the third supply object, the remaining delivery quantity and the second supply object, the selected fourth sub-coding body is updated to generate a third sub-coding body located at the same coding position.

[0119] In some embodiments, the processing module 104 further includes:

[0120] The acquisition unit is used to acquire an initial particle population including a plurality of candidate particles, and determine an initial particle position and a preset initial particle speed of each candidate particle, wherein the initial particle position is used to characterize the material replenishment decision information corresponding to the candidate sub-encoder.

[0121] The determination unit is coupled to the acquisition unit and is used to determine the dynamic adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient corresponding to the current particle search based on a preset dynamic adaptive change method, wherein the dynamic sub-adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient are generated by adaptively changing according to the fitness corresponding to each candidate sub-encoder and the number of search iterations.

[0122] The search unit is coupled to the determination unit and is used to perform at least one population search iteration on the initial particle population using a particle swarm optimization algorithm based on the initial particle position, the initial particle velocity, the dynamic adaptive inertia factor, the first acceleration coefficient, and the second acceleration coefficient, until the total fitness corresponding to the matrix-type coded individual corresponding to the generated particle population is no greater than a preset fitness threshold, and at least one target matrix-type coded individual including multiple candidate particles is obtained.

[0123] In some of the embodiments, the processing module 104 is further used to obtain a candidate particle corresponding to at least one target matrix-type coding individual, and calculate the fitness of the candidate sub-coding body corresponding to each candidate particle; according to the fitness of the candidate sub-coding body, the candidate particle with the smallest fitness is selected from multiple candidate particles to obtain the target particle, wherein the material replenishment decision information corresponding to the candidate sub-coding body corresponding to the target particle is the target material replenishment decision information.

[0124] This embodiment further provides a service platform, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0125] Optionally, the service platform may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0126] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0127] S1, in the preset multiple ordering cycles, determine the multiple delivery cycles agreed upon in each ordering cycle, and obtain the material ordering quantities corresponding to the multiple target materials to be replenished in each ordering cycle. The material ordering quantity is used to represent the total amount of materials that need to be delivered for each target material in the corresponding multiple delivery cycles.

[0128] S2, encode the number of delivery cycles and the material order quantities corresponding to the multiple target materials, and generate multiple first matrix-type coding individuals, the first matrix-type coding individuals include multiple first matrix-type sub-coding bodies with multiple first sub-coding bodies, the first sub-coding bodies are used to represent the material replenishment decision information of a target material in multiple delivery cycles corresponding to an ordering cycle, the material replenishment decision information includes the replenishment supply object and the replenishment supply quantity.

[0129] S3, after performing at least one genetic operation based on multiple first matrix-type coding individuals using a preset genetic algorithm GA, determine the fitness of the second sub-coding body corresponding to each second matrix-type coding individual generated in the current iteration, and perform genetic operation iterations based on the fitness and the second matrix-type coding individuals until multiple candidate matrix-type coding individuals are generated.

[0130] S4, after taking the candidate sub-coding bodies in the candidate matrix-type coding individuals as candidate particles, based on the preset particle swarm optimization algorithm PSO, a particle swarm search iterative process is performed on multiple candidate particles, and a target particle is selected from the generated multiple candidate particles to obtain a decision result, which includes the target material replenishment decision information corresponding to the target particle. The particle swarm optimization algorithm PSO adopts a dynamic adaptive inertia factor and a dynamically adjusted acceleration coefficient.

[0131] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0132] In addition, in combination with the GA-PSO-based material ordering and replenishment decision-making method in the above embodiments, the present application embodiment can provide a storage medium for implementation. The storage medium stores a computer program; when the computer program is executed by the processor, any one of the GA-PSO-based material ordering and replenishment decision-making methods in the above embodiments is implemented.

[0133] Those skilled in the art should understand that the technical features of the above embodiments may be arbitrarily combined. 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 specification.

[0134] The above 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 invention patent. 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 patent of the present application shall be subject to the attached claims.

Claims

1. A material ordering and replenishment decision-making method based on GA-PSO, characterized in that: include: In the preset multiple ordering cycles, multiple delivery cycles agreed upon in each ordering cycle are determined, and the material ordering quantities corresponding to the multiple target materials to be replenished in each ordering cycle are obtained, wherein the material ordering quantities are used to represent the total amount of materials required to be delivered for each target material in the corresponding multiple delivery cycles; Encode the number of the delivery cycles and the material order quantities corresponding to the multiple target materials respectively, and generate a plurality of first matrix-type code individuals, wherein the first matrix-type code individuals include a plurality of first matrix-type sub-code bodies having a plurality of first sub-code bodies, and the first sub-code bodies are used to represent material replenishment decision information of a target material in a plurality of the delivery cycles corresponding to one order cycle, and the material replenishment decision information includes a replenishment supply object and a replenishment supply quantity; After performing at least one genetic operation using a preset genetic algorithm GA based on a plurality of the first matrix-type coding individuals, determining the fitness of a second sub-coding body corresponding to each second matrix-type coding individual generated in the current iteration, and performing genetic operation iterations based on the fitness and the second matrix-type coding individuals until a plurality of candidate matrix-type coding individuals are generated, wherein the genetic evolution operation includes: individual selection based on a roulette wheel method and an elite retention strategy, adaptive crossover based on gene position, and dynamic mutation, and the fitness is determined based on the ordering cost, transportation cost, storage cost, and emergency transfer cost generated by the replenishment supply object corresponding to the second sub-coding body in replenishing the target material of the replenishment supply quantity; After taking the candidate sub-coding bodies in the candidate matrix-type coding individuals as candidate particles, based on the preset particle swarm optimization algorithm PSO, a particle swarm search iterative process is performed on the multiple candidate particles, and a target particle is selected from the generated multiple candidate particles to obtain a decision result, wherein the decision result includes the target material replenishment decision information corresponding to the target particle, and the particle swarm optimization algorithm PSO adopts a dynamic adaptive inertia factor and a dynamically adjusted acceleration coefficient.

2. The method according to claim 1, characterized in that: The number of the delivery cycles and the material order quantities corresponding to the multiple target materials are encoded to generate a plurality of first matrix-type encoding individuals, including: Determine target information of multiple candidate supply objects, and randomly distribute the material order quantity corresponding to each target material in multiple delivery cycles of one order cycle to obtain a first delivery quantity of each target material in one delivery cycle, wherein the target information includes a material discount rate and a transportation cost parameter; Based on the material discount rate and the transportation cost parameter, an alternative supply object that delivers the target material of the first delivery quantity is determined from among the multiple alternative supply objects by using an exhaustive method, and the determined alternative supply object and the first delivery quantity are used as the replenishment supply object and the replenishment supply quantity, respectively, to generate a first sub-code body corresponding to the corresponding target material, wherein the replenishment supply object is the alternative supply object that generates the minimum first cost for supplying the target material of the first delivery quantity, and the first cost is the sum of the ordering cost and the transportation cost determined respectively according to the material discount rate and the transportation cost parameter; All the first sub-encoding bodies corresponding to multiple target materials in the same ordering cycle are encoded into a first matrix-type sub-encoding body, and the first matrix-type sub-encoding bodies corresponding to multiple ordering cycles are respectively encoded into a first matrix-type encoding individual, and multiple first matrix-type encoding individuals are generated through population initialization.

3. The method according to claim 2, characterized in that After generating a plurality of first matrix-type coding individuals, the method further comprises: Determine a current matrix-type coding individual for individual selection based on the roulette method and the elite retention strategy, and calculate the total fitness corresponding to each current matrix-type coding individual according to a preset objective function based on the fitness of all current sub-coding bodies corresponding to each current matrix-type coding individual, wherein the current matrix-type coding individual includes one of the following: the first matrix-type coding individual, the second matrix-type coding individual; According to the total fitness, from the plurality of current matrix-type coding individuals, the current matrix-type coding individuals whose total fitness is less than a preset fitness threshold are screened out to obtain a first candidate coding individual, wherein the first candidate coding individual is used to represent the best individual in the current group; After calculating the first selection probability based on the transposed fitness generated by transposing each of the total fitnesses, the first selection probability is used to perform individual selection based on the roulette method on all of the current matrix-type coding individuals except the first candidate coding individuals to obtain a preset number of second candidate coding individuals, and all of the first candidate coding individuals and all of the second candidate coding individuals are used as the first intended coding individuals corresponding to the selection result of completing the individual selection based on the roulette method and the elite retention strategy.

4. The method according to claim 3, characterized in that The total fitness F is calculated according to the following formula: F=Min(PC+SC+TC+EC) Among them, PC represents the total material ordering cost corresponding to a corresponding matrix-coded individual; SC represents the total target material storage cost corresponding to a corresponding matrix-coded individual; TC represents the total material transportation cost corresponding to a corresponding matrix-coded individual; EC represents the total material emergency replenishment cost corresponding to a corresponding matrix-coded individual; i represents the i-th target material, I represents the set of target materials, i∈I; j represents the j-th replenishment supply object, J represents the set of replenishment supply objects, j∈J; q represents the q-th ordering cycle, Q represents the set of ordering cycles, q∈Q; k represents the k-th discount plan of the target material, K represents the set of discount plans, k∈K; m represents the m-th delivery cycle, M represents the set of delivery cycles, m∈M; v represents the v-th vehicle, V represents the set of vehicles, v∈V; is the discount rate corresponding to the kth discount scheme k for the i-th target material by the j-th replenishment supply object in the q-th ordering period; is the base price of the i-th target material for the j-th replenishment supply object in the q-th ordering cycle; is the unit storage cost of the i-th target material; is the actual storage quantity of the i-th target material in the m-th delivery cycle under the q-th ordering cycle; is the unit rental warehouse storage cost of the i-th target material; is the storage capacity of the rented warehouse for the i-th material in the m-th delivery cycle under the q-th ordering cycle; is the distance cost corresponding to the replenishment supply object j; is the rental cost corresponding to the vth vehicle; is the unit emergency transfer cost of the i-th target material; is the out-of-stock quantity of the i-th target material in the m-th delivery cycle under the q-th ordering cycle; is a decision variable, indicating whether the vth vehicle is selected when the i-th target material in the m-th delivery cycle under the q-th ordering cycle is delivered to the j-th replenishment supply object; is a decision variable, which indicates the quantity of the i-th target material selected to be ordered from the j-th replenishment supply object in the q-th ordering cycle; is a decision variable, indicating whether the i-th target material in the m-th delivery cycle under the q-th ordering cycle chooses to be replenished from the j-th replenishment supply object; is a decision variable, indicating whether the vth vehicle is selected when replenishing the i-th target material in the m-th delivery cycle under the q-th ordering cycle from the j-th replenishment supply object.

5. The method according to claim 3, characterized in that: After generating the first intention coding individual, the method further includes: Selecting the maximum total fitness, the minimum total fitness and the average total fitness from the total fitness corresponding to all the first intention coding individuals, and calculating the current crossover probability corresponding to each first intention coding individual in the iteration based on the total fitness corresponding to each first intention coding individual and the maximum total fitness, the minimum total fitness and the average total fitness; A preset number of the intended sub-encoding bodies randomly selected from all the intended sub-encoding bodies corresponding to the first intended encoding individual are used as the first cross-sub-encoding body group corresponding to each of the first intended encoding individuals, and the intended sub-encoding bodies except the first cross-sub-encoding body group are used as the second cross-sub-encoding body group corresponding to the first intended encoding body; After selecting two of the first intended coding individuals from all the first intended coding individuals one by one based on the current crossover probability to obtain a first parent coding individual and a second parent coding individual, the first crossover sub-coding body group corresponding to the first parent coding individual and the second crossover sub-coding body group corresponding to the second parent coding individual are filled into the first child coding individual according to the corresponding coding positions in the first parent coding individual and the second parent coding individual, and the first crossover sub-coding body group corresponding to the second parent coding individual and the second crossover sub-coding body group corresponding to the first parent coding individual are filled into the second child coding individual according to the corresponding coding positions in the first parent coding individual and the second parent coding individual; The first offspring coding individual and the second offspring coding individual generated each time are used as second intended coding individuals, and the corresponding second intended coding individuals are used as the result of completing the genetic operation of adaptive crossover based on gene position at that time.

6. The method according to claim 5, characterized in that After generating the second intention coding individual, the method further includes: According to the preset dynamic adjustment method, determine the current mutation probability corresponding to the current mutation operation; Based on the current mutation probability, the second intention coding individual currently undergoing dynamic mutation is selected from all the second intention coding individuals, and a target mutation operation is randomly selected from a plurality of preset mutation operations, wherein the mutation operation includes one of the following: random change, adjacent cycle exchange, random sharing, and random merging; The edge operation corresponding to the target mutation operation is performed on the selected second intention coding individual to generate the current matrix-type coding individual that completes the genetic operation, wherein the current matrix-type coding individual includes multiple third sub-coding bodies, and the third sub-coding body includes the second sub-coding body and one of the candidate sub-coding bodies.

7. The method according to claim 6, characterized in that When the mutation operation is a random change, the method includes: Randomly select one of the fourth sub-encoding bodies corresponding to at least one target material from all the fourth sub-encoding bodies corresponding to the selected second intention encoding bodies; Randomly select at least one delivery cycle from among the multiple delivery cycles corresponding to the selected fourth sub-code body, and randomly change the first delivery quantity corresponding to the corresponding delivery cycle to obtain a new delivery quantity, and determine the alternative supplier of the target material for delivering the new delivery quantity by using an exhaustive method based on the material discount rate and the transportation cost parameter; Based on the new delivery amount and the determined candidate supply object, the selected fourth sub-encoding body is updated to generate the third sub-encoding body located at the same encoding position; When the mutation operation is adjacent cycle exchange, the method includes: Randomly select one of the fourth sub-encoding bodies corresponding to at least one target material from all the fourth sub-encoding bodies corresponding to the selected second intention encoding bodies; Randomly select two adjacent delivery cycles from among the multiple delivery cycles corresponding to the selected fourth sub-code body, and exchange the replenishment supply objects and the replenishment supply quantities corresponding to the two adjacent delivery cycles to generate the third sub-code body located at the same coding position; When the mutation operation is random sharing, the method includes: Randomly select one of the fourth sub-encoding bodies corresponding to at least one target material from all the fourth sub-encoding bodies corresponding to the selected second intention encoding bodies; Randomly select a target delivery cycle and a shared delivery cycle from the multiple delivery cycles corresponding to the selected fourth sub-code body, randomly allocate a first shared delivery amount from the first delivery amount corresponding to the target delivery cycle, and allocate the first shared delivery amount to the shared delivery cycle; Based on the material discount rate and the transportation cost parameter, the target material for delivering the remaining delivery quantity of the target delivery period and the candidate supply object of the first shared delivery quantity are determined by exhaustive method to obtain a first supply object and a second supply object; According to the remaining delivery amount of the target delivery period and the first supply object, the first shared delivery amount and the second supply object, the selected fourth sub-encoding body is updated to generate the third sub-encoding body located at the same encoding position; When the mutation operation is random merging, the method includes: Randomly select one of the fourth sub-encoding bodies corresponding to at least one target material from all the fourth sub-encoding bodies corresponding to the selected second intention encoding bodies; Randomly select two delivery cycles from among the multiple delivery cycles corresponding to the selected fourth sub-code body, merge the first delivery quantity of one of the two delivery cycles into the other delivery cycle, and determine the merged delivery quantity and the remaining delivery quantity respectively corresponding to the two delivery cycles after the random merger; Based on the material discount rate and the transportation cost parameter, the candidate supply objects corresponding to the target materials for delivering the combined delivery quantity and the remaining delivery quantity are determined by exhaustive method to obtain a third supply object and a fourth supply object; The selected fourth sub-encoding body is updated according to the remaining delivery amount and the third supply object, and the remaining delivery amount and the second supply object, so as to generate the third sub-encoding body located at the same encoding position.

8. The method according to claim 5, characterized in that After the candidate sub-encoding bodies in the candidate matrix-type encoding individuals are taken as candidate particles, a particle swarm search process is performed on the plurality of candidate particles based on a preset particle swarm optimization algorithm PSO, including: Acquire an initial particle population including a plurality of candidate particles, and determine an initial particle position and a preset initial particle speed of each candidate particle, wherein the initial particle position is used to characterize the material replenishment decision information corresponding to the candidate sub-encoder; Based on a preset dynamic adaptive change method, respectively determine the dynamic adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient corresponding to the current particle search, wherein the dynamic sub-adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient are generated by adaptively changing according to the fitness corresponding to each of the candidate sub-encoders and the number of search iterations; Based on the initial particle position, the initial particle velocity, the dynamic adaptive inertia factor, the first acceleration coefficient and the second acceleration coefficient, the particle swarm optimization algorithm is used to perform at least one population search iteration on the initial particle population until the total fitness corresponding to the matrix-type coded individual corresponding to the generated particle population is not greater than a preset fitness threshold, and at least one target matrix-type coded individual including a plurality of the candidate particles is obtained.

9. The method according to claim 8, characterized in that Select the target particle from the generated multiple candidate particles and get the decision result, including: Acquire at least one of the target matrix-type coding individuals corresponding to the candidate particle, and calculate the fitness of the candidate sub-coding body corresponding to each of the candidate particles; According to the fitness of the candidate sub-encoder, the candidate particle with the smallest fitness is selected from the multiple candidate particles to obtain the target particle, wherein the material replenishment decision information corresponding to the candidate sub-encoder corresponding to the target particle is the target material replenishment decision information.

10. A service platform, comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps of the GA-PSO-based material ordering and replenishment decision-making method according to any one of claims 1 to 9.

Images