Production scheduling method, production scheduling system and electronic equipment

Abstract

The invention discloses a production scheduling method, a production scheduling system and electronic equipment, and belongs to the technical field of production management and scheduling. The production scheduling method comprises the steps that according to production requirements and a preset production correlation model, a fitness function and an initial production scheduling scheme set are determined, and the fitness function comprises at least two of the following production scheduling strategies: total delay time minimization, resource utilization rate maximization, production line load balancing and intermediate product minimization; at least two preset algorithms are fused, the initial production scheduling scheme set is optimized based on the fitness function, the optimal production scheduling scheme is obtained, and the preset algorithms comprise any one of a genetic algorithm, a simulated annealing algorithm and a tabu search algorithm. The method can solve the problems of insufficient processing of resource constraints, low production scheduling efficiency and incapability of effectively finding the optimal production scheduling scheme in related technologies.

Description

Technical Field

[0001] The present invention belongs to the technical field of production management and scheduling, and specifically relates to a production scheduling method, a production scheduling system and electronic equipment. Background Art

[0002] With the rapid development of automation, digitalization, and intelligent technologies in the nuclear chemical industry, numerous flexible automated production lines and intelligent production cells have been established, enabling high-efficiency, high-quality, and mass-produced nuclear chemical products. Due to the unique and complex nature of nuclear chemical conversion processes, traditional production scheduling methods struggle to meet the demands of efficient, flexible, and intelligent production management. In particular, nuclear chemical production involves complex factors such as material form, process routing, equipment utilization, and process cycle time. How to comprehensively consider these factors to achieve optimal scheduling and enhance production line automation and digitalization capabilities has become a major challenge in production management.

[0003] At present, a production scheduling method based on a single algorithm is usually adopted. This method cannot solve the complex problems of material form transformation specific to the nuclear chemical industry (such as the transformation of materials from liquid to solid in the dry water process), material tracking and resource scheduling. It has problems such as insufficient handling of resource constraints, low production scheduling efficiency, and inability to effectively find the optimal production scheduling plan. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address the above-mentioned shortcomings of the existing technology and provide a production scheduling method, a production scheduling system and electronic equipment that can balance multiple resource goals in the nuclear chemical production process and can quickly and effectively find the optimal scheduling plan.

[0005] In the first aspect, the present invention provides a production scheduling method, comprising: determining a fitness function and an initial scheduling plan set based on production demand and a preset production association model, wherein the fitness function includes at least two of the following scheduling strategies: minimizing total delay time, maximizing resource utilization, balancing production line load, and minimizing intermediate products; integrating at least two preset algorithms, and optimizing the initial scheduling plan set based on the fitness function to obtain an optimal scheduling plan, wherein the preset algorithm includes any one of the following: genetic algorithm, simulated annealing algorithm, and taboo search algorithm.

[0006] In some embodiments, after optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, the production scheduling method also includes: real-time monitoring of changes in production factors during the production process; dynamically adjusting the optimal production scheduling plan or re-optimizing to obtain a new optimal production scheduling plan based on changes in production factors.

[0007] In some embodiments, after optimizing the initial production scheduling set based on the fitness function to obtain the optimal production scheduling, the production scheduling method further includes: visually displaying the optimal production scheduling and real-time changes in production factors. The production scheduling method further includes: validating the optimal production scheduling using actual production data based on supervised learning or reinforcement learning, obtaining model parameter adjustment results, and adjusting the model parameters of the preset algorithm.

[0008] In some embodiments, at least two preset algorithms are integrated, and the initial production scheduling plan set is optimized based on the fitness function to obtain the optimal production scheduling plan, specifically including: based on the multi-stage optimization method or the weighted fusion method, integrating at least two of the genetic algorithm, the simulated annealing algorithm, and the taboo search algorithm, and optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan.

[0009] In some embodiments, a genetic algorithm, a simulated annealing algorithm, and a taboo search algorithm are integrated based on a multi-stage optimization method, and the initial production scheduling plan set is optimized based on a fitness function to obtain the optimal production scheduling plan, specifically including: performing a global search on the initial production scheduling plan set based on a genetic algorithm to generate a first production scheduling plan set; performing local optimization on the preset stages in the first production scheduling plan set based on a simulated annealing algorithm to obtain a second production scheduling plan set; and performing a refined search on the second production scheduling plan set based on a taboo search algorithm to obtain the optimal production scheduling plan.

[0010] In some embodiments, a genetic algorithm, a simulated annealing algorithm, and a taboo search algorithm are integrated based on a weighted fusion method, and the initial production scheduling plan set is optimized based on the fitness function to obtain the optimal production scheduling plan, specifically including: optimizing the initial production scheduling plan set based on the genetic algorithm, the simulated annealing algorithm, and the taboo search algorithm to obtain the corresponding production scheduling plan; calculating the first fitness function value of the production scheduling plan corresponding to the genetic algorithm, the second fitness function value of the production scheduling plan corresponding to the simulated annealing algorithm, and the third fitness function value of the production scheduling plan corresponding to the taboo search algorithm; weightedly integrating the first fitness function value, the second fitness function value, and the third fitness function value to obtain a comprehensive fitness value; and selecting the optimal production scheduling plan based on the comprehensive fitness value.

[0011] In some implementations, the production requirements include production orders and scheduling strategies.

[0012] According to production demand and the preset production association model, the fitness function and the initial production scheduling plan set are determined, specifically including: determining the fitness function according to the production scheduling strategy; determining the initial production scheduling plan set according to the preset mapping relationship between the production scheduling strategy and the algorithm, the production order, and the preset production association model.

[0013] In some embodiments, before determining the fitness function and the initial production scheduling plan set based on production demand and a preset production association model, the production scheduling method also includes: constructing an association model between production factors and production orders based on constraints and production process flow to obtain a preset production association model, wherein the constraints include resource constraints, process constraints, and material constraints, and the production factors include processes, equipment, raw materials, containers, and warehousing.

[0014] In a second aspect, the present invention further provides a production scheduling system, comprising: a determination module for determining a fitness function and a set of initial scheduling plans based on production demand and a preset production association model, wherein the fitness function includes at least two of the following: minimizing total delay time, maximizing resource utilization, balancing production line load, and minimizing intermediate products; and an optimization module, connected to the determination module, for integrating at least two preset algorithms and optimizing the set of initial scheduling plans based on the fitness function to obtain an optimal scheduling plan, wherein the preset algorithms include any of the following: a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm.

[0015] In a third aspect, the present invention further provides an electronic device comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the production scheduling method described in the first aspect by executing the computer instructions.

[0016] The present invention provides a production scheduling method, production scheduling system, and electronic equipment. Based on production demand and a production association model, the method determines a fitness function and an initial scheduling plan set containing multiple scheduling strategies, and integrates at least two algorithms to optimize the initial scheduling plan set to obtain the optimal scheduling plan. Since the optimization is performed using a fitness function based on multiple scheduling strategies, multiple resource objectives in the nuclear chemical production process can be weighed and complex scheduling problems in the nuclear chemical production process can be solved. Since the optimization is performed by integrating multiple preset algorithms, the advantages of different algorithms can be combined, thereby improving the optimization efficiency and effectively finding the optimal scheduling plan. Moreover, the optimization of the fitness function based on multiple scheduling strategies and the integration of multiple preset algorithms are interrelated and interact with each other, thereby improving the global optimization capability of the scheduling plan on the basis of achieving a multi-objective balance, thereby improving the scheduling efficiency and quality, and avoiding the situation where a single optimization algorithm cannot effectively find the optimal solution when weighing multiple objectives. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a production scheduling method according to embodiment 1 of the present invention;

[0018] Figure 2 This is a flowchart of a multi-algorithm fusion production scheduling logic according to embodiment 1 of the present invention;

[0019] Figure 3 This is a structural diagram of a production scheduling system according to Example 2 of the present invention. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0021] It should be understood that the specific embodiments and drawings described herein are only used to explain the present invention rather than to limit the present invention.

[0022] It is understood that, in the absence of conflict, the various embodiments of the present invention and the various features in the embodiments may be combined with each other.

[0023] It can be understood that, for the convenience of description, the drawings of the present invention only show parts related to the present invention, while parts unrelated to the present invention are not shown in the drawings.

[0024] It can be understood that each unit and module involved in the embodiments of the present invention may correspond to only one physical structure, or may be composed of multiple physical structures, or multiple units and modules may be integrated into one physical structure.

[0025] It will be understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of the present invention may occur in an order different from that marked in the drawings.

[0026] It is understood that the flowcharts and block diagrams of the present invention illustrate the possible architectures, functions, and operations of the systems, devices, equipment, and methods according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a unit, module, program segment, or code, which contains executable instructions for implementing the specified functions. Moreover, each box or combination of boxes in the block diagram and flowchart may be implemented using a hardware-based system that implements the specified functions, or may be implemented using a combination of hardware and computer instructions.

[0027] It can be understood that the units and modules involved in the embodiments of the present invention can be implemented by software or hardware. For example, the units and modules can be located in a processor.

[0028] Example 1:

[0029] like Figure 1 As shown, this embodiment provides a production scheduling method, which is suitable for scenarios where production scheduling plans are automatically generated, and is particularly suitable for application scenarios in nuclear chemical production. The production scheduling method can be executed by a computer and outputs a scheduling plan. The production scheduling method includes:

[0030] Step 101, determine the fitness function and the initial production scheduling plan set based on production demand and a preset production association model, wherein the fitness function includes at least two of the following scheduling strategies: minimizing total delay time, maximizing resource utilization, balancing production line load, and minimizing intermediate products.

[0031] Step 102: Fusing at least two preset algorithms, and optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, wherein the preset algorithms include any one of the following: genetic algorithm, simulated annealing algorithm, and taboo search algorithm.

[0032] Production demand includes production orders and scheduling strategies. An example of a production order is a collection of products with quantity and deadline requirements, which can also be understood as a production task. Scheduling strategies include, but are not limited to, minimizing total delay time, maximizing resource utilization, balancing production line load, minimizing intermediate products, prioritizing orders, optimizing energy consumption, maximizing efficiency, optimizing delivery value, and minimizing production cycle time. The integration of at least two pre-set algorithms can include a genetic algorithm and a simulated annealing algorithm, a genetic algorithm and a tabu search algorithm, a simulated annealing algorithm and a tabu search algorithm, or a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm. Genetic algorithms can effectively find global optimal solutions, simulated annealing algorithms are suitable for escaping local optima within a large solution space, and tabu search algorithms can avoid invalid solutions that have already been explored. It should be noted that pre-set algorithms are not limited to genetic algorithms, simulated annealing algorithms, and tabu search algorithms, but also include ant colony algorithms, particle swarm optimization algorithms, linear programming, integer programming, heuristic algorithms, and the like.

[0033] As one approach in this embodiment, optimization is performed based on the fitness functions of multiple scheduling strategies. This balances multiple resource objectives within the nuclear chemical production process and addresses complex scheduling issues within the process. By integrating multiple preset algorithms for optimization, the advantages of different algorithms can be combined, thereby improving optimization efficiency and effectively finding the optimal scheduling solution. Furthermore, the optimization based on the fitness functions of multiple scheduling strategies and the integration of multiple preset algorithms are interrelated and interactive. This improves the global optimization capabilities of the scheduling solution while achieving a multi-objective balance, thereby enhancing scheduling efficiency and quality. This avoids the situation where a single optimization algorithm cannot effectively find the optimal solution when balancing multiple objectives.

[0034] The following uses the nuclear chemical production process as an example to explain the business logic of its production process and production scheduling method:

[0035] The nuclear chemical production process consists of two major process sections:

[0036] (1) Raw material processing section (water method section) includes six different types of raw materials. The material forms (powder, liquid, gas) and management methods of different raw materials are very different. Each raw material corresponds to a dedicated raw material processing line. After the raw materials are processed, the same type of semi-finished products are generated and injected into the semi-finished product intermediate warehouse through intermediate storage tanks and pipelines.

[0037] (2) The processing section from semi-finished products to finished products (dry process section) is a fixed raw material formula, fixed production equipment production line, and fixed production and processing technology, which ultimately produces finished products of the same model, which are circulated through relevant containers and packaged / canned into the finished product warehouse.

[0038] Production scheduling based on this production process requires integrating the typical nuclear chemical conversion processes described above. By comprehensively considering constraints such as process, production capacity, tooling, equipment, manpower, shifts, work calendars, molds, containers, storage locations, and processing batches, it primarily addresses multiple scheduling strategies, including "accurately predicting delivery capacity and optimizing process production and material supply under limited production capacity." This embodiment is limited to the production scheduling of specialized materials in the nuclear chemical industry, characterized by a fixed production plan, fixed processing flow, fixed production line equipment, and a single product specification. Based on the various material forms (powder, liquid, solid) and various process flows (water, dry, multi-input, multi-output, etc.), combined with on-site production equipment (storage tanks, pipelines, liquid level metering, energy consumption, and environmental monitoring), comprehensive planning and scheduling, as well as task scheduling rule calculation, are implemented to support improved production efficiency in the nuclear chemical industry. Furthermore, because product processing involves multiple materials, multiple forms, and states, particularly the circulation of chemical powders and liquids through multiple storage tanks and pipelines, material metering, batch tracking, and component tracking require specialized data collection devices. Therefore, the key point of constructing a production association model is that the state changes of multiple materials caused by the progress of the processing technology and the constraint logic between upstream and downstream materials need to be given priority consideration. For example, multiple storage tanks transport materials to one storage tank, and semi-finished products are consumed as raw materials by another storage tank. In the process of outputting the final finished product, the constraint logic between materials at all levels needs to be constrained by BOM (Bill of Materials) and process effectiveness. Therefore, in the production scheduling method of this embodiment, the focus will be on developing intelligent algorithms and models based on constraint theory, training and knowledge extraction, etc., and integrating scheduling strategies such as maximizing production capacity, maximizing benefits, minimizing intermediate products, optimizing energy consumption, comprehensively optimizing delivery time, minimizing total delay time, maximizing resource utilization, and balancing production line load into the APS (Advanced Planning and Scheduling) scheduling module, and using machine learning, artificial intelligence and other means to train and optimize the model algorithm, providing basic support for precise scheduling analysis based on the characteristics of nuclear industry chemical production and the needs of planning management. Specifically, such as Figure 2 As shown, the production scheduling method of this embodiment is a scheduling logic based on the fusion of multiple algorithms. It receives production orders, order priorities, and economic insertion tasks, generates planning and scheduling elements and constraints, and then obtains planning and scheduling strategies and algorithm packaging through algorithm selection. It then performs knowledge iteration based on artificial intelligence technology, which includes artificial intelligence AI and machine learning ML. The optimal scheduling plan is then planned and released, and plan intervention and adjustment, production operation, and status are performed during the plan release stage.

[0039] In some embodiments, after optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, the production scheduling method also includes: real-time monitoring of changes in production factors during the production process; dynamically adjusting the optimal production scheduling plan or re-optimizing to obtain a new optimal production scheduling plan based on changes in production factors.

[0040] Among them, production factors include processes, equipment, raw materials, containers, and warehousing. Real-time monitoring of changes in production factors during the production process, an example of which is real-time monitoring of the actual operating time of the equipment, the current inventory status, etc. Dynamic adjustment of the optimal production schedule can be understood as incremental adjustment, that is, only adjusting the affected part of the production schedule on the basis of the optimal production schedule, rather than recalculating the entire production schedule. Re-optimization to obtain a new optimal production schedule can be understood as a global adjustment, that is, when the impact caused by the change in production factors is large, it is necessary to re-schedule the global production through an optimization algorithm. It should be noted that there may be various complex situations such as demand changes, equipment failures, raw material shortages, process adjustments, etc. during the production process. Therefore, the production scheduling method of this embodiment needs to be linked with the on-site equipment IOT (Internet of Things) data acquisition system and MES (Manufacturing Execution System) to realize the dynamic scheduling function. During the production process, by real-time monitoring of production conditions and proactively adjusting production plans based on actual demand changes, abnormal work orders and tasks can be paused, rerouted, repaired, and restarted to ensure the continuity and stability of the production process, thereby resolving the volatile planned orders in the nuclear chemical industry.

[0041] As one approach of this embodiment, by real-time monitoring of changes in production factors during the production process and dynamically adjusting the optimal production scheduling plan based on these changes, production flexibility can be enhanced. This is primarily reflected in the following two aspects: Because of real-time monitoring and dynamic adjustment, it has real-time data collection and monitoring capabilities, enabling timely understanding of information such as production progress, equipment status, and raw material supply, providing a panoramic view of the production process. For example, the visual interface is used to display information such as the production scheduling sequence of the process, resource usage, and production bottlenecks. Common charts include Gantt charts and equipment utilization heat maps. In the event of an emergency in production (such as equipment failure, staff shortages, raw material delays, etc.), it can respond quickly and automatically make dynamic adjustments to ensure production continuity and flexibility. This embodiment also has the function of intelligently responding to sudden changes. By quickly processing and feedback real-time data, and based on historical experience and real-time monitoring information, it can predict and respond to possible production fluctuations and sudden changes, thereby ensuring that production is not affected.

[0042] In some embodiments, after optimizing a set of initial production scheduling plans based on a fitness function to obtain an optimal production scheduling plan, the production scheduling method further includes visually displaying the optimal production scheduling plan and real-time changes in production factors. This visual display of the optimal production scheduling plan and real-time changes in production factors provides an intuitive visualization interface. Production scheduling results can be displayed through a graphical interface, allowing users to view the production schedule in real time and perform operations and adjustments as needed. This visualization not only helps production managers understand the overall production progress but also helps quickly identify and address unexpected issues.

[0043] In some embodiments, the production scheduling method further includes: based on supervised learning or reinforcement learning, using actual production data to verify the optimal scheduling plan, and obtaining model parameter adjustment results to adjust the model parameters of the preset algorithm.

[0044] As one approach in this embodiment, adjusting the model parameters of a preset algorithm based on supervised learning or reinforcement learning can optimize the performance of the preset algorithm, thereby improving scheduling efficiency and the accuracy of the scheduling plan. For example, by leveraging big data and artificial intelligence technologies to repeatedly solve multiple preset algorithms for multiple orders, and using actual production data as a verification method to obtain the optimal scheduling plan and recommend it to the user, the organic integration of these three makes the scheduling process more comprehensive and accurate.

[0045] In some embodiments, at least two preset algorithms are integrated, and the initial production scheduling plan set is optimized based on the fitness function to obtain the optimal production scheduling plan, specifically including: based on the multi-stage optimization method or the weighted fusion method, integrating at least two of the genetic algorithm, the simulated annealing algorithm, and the taboo search algorithm, and optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan.

[0046] To fully leverage the advantages of each preset algorithm to solve the scheduling problem of balancing multiple scheduling strategies, a multi-stage optimization method can be used to integrate at least two of the genetic algorithm, simulated annealing algorithm, and tabu search algorithm. The initial set of scheduling plans can be optimized based on a fitness function to obtain the optimal scheduling plan. A weighted fusion method can also be used to integrate at least two of the genetic algorithm, simulated annealing algorithm, and tabu search algorithm. A hybrid search strategy can also be used to integrate at least two of the genetic algorithm, simulated annealing algorithm, and tabu search algorithm. By organically integrating at least two preset algorithms through a multi-stage optimization method, a weighted fusion method, or a hybrid search strategy, the advantages of different preset algorithms can be combined, making the scheduling process more comprehensive and the scheduling plan more accurate.

[0047] The following describes the application of the genetic algorithm, simulated annealing algorithm, and taboo search algorithm in the nuclear chemical industry water drying process in this embodiment:

[0048] (1) Genetic algorithms search for the global optimal solution by simulating the natural selection mechanism (selection, crossover, mutation). Each production schedule is represented as a "chromosome", which includes the order of processes, the allocation of equipment (machines), etc. The quality of each solution is evaluated by the fitness function value (such as total delay time or resource utilization). Specifically, chromosome encoding: each "individual" represents a production schedule. The chromosome can be represented by an array with a length equal to the number of processes, and each position represents a process. Each gene of the chromosome can represent a process step (such as reaction, separation, heat exchange, etc.), and the position of each gene represents the order of the process. For example, for three processes A, B, and C, a possible chromosome is represented as: [A, C, B]. The fitness function can be the weighted sum of multiple production schedule strategies. The fitness function may include the following parts: Production delay time: the deviation of the completion time of each process from the scheduled time. Resource utilization: calculate the frequency of use of each device in the entire production process. The fitness function should comprehensively consider production efficiency, resource utilization, cost, equipment utilization and other scheduling strategies, usually implemented through weighted summation. The fitness function example is:

[0049] [F = w_1\times\text{ProductionEfficiency}+w_2\times\text{ResourceUtilization}+w_3\times\text{Cost}+w_4\times\text{Equipment Utilization}], where \times\text{ProductionEfficiency} refers to production efficiency, \times\text{ResourceUtilization} refers to resource utilization, \times\text{Cost} refers to cost, and \times\text{Equipment Utilization} refers to equipment utilization. w_1, w_2, w_3, and w_4 are weight coefficients, where w_1+w_2+w_3+w_4=1. The weight coefficients can be adjusted according to actual needs. By adjusting the weight coefficients, the balance between different scheduling strategies can be flexibly optimized. In selection, crossover, and mutation, selection refers to selecting individuals with better fitness as parents to generate the next generation. Crossover refers to randomly selecting two individuals to exchange some information (for example, swapping the order of processes). Mutation refers to randomly changing the production order of certain individuals to increase the diversity of the search. Termination condition: When a certain number of generations is reached or the fitness reaches a threshold, the algorithm stops and gradually finds the optimal production plan through multiple generations of evolution.

[0050] (2) The simulated annealing algorithm simulates the solid annealing process in physics, starting from a random solution, and gradually reducing the "temperature" to reduce the search space, and finally converges to the global optimal solution. The simulated annealing algorithm can help find the global optimal solution, especially when dealing with complex problems with multiple local optimal solutions. For the optimization of multiple scheduling strategies in the dry water process, simulated annealing can be implemented in the following ways: State space: Each "state" represents a specific production scheduling plan. The state space is very large and involves multiple factors such as production tasks, process sequence, equipment allocation, etc. Neighborhood structure: Starting from the current plan, generate neighborhood solutions (for example, adjust the order of two processes, exchange the use of equipment, etc.). Each neighborhood solution may differ in production efficiency, cost, etc. Acceptance criterion: The probability of accepting a worse solution is higher at the beginning, which can avoid falling into the local optimal solution. As the algorithm iterates, the temperature is gradually reduced, and finally converges to the global optimal solution. Fitness function: Find a suitable balance point between multiple scheduling strategies, taking into account factors such as production delay, resource utilization, and equipment load.

[0051] (3) The taboo search algorithm maintains a "tabu table" to avoid returning to solutions that have already been searched during the search process, thereby improving search efficiency. The taboo search algorithm maintains a taboo table during the search process to avoid repeated searches for solutions that have already been explored, thereby speeding up the search and avoiding falling into local optimal solutions. Its application in the dry water process includes: Neighborhood generation: making small-scale adjustments to the current production schedule, such as exchanging the execution order of two processes, reallocating equipment, etc. Among them, processes in different spatial areas of the same product or processes that have no influence on each other are parallel processes. For parallel processes, the execution order can be adjusted. Taboo table: records the search history within a certain number of times to avoid returning to solutions that have already been visited and increase the diversity of the search. Termination condition: When the quality of the solution is no longer significantly improved, the algorithm stops and outputs the optimal production schedule.

[0052] In some embodiments, a multi-stage optimization method is used to integrate a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm, and an initial production scheduling set is optimized based on a fitness function to obtain an optimal production scheduling plan, specifically including steps 201 to 203:

[0053] Step 201: Perform a global search on the initial production scheduling plan set based on a genetic algorithm to generate a first production scheduling plan set.

[0054] Step 202 : Locally optimize the preset stages in the first production scheduling plan set based on a simulated annealing algorithm to obtain a second production scheduling plan set.

[0055] Step 203: Perform a refined search on the second production scheduling plan set based on the tabu search algorithm to obtain the optimal production scheduling plan.

[0056] The order of the preset algorithms in the multi-stage optimization is not fixed, and is not limited to three preset algorithms for multi-stage optimization, but can also be other preset algorithms.

[0057] In some embodiments, the goal of step 201 is to generate multiple possible production scheduling plans to avoid falling into a local optimum. Step 201 includes S11-S15:

[0058] S11, initializing the population. For example, the initial production scheduling plan includes three different production scheduling plans.

[0059] S12, fitness evaluation. Evaluate each of the three production scheduling plans and calculate its fitness function value (such as the comprehensive score of production cycle, equipment load, etc.).

[0060] S13, selection, crossover and mutation: select the production scheduling plan with a fitness function value greater than the first threshold, perform crossover and mutation operations, and generate a new candidate solution.

[0061] S14, Iteration. Repeat the selection, crossover, and mutation operations. After multiple generations of evolution, multiple possible production scheduling plans are obtained.

[0062] S15: Output a first production scheduling plan set. Multiple candidate production scheduling plans constitute the first production scheduling plan set, which represent different global search results.

[0063] For example, consider the following production scheduling problem: processes A, B, and C, and two pieces of equipment E1 and E2. Process times: Process A takes 5 hours and uses equipment E1 and E2; Process B takes 3 hours and uses equipment E1; Process C takes 4 hours and uses equipment E2. The scheduling strategy is to minimize delays and maximize equipment utilization.

[0064] S11, initialize the population. The initial production schedule includes the following individuals:

[0065] Individual 1: [A, B, C],

[0066] Individual 2: [B, A, C],

[0067] Individual 3: [C, A, B],

[0068] S12, evaluate fitness (calculate each individual's delay time and equipment utilization rate). For each individual, first calculate the delay time and equipment utilization rate, and then evaluate its quality based on the fitness function value.

[0069] Individual 1: Calculation of delay time for [A, B, C]: Process A starts at time 0 and ends in 5 hours (equipment E1 and E2 are used simultaneously). Process B requires equipment E1, but equipment E1 is occupied by process A. Therefore, process B needs to wait until process A is completed, i.e., process B starts at 5 hours and ends in 8 hours. Process C starts at time 0, but equipment E2 is occupied by process A. Therefore, process C waits for equipment E2. Process C starts at 5 hours and ends in 9 hours. Delay time: Process B waits for E1: delay of 5 hours (process B has to wait from 0 to 5 before it can start). Process C waits for E2: delay of 5 hours (process C has to wait from 0 to 5 before it can start). Therefore, the total delay time for individual 1 = 5 + 5 = 10 hours. Calculation of equipment utilization for individual 1: E1: usage 5 hours (process A) + 3 hours (process B) = 8 hours, E2: usage 5 hours (process A) + 4 hours (process C) = 9 hours. Total equipment utilization: E1 utilization

[0070] =8 / (5+3+4)=8 / 12=0.67, E2 utilization rate=9 / 12=0.75. Therefore, the total equipment utilization rate of individual 1 = (0.67+0.75) / 2=0.71. The fitness function of individual 1 is:

[0071] [F = w_1\times\frac{1}{\text{delay time}}+w_2\times\text{equipment utilization rate}], assuming that (w_1=0.6) and (w_2=0.4) are selected.

[0072] [F=0.6\times\frac{1}{10}+0.4\times0.71=0.06+0.284=0.344].

[0073] Individual 2: Calculation of delay time for [B, A, C]: Process B starts at time 0 and ends in 3 hours (equipment E1). Process A requires equipment E1, but equipment E1 is occupied by process B, so process A waits for E1. Process A starts at 3 hours and ends in 8 hours. Process C starts at time 0 and ends in 4 hours (equipment E2). Delay time: Process A waits for E1: Delay of 3 hours (process A has to wait from 0 to 3 before it can start), so the total delay time of individual 2 = 3 hours. Calculation of equipment utilization: E1: 3 hours (process B) + 5 hours (process A) = 8 hours, E2: 4 hours (process C) = 4 hours, total equipment utilization: E1 utilization = 8 / 12 = 0.67, E2 utilization = 4 / 12 = 0.33, so the total equipment utilization of individual 2

[0074] =(0.67+0.33) / 2=0.5. The fitness function value of individual 2 is:

[0075] [F=0.6\times\frac{1}{3}+0.4\times0.5=0.2+0.2=0.4].

[0076] Individual 3: Calculation of delay time for [C, A, B]: Process C starts at time 0 and ends at 4 hours (equipment E2). Process A requires equipment E1 and E2. Equipment E1 is idle during the execution of process C, but E2 is occupied by process C. Therefore, process A waits for equipment E2. Process A starts at 4 hours and ends at 9 hours. Process B starts at time 9 and ends at 12 hours (equipment E1). Delay time: Process A waits for E2: delay of 4 hours, process B has no delay. The total delay time of individual 3 = 4 hours. Calculation of equipment utilization: E1: uses 4 hours (process A) + 3 hours (process B) = 7 hours, E2: uses 4 hours (process C) + 5 hours (process A) = 9 hours, total equipment utilization: E1 utilization = 7 / 12 = 0.58, E2 utilization =

[0077] =9 / 12=0.75, the total equipment utilization rate of individual 3 = (0.58+0.75) / 2=0.665. The fitness function value is:

[0078] [F=0.6\times\frac{1}{4}+0.4\times0.665=0.15+0.266=0.416].

[0079] S13, selection operation. Based on fitness, the individual with the highest fitness is selected to enter the next generation. In this generation, individual 3 ([C, A, B]) has the highest fitness (0.416), so it will become one of the parents.

[0080] S14, crossover and mutation operations. Two individuals with higher fitness are selected for crossover (e.g., individuals 2 and 3), and new individuals may be generated through mutation. Through these operations, a new scheduling solution will be generated in the next generation.

[0081] S14, Iteration. The termination condition for a genetic algorithm is usually reaching a predetermined number of generations or reaching fitness convergence. If fitness does not improve significantly after multiple generations of evolution, the algorithm can be terminated and the current optimal scheduling solution can be output.

[0082] S15, output the final result (i.e., the optimal production scheduling plan). The optimal production scheduling plan (based on the fitness evaluation results): process sequence: [C, A, B] delay time = 4 hours, equipment utilization rate = 0.665. Compared with other plans (such as individual 1 and individual 2), this production scheduling plan has lower delay time and higher equipment utilization rate, indicating that it is a relatively optimal plan.

[0083] In some embodiments, the goal of step 202 is to perform local optimization based on multiple production scheduling plans generated by the genetic algorithm to avoid local optimality. Step 202 includes S21-S25:

[0084] S21, selecting a production scheduling plan: Select one from the first production scheduling plan set output by the genetic algorithm as the current production scheduling plan (ie, the current solution).

[0085] S22: Generate a neighborhood solution. Make minor adjustments to the current production schedule (e.g., swap the order of two tasks, change the start time of a task, etc.) to generate a new neighborhood solution.

[0086] S23, the Metropolis criterion. It determines whether to accept a new solution based on the fitness of neighboring solutions compared to the fitness of the current solution. If the neighboring solution is superior, it is accepted; if it is inferior, it is accepted with a certain probability, which decreases gradually with temperature. Temperature decay: Gradually lowering the temperature reduces the probability of accepting inferior solutions.

[0087] S24, iterative termination condition: When the temperature drops to the minimum or reaches the predetermined number of iterations, the optimization process is stopped.

[0088] S25: Output the second production scheduling plan set. The production scheduling plan optimized by simulated annealing is the local optimal solution.

[0089] As another example, if the genetic algorithm optimizes the initial production schedule set and obtains the first production schedule (A, B, C, D); the simulated annealing algorithm only performs local optimization on the (A, B) water method stage to obtain (B, A), so the second production schedule is (B, A, C, D); the taboo search algorithm continues to optimize, and may eventually obtain (B, A, D, C) as the optimal production schedule.

[0090] In some embodiments, the goal of step 203 is to further refine the local optimal solution through tabu search, avoid falling into the local optimal solution, and enhance the diversity of solutions. Step 203 includes S31-S36:

[0091] S31, selecting a production scheduling plan. Select one from the second production scheduling plan set obtained by simulated annealing optimization as an initial solution.

[0092] S32: Generate a neighborhood solution and make some local adjustments to the current solution (such as changing the task scheduling order, adjusting the task time window, etc.).

[0093] S33, taboo table update. The solutions that have been visited are recorded in the taboo table to prevent returning to the solutions that have been tried previously.

[0094] S34, neighborhood selection. Select the optimal neighborhood solution based on the fitness value. If the improvement conditions are met, update the current solution.

[0095] S35, Iteration. Continue searching until the maximum number of iterations is reached or the quality of the solution converges.

[0096] S36, outputs the optimal production scheduling plan, which is the result of further optimization through taboo search.

[0097] In step 201, the genetic algorithm performs a global search on the initial set of production scheduling solutions through selection, crossover, and mutation operations, which can explore a wide area of the solution space and avoid falling into local optimality. In step 202, the simulated annealing algorithm performs local optimization on the solution generated by the genetic algorithm, further optimizing the quality of the solution by probabilistically accepting inferior solutions while avoiding premature convergence. In step 203, the taboo search algorithm refines the solution through taboo tables and neighborhood searches to further improve the accuracy of the solution. Therefore, the multi-stage optimization method combines the advantages of global search and local optimization, and can efficiently find the global optimal or near-global optimal production scheduling solution in the solution space. This embodiment integrates multiple preset algorithms through a staged optimization method, that is, using different algorithms in stages to gradually find the optimal production scheduling solution. This can significantly improve the optimization effect of the production scheduling results, achieve more efficient, flexible, and intelligent production scheduling, and is also suitable for complex large-scale production scheduling problems, especially complex scheduling scenarios such as nuclear chemical industry.

[0098] In some implementations, a weighted fusion method is used to integrate a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm, and an initial production scheduling plan set is optimized based on a fitness function to obtain an optimal production scheduling plan, specifically including steps 301 to 304:

[0099] Step 301 : Optimize the initial production scheduling plan set based on the genetic algorithm, simulated annealing algorithm, and tabu search algorithm to obtain the corresponding production scheduling plan.

[0100] Step 302 , respectively calculate the first fitness function value of the production scheduling plan corresponding to the genetic algorithm, the second fitness function value of the production scheduling plan corresponding to the simulated annealing algorithm, and the third fitness function value of the production scheduling plan corresponding to the tabu search algorithm.

[0101] Step 303: Perform weighted fusion on the first fitness function value, the second fitness function value, and the third fitness function value to obtain a comprehensive fitness value.

[0102] Step 304: Select the optimal production scheduling plan based on the comprehensive fitness value.

[0103] The core idea of the weighted fusion method is to combine the fitness function values of each pre-set algorithm to form a final comprehensive fitness value. Weight coefficients are then used to balance the contributions of different pre-set algorithms to obtain the optimal production schedule. This, in turn, integrates the advantages of different pre-set algorithms to find the optimal schedule. For example, for a production schedule, the fitness function values of the genetic algorithm, simulated annealing, and tabu search can be calculated separately. These results are then comprehensively considered to select the optimal schedule. An example weighted fusion formula is: [F = m_1\times F_{\text{GA}} + m_2\times F_{\text{SA}} + m_3\times F_{\text{TS}}]. Where: (F_{\text{GA}}) refers to the fitness function value of the schedule generated by the genetic algorithm; (F_{\text{SA}}) refers to the fitness function value of the schedule optimized by the simulated annealing algorithm; and (F_{\text{TS}}) refers to the fitness function value of the schedule further optimized by the tabu search algorithm. m_1, m_2, and m_3 are weight coefficients of each algorithm, and m_1+m_2+m_3=1.

[0104] For example, if there is a production scheduling requirement, the scheduling strategy is mainly to minimize the production cycle. To simplify the calculation process of the example, the weight coefficients of other scheduling strategies are assumed to be small enough and thus negligible.

[0105] Calculate the fitness function values for the production scheduling plans output by the genetic algorithm, simulated annealing algorithm, and tabu search algorithm. For example, if the first fitness function value F_{\text{GA}} = 120, the production cycle is 120 days; the second fitness function value F_{\text{SA}} = 110, the production cycle is 110 days; and the third fitness function value F_{\text{TS}} = 115, the production cycle is 115 days. Based on experimentation or experience, set appropriate weighting coefficients. For example, if m_1 = 0.4, the genetic algorithm contributes significantly; m_2 = 0.3, the simulated annealing algorithm contributes moderately; and m_3 = 0.3, the tabu search algorithm contributes moderately. A weighted fusion calculation yields a comprehensive fitness value of 115.5: [F = 0.4\times120 + 0.3\times110 + 0.3\times115 = 48 + 33 + 34.5 = 115.5]. Select the optimal plan: Select the production schedule with the smallest fitness function value. Assuming that the production cycle of a certain production schedule is 115.5 days, this production schedule is the optimal production schedule.

[0106] By combining the strengths of three algorithms, the weighted fusion method strikes a balance between global search, local optimization, and refined search, generating a more optimal scheduling solution. Specifically, the weighted fusion method balances the contributions of different algorithms using weight coefficients, adapting to different scheduling problems and constraints (such as resource limits, time windows, and priorities).

[0107] In some implementations, a hybrid search strategy is used to integrate a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm, and to optimize the initial production scheduling plan set based on a fitness function to obtain the optimal production scheduling plan. The hybrid search strategy refers to combining the fitness function values of different algorithms through weighted average or combination strategies, thereby integrating the advantages of different preset algorithms to achieve better optimization results. Specifically, it includes steps 401 to 404:

[0108] Step 401: Initial plan selection: First, select one of the multiple production scheduling plans generated by optimizing the initial production scheduling plan using a genetic algorithm as the initial plan.

[0109] Step 402: Simulated annealing algorithm optimization: Perform simulated annealing on the initial plan to obtain an optimized production scheduling plan.

[0110] Step 403: Refine the tabu search algorithm. Perform a tabu search based on the simulated annealing optimization results to further refine the production scheduling plan.

[0111] Step 404: Weighted fusion: From the multiple candidate solutions obtained in the tabu search phase, a weighted fusion strategy is used to select the final optimal solution as the optimal production scheduling solution.

[0112] Example: Suppose we have the following three optimization results: A scheduling plan optimized by a genetic algorithm (GA) results in a production cycle of 120 days. A scheduling plan optimized by simulated annealing (SA) results in a production cycle of 110 days. A scheduling plan further optimized by tabu search (TS) results in a production cycle of 115 days. Through weighted fusion calculation, the final optimized result is 115.5 days.

[0113] The hybrid search strategy combines the advantages of multi-stage optimization and weighted fusion methods, achieving a balance between global and local search, diversity and centralization, and robustness and stability. The weighted fusion strategy further improves solution quality. This strategy can achieve better optimization results in complex optimization problems and has strong practicality and broad application prospects.

[0114] In some embodiments, the production demand includes a production order and a scheduling strategy. Determining a fitness function and an initial set of scheduling solutions based on the production demand and a preset production association model specifically includes: determining the fitness function based on the scheduling strategy; and determining the initial set of scheduling solutions based on a preset mapping relationship between the scheduling strategy and the algorithm, the production order, and the preset production association model.

[0115] Among them, the preset mapping relationship between the production scheduling strategy and the algorithm is shown in Table 1.

[0116] Table 1 Mapping relationship between scheduling strategy and algorithm

[0117]

[0118]

[0119] For example, if the production scheduling strategy includes energy consumption optimization, then according to the mapping relationship in Table 1, the dynamic programming algorithm can be used first, combined with the production order and the preset production association model, to determine the initial production scheduling plan set, and then a variety of preset algorithms can be integrated to optimize the initial production scheduling plan set to obtain the optimal production scheduling plan. It should be noted that the initial production scheduling plan set can be generated by calling a matching algorithm based on the preset mapping relationship between the production scheduling strategy and the algorithm, or it can be generated randomly. In this embodiment, first determining the initial production scheduling plan set based on the preset mapping relationship between the production scheduling strategy and the algorithm has the following effects:

[0120] (1) Higher optimization efficiency: Since the initial production scheduling plan is relatively reasonable, optimization algorithms such as genetic algorithms, simulated annealing algorithms, and taboo search algorithms can converge to the optimal solution more quickly during the optimization process, reducing the search space and improving optimization efficiency.

[0121] (2) Combining global optimization with local optimization to obtain an effective production scheduling plan: The algorithms in Table 1 are usually used for local optimization (such as order priority sorting, equipment capacity maximization, etc.), while global optimization algorithms such as genetic algorithms, simulated annealing algorithms, and taboo search algorithms can further optimize the production scheduling plan on a global scale. This combination can take into account both local and global optimization needs and obtain a better production scheduling plan.

[0122] (3) Strong adaptability: The initial production scheduling plan has taken into account a variety of production scheduling strategy requirements and constraints. The optimization algorithm is further optimized on this basis, which can better adapt to the complex production environment and changing production scheduling needs.

[0123] (4) Reduce computational complexity: Since the initial production scheduling plan is already relatively reasonable, the optimization algorithm does not need to start searching from a completely random initial solution, which reduces computational complexity and saves computing resources.

[0124] In some embodiments, before determining the fitness function and the initial production scheduling plan set based on production demand and a preset production association model, the production scheduling method also includes: constructing an association model between production factors and production orders based on constraints and production process flow to obtain a preset production association model, wherein the constraints include resource constraints, process constraints, and material constraints, and the production factors include processes, equipment, raw materials, containers, and warehousing.

[0125] Among them, building a production association model includes:

[0126] Model production processes, equipment capabilities, process flows, and material inventory. Each process's duration (takt time), equipment capabilities, personnel hours, and material availability all need to be quantified. A production process typically consists of multiple steps, each requiring specific resources (e.g., equipment, personnel, and raw materials). These steps have specific durations and sequences.

[0127] Obtain constraints. Resource constraints include the number of devices and the maximum processing capacity of each device (maximum hourly output). Process constraints include dependencies between processes (some processes must be performed before or after others). Material constraints include formula-based constraints such as various raw materials and auxiliary materials (gas, metal powder, and chemical materials). Constraints also include personnel constraints, which include the work schedule of personnel involved in the production process.

[0128] The production scheduling method of this embodiment optimizes based on the fitness functions of multiple scheduling strategies. This method balances multiple resource objectives within the nuclear chemical production process and addresses complex scheduling issues within the process. By integrating multiple optimization algorithms for optimal search, it leverages the strengths of different algorithms, improving optimization efficiency and effectively finding the optimal scheduling solution. Furthermore, the fitness functions of multiple scheduling strategies and the integration of multiple optimization algorithms are interrelated and interactive. This improves the global optimization capabilities of the scheduling solution while achieving a multi-objective balance, thereby enhancing scheduling efficiency and quality. This avoids the inability to effectively find the optimal solution when using a single optimization algorithm to balance multiple objectives. Furthermore, higher-quality scheduling solutions are obtained through multi-stage optimization methods, weighted fusion methods, or hybrid search strategies that integrate genetic algorithms, simulated annealing algorithms, and tabu search algorithms. Furthermore, by generating an initial set of scheduling solutions based on a pre-defined mapping between scheduling strategies and algorithms, and then optimizing based on the integrated algorithm, the scheduling method achieves higher optimization efficiency, combines global and local optimization, enhances adaptability, and reduces computational complexity.

[0129] Example 2:

[0130] like Figure 3As shown, this embodiment provides a production scheduling system, including:

[0131] Determination module 31 is used to determine the fitness function and the initial production scheduling plan set based on production demand and a preset production association model, wherein the fitness function includes at least two of the following: minimizing total delay time, maximizing resource utilization, balancing production line load, and minimizing intermediate products.

[0132] The optimization module 32 is connected to the determination module 31 and is used to integrate at least two preset algorithms and optimize the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, wherein the preset algorithm includes any one of the following: genetic algorithm, simulated annealing algorithm, and taboo search algorithm.

[0133] In some embodiments, the production scheduling system further includes an adjustment module.

[0134] The adjustment module is used to monitor the changes in production factors during the production process in real time, and dynamically adjust the optimal production scheduling plan or re-optimize to obtain a new optimal production scheduling plan based on the changes in production factors.

[0135] In some embodiments, the production scheduling system further includes a display module.

[0136] The display module is connected to the optimization module and the adjustment module to visually display the optimal production scheduling plan and real-time changes in production factors.

[0137] In some embodiments, the production scheduling system further includes a parameter adjustment module.

[0138] The parameter adjustment module is connected to the optimization module and is used to verify the optimal production scheduling plan based on supervised learning or reinforcement learning using actual production data, obtain the model parameter adjustment results, and adjust the model parameters of the preset algorithm.

[0139] In some embodiments, the optimization module is also used to integrate at least two of the genetic algorithm, simulated annealing algorithm, and taboo search algorithm based on a multi-stage optimization method or a weighted fusion method, and optimize the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan.

[0140] In some embodiments, the optimization module is also used to integrate genetic algorithm, simulated annealing algorithm, taboo search algorithm based on multi-stage optimization method, and optimize the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan.

[0141] The optimization module is specifically used to perform a global search on the initial production scheduling plan set based on a genetic algorithm to generate a first production scheduling plan set, and to perform local optimization on the preset stages in the first production scheduling plan set based on a simulated annealing algorithm to obtain a second production scheduling plan set, and to perform a refined search on the second production scheduling plan set based on a taboo search algorithm to obtain the optimal production scheduling plan.

[0142] In some embodiments, the optimization module is further used to fuse a genetic algorithm, a simulated annealing algorithm, and a taboo search algorithm based on a weighted fusion method, and optimize the initial production scheduling plan set based on a fitness function to obtain an optimal production scheduling plan.

[0143] The optimization module is specifically used to optimize the initial production scheduling plan set based on the genetic algorithm, simulated annealing algorithm, and taboo search algorithm to obtain the corresponding production scheduling plan. It is also used to calculate the first fitness function value of the production scheduling plan corresponding to the genetic algorithm, the second fitness function value of the production scheduling plan corresponding to the simulated annealing algorithm, and the third fitness function value of the production scheduling plan corresponding to the taboo search algorithm, and to weightedly fuse the first fitness function value, the second fitness function value, and the third fitness function value to obtain a comprehensive fitness value, and to select the optimal production scheduling plan based on the comprehensive fitness value.

[0144] In some implementations, the production requirements include production orders and scheduling strategies.

[0145] The determination module is used to determine the fitness function according to the production scheduling strategy, and is also used to determine the initial production scheduling plan set according to the preset mapping relationship between the production scheduling strategy and the algorithm, the production order, and the preset production association model.

[0146] The production scheduling system of this embodiment involves optimizing multi-resource scheduling for dry-water hybrid production scenarios in the nuclear chemical industry, specifically researching a production scheduling algorithm based on the integration of multiple scenarios, multiple resources, and multiple algorithms. This system comprehensively considers multiple factors, including order priority, resource utilization, process flow and resource constraints, and multi-material form tracking, to achieve more efficient and flexible production scheduling. The algorithm in this embodiment integrates multiple sub-algorithms and combines them with optimization strategies such as AI for comprehensive decision-making, achieving automated and intelligent production scheduling. This effectively improves the accuracy and flexibility of production scheduling, optimizes resource allocation, and reduces production costs. The production scheduling system of this embodiment not only achieves the same benefits as Example 1 but also supports plant-wide production planning and process-level operation plan issuance. The algorithm in this embodiment supports three-level planning and scheduling functions: plant-level, production line-level, and equipment-level. From the compilation of plant-wide production plans to the issuance of individual process-level operation plans, it provides data support for all production links, ensuring the smooth execution of all tasks in the production process. Based on resource constraints and scheduling strategies for production scheduling, the system encapsulates different algorithm application modules and optimizes algorithms for different production modes (dry and wet). This allows planners to optimize production plans by selecting from a variety of algorithms. It can recommend the optimal scheduling plan while ensuring production progress and maximizing the use of limited production resources. The system also provides detailed scheduling management, encompassing not only production line scheduling but also team / personnel scheduling, material preparation and transfer scheduling, equipment inspection / startup / operation scheduling, energy supply scheduling, and warehouse operation scheduling. Each link can be precisely scheduled and monitored in real time within the system. Furthermore, the system can reduce costs. Specifically, it reduces ineffective downtime: Through precise scheduling and resource allocation, ineffective production stoppages and waiting times can be avoided. The optimized production plan minimizes idle time for equipment and workers, reduces production downtime caused by improper scheduling, and reduces operating costs. Rational resource allocation: Accurate scheduling of materials and equipment avoids over-purchasing and excess inventory, helping companies achieve efficient resource utilization and reduce inventory costs. Energy and material consumption optimization: Utilize intelligent scheduling and forecasting functions to reduce unnecessary energy and raw material waste and lower production costs while ensuring production efficiency. The system is also scalable and versatile. Specifically, cross-industry applicability: The algorithm in this embodiment is not only optimized in the nuclear industry, but other complex production environments, such as manufacturing, chemical, automotive, aerospace and other industries require integrated business optimization. Modular design: The system adopts a modular structure and can flexibly adjust and expand the functions of each module according to different production environments and needs. Continuous optimization capabilities: Through data accumulation and machine learning, it can continuously optimize its own decision-making model, continuously improve production efficiency and reduce costs during the application process.

[0147] Example 3:

[0148] This embodiment provides an electronic device, including: a memory and a processor, the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the production scheduling method described in Example 1 by executing the computer instructions.

[0149] It will be understood that the above embodiments are merely exemplary embodiments for illustrating the principles of the present invention, and the present invention is not limited thereto. Those skilled in the art will appreciate that various modifications and improvements can be made without departing from the spirit and substance of the present invention, and such modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A production scheduling method, characterized in that: include: Determine a fitness function and an initial production scheduling plan set based on production demand and a preset production association model, where the fitness function includes at least two of the following scheduling strategies: minimizing total delay time, maximizing resource utilization, balancing production line load, and minimizing intermediate products; At least two preset algorithms are integrated, and the initial production scheduling plan set is optimized based on the fitness function to obtain the optimal production scheduling plan, wherein the preset algorithms include any one of the following: genetic algorithm, simulated annealing algorithm, and taboo search algorithm.

2. The method according to claim 1, characterized in that After optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, the method further includes: Real-time monitoring of changes in production factors during the production process; According to the changes in the production factors, the optimal production scheduling plan is dynamically adjusted or a new optimal production scheduling plan is obtained by re-optimization.

3. The method according to claim 2, characterized in that After optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, the method further includes: Visually display the optimal production scheduling plan and real-time changes in production factors, The production scheduling method further includes: based on supervised learning or reinforcement learning, using actual production data to verify the optimal scheduling plan, obtaining model parameter adjustment results, and adjusting the model parameters of the preset algorithm.

4. The method according to claim 1, wherein The fusing of at least two preset algorithms and optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan specifically includes: Based on a multi-stage optimization method or a weighted fusion method, at least two of a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm are integrated, and based on the fitness function, the initial production scheduling plan set is optimized to obtain an optimal production scheduling plan.

5. The method according to claim 4, characterized in that The multi-stage optimization method integrates a genetic algorithm, a simulated annealing algorithm, and a tabu search algorithm, and optimizes the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, specifically including: Performing a global search on the initial production scheduling plan set based on a genetic algorithm to generate a first production scheduling plan set; Performing local optimization on the preset stages in the first production scheduling plan set based on a simulated annealing algorithm to obtain a second production scheduling plan set; The second production scheduling plan set is refined and searched based on the tabu search algorithm to obtain the optimal production scheduling plan.

6. The method according to claim 4, characterized in that The weighted fusion method is based on fusing the genetic algorithm, the simulated annealing algorithm, and the tabu search algorithm, and optimizing the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, specifically including: Optimizing the initial production scheduling plan set based on genetic algorithm, simulated annealing algorithm and tabu search algorithm to obtain corresponding production scheduling plans; Calculate the first fitness function value of the production scheduling plan corresponding to the genetic algorithm, the second fitness function value of the production scheduling plan corresponding to the simulated annealing algorithm, and the third fitness function value of the production scheduling plan corresponding to the tabu search algorithm respectively; Performing weighted fusion on the first fitness function value, the second fitness function value, and the third fitness function value to obtain a comprehensive fitness value; An optimal production scheduling plan is selected based on the comprehensive fitness value.

7. The method according to claim 1, characterized in that Production demand includes production orders and scheduling strategies. Determining the fitness function and the initial production scheduling plan set based on production demand and a preset production association model specifically includes: Determine the fitness function based on the production scheduling strategy; An initial production scheduling plan set is determined based on the preset mapping relationship between the production scheduling strategy and the algorithm, the production order, and the preset production association model.

8. The method according to claim 1, characterized in that Before determining the fitness function and the initial production scheduling plan set according to the production demand and the preset production association model, the method further includes: Based on the constraints and production process flow, a correlation model between production factors and production orders is constructed to obtain a preset production correlation model, where the constraints include resource constraints, process constraints, and material constraints, and the production factors include processes, equipment, raw materials, containers, and warehousing.

9. A production scheduling system, characterized in that: include: The determination module is used to determine the fitness function and the initial production scheduling plan set based on production demand and a preset production association model, wherein the fitness function includes at least two of the following: minimizing total delay time, maximizing resource utilization, balancing production line load, and minimizing intermediate products. The optimization module is connected to the determination module and is used to integrate at least two preset algorithms and optimize the initial production scheduling plan set based on the fitness function to obtain the optimal production scheduling plan, wherein the preset algorithm includes any one of the following: genetic algorithm, simulated annealing algorithm, and taboo search algorithm.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes a production scheduling method according to any one of claims 1 to 8 by executing the computer instructions.

Images