Dynamic self-adaptive production scheduling method and system

Abstract

The invention discloses a dynamic self-adaptive production scheduling method and system, and the method comprises the steps: collecting and sorting data needed by production scheduling, and carrying out the preprocessing of the collected data, and obtaining the preprocessed data; the preprocessed data are screened and extracted to obtain effective entries, and the effective entries comprise work orders to be scheduled, production equipment states, product batch information and product procedures; setting a matching rule, matching the effective items based on the matching rule, and screening out batches meeting production conditions; collecting production efficiency data of the production equipment, and selecting optimal production equipment for the batches meeting the production conditions based on the production efficiency data; and collecting the current state of the equipment and the current state of the batch in real time, and generating an optimal scheduling plan based on the optimal production equipment by adopting a dynamic adaptive algorithm. Through the synergistic effect of the steps, an efficient and flexible production scheduling system is formed, production changes can be dynamically responded, resource configuration is optimized, and the production efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of production scheduling, and in particular relates to a dynamic and adaptive production scheduling method and system. Background Art

[0002] In the manufacturing industry, production scheduling is a key link to ensure that the production process is efficient and completed on time. Traditional production scheduling systems usually rely on static rules and sequential scheduling methods, which often lead to insufficient resource utilization, extended production cycles and unguaranteed delivery dates in the face of rapidly changing market demands and complex production environments.

[0003] In the existing production scheduling methods, due to the lack of overall optimization of the production process, equipment and human resources are not fully utilized, and when faced with urgent orders or production changes, the existing system is difficult to adapt quickly, resulting in the inability to adjust the production plan in a timely manner. At the same time, the traditional methods lack the ability to monitor in real time and adjust flexibly, which may cause the production line to stop and wait, reducing overall efficiency. Many systems fail to effectively analyze and use historical production data to improve scheduling algorithms and predict production cycles.

[0004] Therefore, there is an urgent need for a method and system that can dynamically adapt to changes in the production environment and use historical data to optimize scheduling decisions in order to improve resource utilization and production efficiency. Summary of the invention

[0005] The present invention aims to solve the deficiencies of the prior art and provides the following solutions:

[0006] A dynamic and adaptive production scheduling method comprises the following steps:

[0007] Collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data;

[0008] Screening and extracting the preprocessed data to obtain valid entries, the valid entries including: work orders to be scheduled, production equipment status, product batch information and product process;

[0009] Setting matching rules, matching the valid entries based on the matching rules, and screening out batches that meet production conditions;

[0010] Collecting production efficiency data of the production equipment, and selecting the optimal production equipment for the batch meeting the production conditions based on the production efficiency data;

[0011] The current status of the equipment and the current status of the batch are collected in real time, and a dynamic adaptive algorithm is used to generate an optimal scheduling plan based on the optimal production equipment.

[0012] Preferably, the data required for production scheduling includes: work order information, equipment capacity, product process efficiency and production history records;

[0013] The preprocessing method includes: using data cleaning technology to process the collected data, removing locked data, data with incorrect status, duplicate data, and data field value incorrect data, to obtain the preprocessed data.

[0014] Preferably, the matching method includes:

[0015] Setting a weight factor and a score for each of the valid entries;

[0016] Calculating a weighted score for each of the valid entries based on the weight factor and the score;

[0017] A minimum acceptance score threshold is set, compared with the weighted score, and the valid entries with weighted scores higher than the score threshold are selected as the batches that meet the production conditions.

[0018] Preferably, the method for generating the optimal scheduling plan includes:

[0019] Collect the current status of equipment and batches in real time, analyze historical production records, and obtain analysis results;

[0020] Processing the analysis results using a genetic algorithm to obtain an optimized scheduling solution based on the optimal production equipment;

[0021] The optimized scheduling scheme is dynamically adjusted using a reinforcement learning algorithm to obtain the optimal scheduling scheme.

[0022] Preferably, the method of processing the analysis result using the genetic algorithm includes:

[0023] Generate multiple initial scheduling plans according to the analysis results, each of which serves as a chromosome:

[0024] Initial Population = {S 1 , S 2 , ..., S n}

[0025] Among them, InitialPopulation represents the initial population, S n represents chromosome;

[0026] Design the fitness function based on production efficiency, delivery time achievement rate and resource balance:

[0027] F(S)=w 1 ×Efficiency(S)+w2 ×Delivery(S)+w 3 ×Resource Balance(S)

[0028] Among them, F(S) represents the fitness function, Efficiency(S) represents the production efficiency function, Delivery(S) represents the delivery achievement rate function, Resource Balance(S) represents the resource balance function, and w 1 、w 2 、w 3 represents the weight of the function;

[0029] Using the roulette wheel selection method, a scheduling scheme with better performance is selected from a number of the initial scheduling schemes according to the fitness function, and a new offspring scheme is generated through a crossover operation:

[0030] S child =Crossover(S parent1 , S parent2 )

[0031] Among them, S child represents the offspring scheme, Crossover() represents the crossover operation, S parent1 and S parent2 Indicates the selected scheduling scheme with better performance;

[0032] The offspring plan is mutated with a preset probability to obtain a mutated scheduling plan:

[0033] S mutated =Mutation(S child )

[0034] Among them, S mutated It indicates the scheduling plan after mutation, and Mutation() indicates the mutation operation;

[0035] The crossover operation and the mutation operation are repeatedly performed to optimize the scheduling scheme generation by generation until the fitness function converges to an optimal solution that meets the conditions, thereby obtaining the optimized scheduling scheme.

[0036] Preferably, the method of dynamically adjusting the optimized scheduling scheme using a reinforcement learning algorithm includes:

[0037] Based on the current state of the device and the current state of the batch, the state representation at each moment is obtained:

[0038] S t =(Equipment Status, Order Queue, Priority,...)

[0039] Among them, S t Indicates the status at time t, Equipment Status indicates the current status of the equipment, OrderQueue indicates the order queue, and Priority indicates the priority;

[0040] Setting an action set, the action set comprising: prioritizing high priority orders and maximizing equipment utilization;

[0041] At each moment, the reward value is obtained based on the actual production efficiency and delivery performance:

[0042] R t =w 1 ×On-time Delivery+w 2 ×Efficiency+w 3 ×Resource Utilization

[0043] Among them, R t represents the reward value, On-time Delivery represents the on-time delivery rate, Efficiency represents the production efficiency, Resource Utilization represents the resource utilization rate, and w 1 、w 2 、w 3 represents the weight of the function;

[0044] The Q-learning algorithm is used to gradually learn how to select the optimal action from the action set under different states:

[0045] Q(S t , A t )=Q(S t , A t )+α[R t +γmax Q(S t+1 , A)-Q(S t , A t )]

[0046] Among them, Q() represents the value function of the state-action pair, A t represents the action taken, α represents the learning rate, γ represents the discount factor, max represents the maximum value of the action for the next state, S t+1 Indicates the state representation at the next moment;

[0047] Through iterative learning, the scheduling strategy is dynamically updated to obtain the optimal scheduling solution.

[0048] The present invention also provides a dynamic and adaptive production scheduling system, the scheduling system applies any of the scheduling methods described above, including: a data collection and processing module, a screening module, a matching module, a device selection module and a scheduling module;

[0049] The data collection and processing module is used to collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data;

[0050] The screening module is used to screen and extract the pre-processed data to obtain valid entries, and the valid entries include: work orders to be scheduled, production equipment status, product batch information and product process;

[0051] The matching module is used to set matching rules, match the valid entries based on the matching rules, and screen out batches that meet the production conditions;

[0052] The equipment selection module is used to collect the production efficiency data of the production equipment, and select the optimal production equipment for the batch that meets the production conditions based on the production efficiency data;

[0053] The scheduling module is used to collect the current status of the equipment and the current status of the batch in real time, and adopts a dynamic adaptive algorithm to generate an optimal scheduling plan based on the optimal production equipment.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The present invention forms an efficient and flexible production scheduling system through the synergistic effect of the steps, which can dynamically respond to production changes, optimize resource allocation, and significantly improve production efficiency. In practical applications, the present invention will greatly reduce the production cycle, improve the accuracy and timeliness of order delivery, and thus enhance the market competitiveness of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0057] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] Embodiment 1

[0061] In this embodiment, if Figure 1 As shown, a dynamic and adaptive production scheduling method comprises the following steps:

[0062] S1. Collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data.

[0063] The data required for production scheduling include: work order information, equipment capacity, product process efficiency and production history. The preprocessing method includes: using data cleaning technology to process the collected data, remove locked, incorrect, duplicated, and incorrect data field values, and obtain preprocessed data.

[0064] S2. Filter and extract the pre-processed data to obtain valid entries, which include: work orders to be scheduled, production equipment status, product batch information and product processes.

[0065] S3. Set matching rules, match valid entries based on the matching rules, and filter out batches that meet production conditions.

[0066] The matching method includes: setting a weight factor and a score for each valid entry; calculating a weighted score for each valid entry based on the weight factor and the score; setting a minimum acceptance score threshold, comparing with the weighted score, and selecting valid entries with weighted scores higher than the score threshold as batches that meet production conditions.

[0067] In this embodiment, weight factors are defined: First, valid items are considered as factors affecting production scheduling, and a weight factor is defined for the factor. These factors may include work order priority, delivery urgency, equipment capacity, product complexity, material availability, etc. The weight factor is usually a value between 0 and 1, and the sum of all weight factors should be equal to 1.

[0068] Scoring criteria: Set a scoring criteria for each factor. For example, the priority of a work order can be scored as high, medium, or low, the urgency of the delivery can be scored based on the number of days until the delivery date, and the equipment capacity can be scored based on the actual operating efficiency of the equipment.

[0069] Calculate weighted score: For each valid item (for example, a work order to be scheduled), calculate its weighted score based on the scoring criteria and weight factor. The calculation formula can be:

[0070] Weighted score=(weight factor 1×score of factor 1)+…+(weight factor n×score of factor n) where n represents a natural number.

[0071] S4. Collect the production efficiency data of the production equipment and select the optimal production equipment for the batch that meets the production conditions based on the production efficiency data. This step ensures that the best processing equipment is selected by considering the maintenance status and load capacity of the equipment.

[0072] S5. Collect the current status of equipment and batches in real time, and use dynamic adaptive algorithms to generate the optimal scheduling plan based on the optimal production equipment.

[0073] The method for generating the optimal scheduling plan includes: collecting the current status of the equipment and the current status of the batch in real time, analyzing the historical production records, obtaining the analysis results, processing the analysis results using a genetic algorithm, and obtaining an optimized scheduling plan based on the optimal production equipment; using a reinforcement learning algorithm to dynamically adjust the optimized scheduling plan to obtain the optimal scheduling plan.

[0074] The method for processing the analysis result by using the genetic algorithm includes: generating a plurality of initial scheduling schemes according to the analysis result, each initial scheduling scheme being a chromosome:

[0075] Initial Population = {S 1 , S 2 , ..., S n}

[0076] Among them, InitialPopulation represents the initial population, S n Represents chromosomes; the fitness function is designed based on production efficiency, delivery time achievement rate and resource balance:

[0077] F(S)=w 1 ×Efficiency(S)+w 2 ×Delivery(S)+w 3 ×Resource Balance(S)

[0078] Among them, F(S) represents the fitness function, Efficiency(S) represents the production efficiency function, Delivery(S) represents the delivery achievement rate function, Resource Balance(S) represents the resource balance function, and w 1 、w 2 、w 3 Represents the weight of the function; uses the roulette wheel selection method to select a better performing scheduling scheme from several initial scheduling schemes according to the fitness function, and generates a new offspring scheme through crossover operation:

[0079] S child =Crossover(S parent1 , S parent2 )

[0080] Among them, S child represents the offspring scheme, Crossover() represents the crossover operation, S parent1 and S parent2 Indicates the selected scheduling scheme with better performance; the offspring scheme is mutated with a preset probability to obtain the mutated scheduling scheme:

[0081] S mutated =Mutation(S child )

[0082] Among them, S mutated Represents the scheduling plan after mutation, and Mutation() represents the mutation operation; repeatedly perform crossover and mutation operations, optimize the scheduling plan generation by generation, until the fitness function converges to the optimal solution that meets the conditions, and obtains the optimized scheduling plan.

[0083] The method of dynamically adjusting the optimized scheduling scheme using the reinforcement learning algorithm includes: obtaining the state representation at each moment based on the current state of the equipment and the current state of the batch:

[0084] S t =(Equipment Status, Order Queue, Priority,...)

[0085] Among them, S t Indicates the status at time t, Equipment Status indicates the current status of the equipment, OrderQueue indicates the order queue, and Priority indicates the priority; set an action set, which includes: giving priority to high-priority orders and maximizing equipment utilization; at each moment, get a reward value based on the actual production efficiency and delivery performance:

[0086] R t =w 1×On-time Delivery+w 2 ×Efficiency+w 3 ×Resource Utilization

[0087] Among them, R t represents the reward value, On-time Delivery represents the on-time delivery rate, Efficiency represents the production efficiency, Resource Utilization represents the resource utilization rate, and w 1 、w 2 、w 3 Represents the weight of the function; the Q-learning algorithm is used to gradually learn how to select the optimal action from the action set under different states:

[0088] Q(S t , A t )=Q(S t , A t )+α[R t +γmaxQ(S t+1 , A)-Q(S t , A t )]

[0089] Among them, Q() represents the value function of the state-action pair, A t represents the action taken, α represents the learning rate, γ represents the discount factor, max represents the maximum value of the action for the next state, S t+1 Represents the state representation at the next moment; through iterative learning, dynamically updates the scheduling strategy to obtain the optimal scheduling solution.

[0090] Embodiment 2

[0091] In this embodiment, the workflow of the present invention will be introduced by taking an actual project as an example.

[0092] S1. Collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data.

[0093] In this embodiment, key data required for production scheduling are collected and sorted, including work order information (such as priority, delivery date), equipment capacity (such as available equipment, production efficiency), product process efficiency, etc. Information is obtained from multiple data sources (such as daily schedule, work order table, equipment capacity table, etc.) to lay the foundation for subsequent scheduling analysis. The specific steps include: extracting valid data for the day from the daily schedule (DailyStepPlan). Obtaining unlocked and high-priority work order information from the work order table (WorkOrder). The collected and sorted data is stored in the SQL or Pandas library. Afterwards, the collected data is processed using data cleaning technology to remove locked, incorrectly coded, duplicated, and incorrect data field values ​​to obtain pre-processed data. The collected and sorted data sources include multiple formats, such as Excel, CSV, and databases, etc., to facilitate the flexibility and compatibility of the system.

[0094] S2. Filter and extract the pre-processed data to obtain valid entries, which include: work orders to be scheduled, production equipment status, product batch information and product processes.

[0095] In this embodiment, relevant data is extracted from the database through SQL or Pandas library, valid data entries are screened out, and valid Planned_Order is formed. Specifically: unfinished records in the daily plan table are selected, all related work orders are screened out and stored as Planned_Order, i.e. valid entries. Locked work orders are further excluded from the work order table to ensure the effectiveness of subsequent data analysis.

[0096] S3. Set matching rules, match valid items based on the matching rules, and screen out batches that meet production conditions. The matching method includes: using weights to match valid items.

[0097] In this embodiment, the information such as the work order to be scheduled, the status of the production equipment, the product batch information and the product process are matched to screen out the batches (Planned_Lot) that meet the production conditions, and ensure that the information of each batch matches the corresponding work order and process. The specific steps include: in the batch information table (LotInfo), matching the batches related to Planned_Order, and excluding the batches that have been detained and not dispatched. Obtain the products, processes and quantities corresponding to each batch to provide basic data for subsequent scheduling. Preferably, the matching process uses a data structure optimization algorithm to improve matching efficiency and accuracy.

[0098] S4. Collect the production efficiency data of production equipment and select the optimal production equipment for batches that meet the production conditions based on the production efficiency data.

[0099] In this embodiment, the data required for production scheduling is summarized, the processing equipment is screened, and the optimal processing equipment is selected for each batch, taking into account the equipment's production capacity, material change time, and unit product efficiency to ensure maximum resource utilization. The specific operations include: screening out available equipment from the product process efficiency table (StepProductUPH) and the equipment capacity table (EquipmentCapacity), and evaluating its daily standard available operating time and production capacity. For each Planned_Lot, calculate the appropriate processing equipment and record the reserved equipment operating time. Combine historical data and real-time monitoring information to dynamically adjust the equipment allocation strategy to adapt to changing production needs.

[0100] S5. Collect the current status of equipment and batches in real time, and use dynamic adaptive algorithms to generate the optimal scheduling plan based on the optimal production equipment.

[0101] Apply dynamic adaptive algorithms to generate efficient production plans based on real-time monitoring data and historical data, calculate the completion date of each order, and optimize according to the work order priority and delivery requirements. The specific implementation steps include: first, combine the availability of equipment with production needs, and use genetic algorithms and reinforcement learning to select the optimal scheduling plan. Genetic algorithms optimize scheduling decisions by simulating natural selection, giving priority to high-priority work orders and work orders that are about to expire, ensuring maximum resource utilization and reducing delivery delays. The optimized scheduling results are output to the database, and an API interface is provided for other systems or users to query and call. Specifically, it includes: outputting the batch information reservation information table (LotReservedInfo) to record the processing status and time schedule of each batch. Provide an interface for calling, input "reservation cycle, process", and return the "latest list of the current reservation batch" to achieve data sharing and system integration.

[0102] In this embodiment, the production process is also monitored in real time, and the scheduling plan is dynamically adjusted according to the actual production situation to improve the flexibility and adaptability of the production line. The specific steps include: real-time tracking of production status, monitoring of equipment load and production efficiency, and timely detection of abnormal situations. Adjust the generated schedule according to real-time data to ensure that production tasks can proceed smoothly and maximize resource utilization. The monitoring system is combined with an early warning mechanism to automatically adjust the schedule when the equipment load exceeds the threshold to ensure production continuity.

[0103] Embodiment 3

[0104] In this embodiment, a dynamic and adaptive production scheduling system includes: a data collection and processing module, a screening module, a matching module, an equipment selection module and a scheduling module.

[0105] The data collection and processing module is used to collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data; the screening module is used to screen and extract the pre-processed data to obtain valid entries, which include: work orders to be scheduled, production equipment status, product batch information and product processes; the matching module is used to set matching rules, match valid entries based on the matching rules, and screen out batches that meet the production conditions; the equipment selection module is used to collect the production efficiency data of production equipment, and select the optimal production equipment for batches that meet the production conditions based on the production efficiency data; the scheduling module is used to collect the current status of the equipment and the current status of the batch in real time, and use a dynamic adaptive algorithm to generate the optimal scheduling plan based on the optimal production equipment.

[0106] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A dynamic and adaptive production scheduling method, characterized in that: The following steps are involved: Collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data; Screening and extracting the preprocessed data to obtain valid entries, the valid entries including: work orders to be scheduled, production equipment status, product batch information and product process; Setting matching rules, matching the valid entries based on the matching rules, and screening out batches that meet production conditions; Collecting production efficiency data of the production equipment, and selecting the optimal production equipment for the batch meeting the production conditions based on the production efficiency data; The current status of the equipment and the current status of the batch are collected in real time, and a dynamic adaptive algorithm is used to generate an optimal scheduling plan based on the optimal production equipment.

2. A dynamic and adaptive production scheduling method according to claim 1, characterized in that: The data required for the production scheduling includes: work order information, equipment capacity, product process efficiency and production history records; The preprocessing method includes: using data cleaning technology to process the collected data, removing locked data, data with incorrect status, duplicate data, and data field value incorrect data, to obtain the preprocessed data.

3. The method for dynamically adaptive production scheduling according to claim 1, characterized in that: The matching method includes: Setting a weight factor and a score for each of the valid entries; Calculating a weighted score for each of the valid entries based on the weight factor and the score; A minimum acceptance score threshold is set, compared with the weighted score, and the valid entries with weighted scores higher than the score threshold are selected as the batches that meet the production conditions.

4. The method for dynamically adaptive production scheduling according to claim 1, characterized in that: The method for generating the optimal scheduling plan includes: Collect the current status of equipment and batches in real time, analyze historical production records, and obtain analysis results; Processing the analysis results using a genetic algorithm to obtain an optimized scheduling solution based on the optimal production equipment; The optimized scheduling scheme is dynamically adjusted using a reinforcement learning algorithm to obtain the optimal scheduling scheme.

5. A dynamic and adaptive production scheduling method according to claim 4, characterized in that: The method of processing the analysis result using the genetic algorithm includes: Generate multiple initial scheduling plans according to the analysis results, each of which serves as a chromosome: Initial Population={S1,S2,...,S n } Among them, InitialPopulation represents the initial population, S n represents chromosome; Design the fitness function based on production efficiency, delivery time achievement rate and resource balance: F(S)=w1×Efficiency(S)+w2×Delivery(S)+w3×Resource Balance(S) Among them, F(S) represents the fitness function, Efficiency(S) represents the production efficiency function, Delivery(S) represents the delivery achievement rate function, Resource Balance(S) represents the resource balance function, and w1, w2, and w3 represent the weights of the functions; Using the roulette wheel selection method, a scheduling scheme with better performance is selected from a number of the initial scheduling schemes according to the fitness function, and a new offspring scheme is generated through a crossover operation: S child =Crossover(S parent1 ,S parent2 ) Among them, S child represents the offspring scheme, Crossover() represents the crossover operation, S parent1 and S parent2 Indicates the selected scheduling scheme with better performance; The offspring plan is mutated with a preset probability to obtain a mutated scheduling plan: S mutated =Mutation(S child ) Among them, S mutated Represents the scheduling plan after mutation, and Mutation() represents the mutation operation; The crossover operation and the mutation operation are repeatedly performed to optimize the scheduling scheme generation by generation until the fitness function converges to an optimal solution that meets the conditions, thereby obtaining the optimized scheduling scheme.

6. A dynamic and adaptive production scheduling method according to claim 4, characterized in that: The method of dynamically adjusting the optimized scheduling scheme using a reinforcement learning algorithm includes: Based on the current state of the device and the current state of the batch, the state representation at each moment is obtained: S t =(Equipment Status,Order Queue,Priority,...) Among them, S t Indicates the status at time t, Equipment Status indicates the current status of the equipment, Order Queue indicates the order queue, and Priority indicates the priority; Setting an action set, the action set comprising: prioritizing high priority orders and maximizing equipment utilization; At each moment, the reward value is obtained based on the actual production efficiency and delivery performance: R t =w1×On-time Delivery+w2×Efficiency+w3×Resource Utilization Among them, R t represents the reward value, On-time Delivery represents the on-time delivery rate, Efficiency represents the production efficiency, Resource Utilization represents the resource utilization rate, and w1, w2, and w3 represent the weights of the function; The Q-learning algorithm is used to gradually learn how to select the optimal action from the action set under different states: Q(S t ,A t )=Q(S t ,A t )+α[R t +γmax Q(S t+1 ,A)-Q(S t ,A t )] Among them, Q( ) represents the value function of the state-action pair, A t represents the action taken, α represents the learning rate, γ represents the discount factor, max represents the maximum value of the action for the next state, S t+1 Indicates the state representation at the next moment; Through iterative learning, the scheduling strategy is dynamically updated to obtain the optimal scheduling solution.

7. A dynamic and adaptive production scheduling system, wherein the scheduling system applies the scheduling method according to any one of claims 1 to 6, characterized in that: include: Data collection and processing module, screening module, matching module, equipment selection module and scheduling module; The data collection and processing module is used to collect and organize the data required for production scheduling, and pre-process the collected data to obtain pre-processed data; The screening module is used to screen and extract the pre-processed data to obtain valid entries, and the valid entries include: work orders to be scheduled, production equipment status, product batch information and product process; The matching module is used to set matching rules, match the valid entries based on the matching rules, and screen out batches that meet the production conditions; The equipment selection module is used to collect the production efficiency data of the production equipment, and select the optimal production equipment for the batch that meets the production conditions based on the production efficiency data; The scheduling module is used to collect the current status of the equipment and the current status of the batch in real time, and adopts a dynamic adaptive algorithm to generate an optimal scheduling plan based on the optimal production equipment.

Images