Energy-saving flexible workshop scheduling method and device considering human factors

Abstract

The invention provides an energy-saving flexible workshop scheduling method and device considering human factors, and belongs to the field of workshop scheduling. The method comprises the following steps: calculating energy consumption information in each processing procedure based on production information of a workshop and energy consumption information of equipment and facilities, and summing the energy consumption information in each processing procedure to obtain total energy consumption information of a machine; establishing a worker learning ability evaluation model, and evaluating the learning ability of each worker; taking the minimum total energy consumption information of the machine and the minimum machining completion time of the machine as a target function, calculating the actual machining time for completing machining of one workpiece on one machine according to the learning ability of each worker, and generating a constraint condition based on the actual machining time and a workshop machining time sequence rule; and solving the objective function based on the constraint condition, and outputting an optimal scheduling scheme of the energy-saving flexible workshop. According to the energy-saving flexible workshop scheduling method and device considering human factors provided by the invention, the energy consumption and human factors are considered at the same time, and the workshop scheduling scheme is optimized.

Description

Technical Field

[0001] This application relates to the technical field of workshop scheduling, and particularly relates to an energy-saving flexible workshop scheduling method and device considering human factors. Background Art

[0002] Traditional manufacturing is the main body of China's manufacturing industry and the foundation of the modern industrial system. Promoting the modernization improvement and upgrading of traditional manufacturing is an active response and forward-looking guidance to the current technological innovation and industrial transformation, which brings numerous opportunities and challenges. In traditional production and manufacturing, the main production mode is mass production of a single product. However, with the continuous increase in product diversification and market personalized demands, the fixed and single production line can no longer meet the requirements of the demand side for product types and delivery times. This prompts enterprises to transform towards a customized production mode, and workshop flexibility has also become an important indicator for evaluating enterprise competitiveness. This gives rise to the Flexible Job Shop Scheduling Problem (FJSP), which evolved from the traditional Job Shop Scheduling Problem (JSP). It not only requires arranging the processing sequence of processes but also puts forward requirements on which processing machine to select. This dual optimization requirement significantly increases the computational complexity of the problem.

[0003] Since the flexible workshop scheduling problem was proposed, more attention has been paid to the improvement of production efficiency. With the deepening of the energy crisis, sustainable development strategies have been proposed at home and abroad for the future, and saving energy consumption has become an urgent problem for manufacturing enterprises. In addition, the people-oriented concept has been increasingly emphasized, and considering the worker state during the processing process and ensuring the well-being of workers have also become key issues. Therefore, it is necessary to study the energy-saving flexible workshop scheduling problem considering human factors. Summary of the Invention

[0004] In view of this, this application provides an energy-saving flexible workshop scheduling method and device considering human factors, which simultaneously consider energy consumption and human factor, and optimize the workshop scheduling plan.

[0005] Specifically, this application is implemented through the following technical solutions:

[0006] The first aspect of this application provides an energy-saving flexible workshop scheduling method considering human factors, and the method includes:

[0007] Calculating the energy consumption information in each processing operation based on the production information of the workshop and the energy consumption information of equipment and facilities, and summing the energy consumption information in each processing operation to obtain the total machine energy consumption information, where the total machine energy consumption information changes with the production information, and the degree of change is affected by the energy consumption information of equipment and facilities;

[0008] Establish a worker learning ability evaluation model to evaluate the learning ability of each worker;

[0009] Among them, the worker learning ability evaluation model includes a machine operation efficiency sub-model and a workpiece operation efficiency sub-model. The machine operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the machine, and the workpiece operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the workpiece. The learning ability is affected by the sum of the machine operation efficiency and the workpiece operation efficiency;

[0010] Taking the minimum of the total machine energy consumption information and the machine processing completion time as the objective function, calculate the actual processing time for processing a workpiece on one machine according to the learning ability of each worker, and generate constraint conditions based on the actual processing time and the workshop processing time sequence rules;

[0011] Solve the objective function based on the constraint conditions and output the optimal scheduling plan for the energy-saving flexible workshop.

[0012] The second aspect of this application provides an energy-saving flexible workshop scheduling device considering human factors. The device includes a calculation module, a establishment module, a generation module, and a solution module; among them,

[0013] The calculation module is used to calculate the energy consumption information in each processing procedure based on the production information of the workshop and the energy consumption information of the equipment and facilities, and sum the energy consumption information in each processing procedure to obtain the total machine energy consumption information. Among them, the total machine energy consumption information changes with the production information, and the degree of change is affected by the energy consumption information of the equipment and facilities;

[0014] The establishment module is used to establish a worker learning ability evaluation model to evaluate the learning ability of each worker;

[0015] Among them, the worker learning ability evaluation model includes a machine operation efficiency sub-model and a workpiece operation efficiency sub-model. The machine operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the machine, and the workpiece operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the workpiece. The learning ability is affected by the sum of the machine operation efficiency and the workpiece operation efficiency;

[0016] The generation module is used to take the minimum of the total machine energy consumption information and the machine processing completion time as the objective function, calculate the actual processing time for processing a workpiece on one machine according to the learning ability of each worker, and generate constraint conditions based on the actual processing time and the workshop processing time sequence rules;

[0017] The solution module is used to solve the objective function based on the constraint conditions and output the optimal scheduling plan for the energy-saving flexible workshop.

[0018] The energy-saving flexible workshop scheduling method and device considering human factors provided by this application. In the first aspect, by accurately calculating the energy consumption information of each processing operation and integrating this information into the total energy consumption of the machine, this process not only considers the energy consumption of the equipment and facilities themselves, but also fully incorporates the energy consumption involved in the production information, and can clearly identify which steps or equipment are the main energy consumption sources during the workshop production process. Taking the total energy consumption information of the machine as one of the objective functions, it can guide the scheduling algorithm to preferentially select processing routes or scheduling strategies with lower energy consumption, thereby reducing the overall energy consumption of the workshop. In addition, since the total energy consumption information of the machine is dynamically adjusted with the change of production information, and this change is affected by the energy consumption information of the equipment and facilities, the energy consumption calculation accuracy is high, and the scheduling plan can adapt to the dynamic changes in the production process in real time, ensuring that the workshop can maintain efficient operation under different production conditions.

[0019] In the second aspect, first, by combining the machine operation efficiency and the workpiece operation efficiency to form a worker learning ability evaluation model, it can more accurately quantify and evaluate the performance of workers in different tasks, and by adjusting the scheduling strategy, workers can give full play to their potential. As workers become more familiar with the operation process, their operation efficiency will gradually increase, thereby shortening the processing time. Secondly, minimizing the total energy consumption of the machine and the processing completion time as the objective function not only considers the energy consumption problem but also ensures production efficiency. In addition, by using the worker learning ability evaluation model to calculate the actual processing time and generating constraint conditions based on this, it can more accurately reflect various limitations in the actual production process. Considering the processing time sequence rules of the workshop, this kind of constraint condition based on the actual worker ability is closer to reality and can effectively avoid the scheduling failure problem caused by the inconsistency between the scheduling plan and the actual production situation. Description of the Drawings

[0020] Figure 1 It is a flowchart of the first embodiment of the energy-saving flexible workshop scheduling method considering human factors provided by this application;

[0021] Figure 2 It is a structural diagram of calculating the total energy consumption information of the machine provided by this application;

[0022] Figure 3 It is a schematic diagram of the principle of solving the objective function provided by this application;

[0023] Figure 4 It is a schematic structural diagram of the first embodiment of the energy-saving flexible workshop scheduling device considering human factors provided by this application. Detailed Embodiments

[0024] Exemplary embodiments will be described in detail herein, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application.

[0025] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit the present application. The singular forms "a", "the", and "said" used in this application are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0026] It should be understood that although the terms first, second, third, etc. may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0027] Specific embodiments are given below to introduce the technical solutions of the present application in detail.

[0028] Figure 1 This is a flowchart of the first embodiment of the energy-saving flexible workshop scheduling method considering human factors provided for the present application. Please refer to Figure 1 , the method provided in this embodiment may include:

[0029] S101. Calculate the energy consumption information in each processing procedure based on the production information of the workshop and the energy consumption information of the equipment and facilities, and sum the energy consumption information in each processing procedure to obtain the total machine energy consumption information, where the total machine energy consumption information changes with the production information, and the degree of change is affected by the energy consumption information of the equipment and facilities.

[0030] It should be noted that the production information of the workshop may include information such as machines, the number of workpieces, machine processing time, the number of machine processing procedures, whether preparation is required for the procedure, whether operation is required between procedures, machine idle time, and the maximum completion time. The energy consumption information of the equipment and facilities in the workshop may include information such as the energy consumption per unit time of machine operation, the energy consumption of machine single start-up and shutdown, the energy consumption of machine single conversion of work, the energy consumption per unit time of machine idle, and the public energy consumption information of the production workshop per unit time.

[0031] Specifically, based on the production information of the workshop and the energy consumption information of equipment and facilities, calculate the energy consumption information in each processing operation, and sum up the energy consumption information in each processing operation to obtain the total machine energy consumption information, including:

[0032] (1) Determine multiple working states of each machine according to the production information of the workshop, divide the processing operations according to the continuous production tasks executed by each machine in different working states, calculate the process energy consumption of different processing operations of each machine, and the process energy consumption of each process is affected by the energy consumption information of the machine facilities corresponding to the process.

[0033] It should be noted that according to the workshop information, different working states of the machine during production can be identified and determined. Generally, these states can include: startup, processing, preparation, transportation, idling, pause, shutdown, etc. Among them, different working states correspond to different energy consumption levels.

[0034] After clarifying the working state of the machine, the entire processing process can be further divided into multiple processes. Specifically, the processing processes can be divided according to the continuous production tasks executed by the machine in different states. The division of these processes can be based on time periods, task types, or changes in machine states. For example, when the machine is in the running state, if a specific processing task is continuously executed, the process from the start to the end of the task can be divided into a processing operation; when the machine pauses and then restarts to execute a new task, the execution process of the new task is divided into a new processing operation. After the processing operations are divided, the energy consumption of each processing operation will be calculated according to the energy consumption information of the corresponding machine facilities. The process energy consumption of each process is affected by the energy consumption information of the machine facilities corresponding to the process. Specifically, the process energy consumption is equal to the energy consumption of the machine facilities in the process per unit time multiplied by the duration of the process.

[0035] Figure 2 This is the structure diagram for calculating the total machine energy consumption information provided by this application. Please refer to Figure 2 , after the processing operations are divided, the energy consumption of each processing operation can include processing energy consumption, preparation energy consumption, transportation energy consumption, idling energy consumption, and common energy consumption. Among them, the processing energy consumption is the energy consumption of the machine to complete the processing operation of the process. In addition to the energy consumption when directly acting on the workpiece, it also includes energy-consuming items such as the necessary movement and rotation of the tool. Specifically, its calculation formula is:

[0036] ;

[0037] Among them, represents the th machine processing the workpiece energy consumption per unit time, represents the th machine processing the workpiece total duration.

[0038] Further, calculating the energy consumption information in each processing step based on the production information of the workshop and the energy consumption information of equipment and facilities includes: determining the machines used in the workshop as the target machines for calculating energy consumption information according to the production information. The production information includes multiple items, each of which involves a different order, and the machines included may be completely the same, completely different, or partially the same. Establishing a work schedule for each machine in the target machines according to the production information. If a target machine is involved in two production information during operation, the time of the target machine participating in each production information does not conflict. Dividing the processing steps of the target machines according to the work schedule to obtain multiple initial states to be calculated. The processing step times corresponding to the same production information are not completely the same, and the processing step times of the same processing step for different production information are not completely the same; merging the same processing steps in the initial states to be calculated to obtain the target states to be calculated; calculating the energy consumption in each target state to be calculated.

[0039] The preparation energy consumption refers to the energy consumed by the machine during the preparation work before processing the process, such as tool changing, tool setting, workpiece clamping, etc. Generally, only when two adjacent processes of the machine processing belong to different tools respectively, adjustment and replacement are required. Usually, the processing operations performed by the same machine are similar, so the energy consumed by the adjustment operations performed by the same machine is simplified and processed into a unified fixed quantity for easy calculation. The calculation formula for the preparation energy consumption is as follows:

[0040] ;

[0041] Among them, represents the energy consumed by the machine for one preparation work, represents the number of processes processed by the machine , .

[0042] The conveying energy consumption refers to the energy consumed when the workpiece is transferred from one machine to another machine, which is the energy consumption of the conveying equipment (such as forklift, conveyor belt, etc.) during the process of conveying the workpiece (or raw material). Generally, the conveying operation only occurs when the adjacent two processes of the workpiece are processed on different machines. The calculation formula for the conveying energy consumption is:

[0043] ;

[0044] Among them, represents the energy consumption of the workpiece from the th process to the th process, represents the workpiece The number of processes, .

[0045] Idle energy consumption refers to the energy consumed by a machine when it is in the on state but not performing any processing operations. This is mainly because the machine is waiting for workpieces or workers. The formula for calculating idle energy consumption is as follows:

[0046] ;

[0047] Among them, represents the energy consumption per unit time of the machine in the idle state, represents the machine idle time.

[0048] It should also be noted that during the production process, there is also a part of the energy consumption for auxiliary facilities to ensure the normal operation of the production workshop. This type of energy consumption is called public energy consumption. It includes, but is not limited to, the energy consumption of air conditioning, ventilation, lighting, and heating systems in the workshop. Although these systems do not directly participate in the processing of products, they are essential for maintaining a suitable working environment in the workshop and ensuring the smooth progress of production activities. In actual production activities, public energy consumption is usually relatively stable because it is mainly affected by the scale of the workshop and operating conditions, and these factors do not change much within a certain period. Specifically, the formula for calculating public energy consumption is:

[0049] ;

[0050] Among them, ce represents the public energy consumption of the production workshop per unit time, represents the maximum completion time.

[0051] (2) For each machine, calculate the total energy consumption information of the machine based on the sum of the energy consumption of different processing processes of the machine.

[0052] Specifically, it can be calculated according to the following formula:

[0053] ;

[0054] Among them, PE, RE, TE, NE, and CE are processing energy consumption, preparation energy consumption, transportation energy consumption, idle energy consumption, and public energy consumption respectively.

[0055] It should be noted that by linking the working state of the machine with specific processing procedures and calculating the energy consumption of each procedure, the accuracy of energy consumption assessment can be improved. After clarifying the energy consumption characteristics of each procedure, the production plan and scheduling strategy can be adjusted more pertinently. For example, for procedures with high energy consumption, they can be scheduled to avoid peak operation periods as much as possible, or concentrated when the machine is in good condition, so as to reduce unnecessary energy consumption. In addition, by understanding how the energy consumption of each procedure is affected by the working state of the machine, the workshop can take measures to ensure that the machine operates in a high-efficiency and low-consumption state as much as possible. For example, avoiding frequent startup and shutdown of the machine can not only reduce energy consumption, but also extend the service life of the equipment.

[0056] Furthermore, for the generated information of the same machine that is adjacent in time, there will be a situation where adjacent procedures overlap in time. For example, during the overlapping time period, the common energy consumption is shared information, the preparation energy consumption is reduced because there is no need to replace equipment, and at the same time, the idling energy consumption can be reduced. Therefore, in order to further improve the accuracy of energy consumption calculation, the present invention fully considers the production information of using the same type of parts and machines and preferentially arranges them together. Specifically, calculating the energy consumption information in each processing procedure based on the production information of the workshop and the energy consumption information of equipment and facilities includes: sorting each production information based on the time information in the production information to obtain ordered production information; traversing the adjacent production information in the ordered production information to determine whether the types of machines and parts used in the adjacent production information are the same; if the same, binding the adjacent production information to obtain the adjusted ordered production information; calculating the energy consumption according to the ordered production information, where the energy consumption of the production information with a binding relationship is calculated collaboratively. For independent production information, it is calculated according to the steps described above. For two production information with a binding relationship, perform a time overlap operation on the head of the first production information and the tail of the second production information, perform a time overlap process on the procedures where the machine usage states do not conflict, and calculate the energy consumption according to the procedures adjusted in time. For the first production information, determine each procedure according to time and calculate the first energy consumption of each procedure; for the second production information, determine the overlapping time period with the first production information, determine the auxiliary parts with different work contents from those in the first production information during the overlapping time period, and calculate the energy consumption of the auxiliary parts as the energy consumption of the overlapping time period of the second production information; calculate the energy consumption of the non-overlapping time period of the second production information; take the sum of the three (the energy consumption of the first production information, the energy consumption of the overlapping time period of the second production information, and the energy consumption of the non-overlapping time period of the second production information) as the total energy consumption of the bound production information. Further, the bound production information is used as a constraint and scheduled on the same machine and adjacent in time.

[0057] S102. Establish a worker learning ability evaluation model to evaluate the learning ability of each worker.

[0058] Among them, the worker learning ability evaluation model includes a machine operation efficiency sub-model and a workpiece operation efficiency sub-model. The machine operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the machine, and the workpiece operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the workpiece. The learning ability is affected by the sum of the machine operation efficiency and the workpiece operation efficiency.

[0059] It should be noted that considering the instantaneous learning effect, the actual processing time of a certain process is taken as a function of the cumulative time of the operations arranged in front of it, that is, learning depends on time, and the efficiency is regarded as a function of the cumulative processing time to establish a worker learning ability evaluation model. Among them, the established worker learning ability evaluation model is as follows:

[0060] ; ; ;

[0061] Among them, represents the label of the workpiece; represents the label of the process; represents the label of the machine; r represents the label of the worker; n represents the total number of workpieces, represents the workpiece contains the number of processes; represents the learning rate parameter of worker r; represents the basic machine operation efficiency of worker r when performing machine processing; represents the basic workpiece operation efficiency of worker r when performing workpiece processing; represents the machine can achieve the maximum efficiency; represents the cumulative operation time of machine by worker r before processing ; represents the workpiece the th process; represents the cumulative processing time of workpiece by worker r before processing ; represents the actual processing time of the process ; represents the basic processing time of worker r for the jth process of the ith workpiece on machine k; the actual processing time of worker r for the jth process of the ith workpiece on machine k after considering the worker learning effect; represents the judgment process Whether it is processed by worker r on machine k; Indicates a judgment process Whether it is processed by worker r; Indicates a judgment process Whether the completion time of the process is greater than the start time of process ;

[0062] Specifically, evaluate the learning ability of each worker, including:

[0063] (1) Determine the basic operation efficiency, which includes the basic machine operation efficiency and the basic workpiece operation efficiency.

[0064] It should be noted that for a specific process, there is a standard working hour , and each worker has a basic operation efficiency for a machine operation or a workpiece processing task (basic machine operation efficiency) or (basic workpiece operation efficiency), and they are not necessarily less than 1.

[0065] (2) Calculate the actual operation efficiency during the processing based on the learning rate parameter and the basic operation efficiency, where the learning rate parameter represents the impact of the cumulative learning experience of the worker on the operation efficiency over time.

[0066] It should be noted that for the work of processing workpieces and operating machines, the learning of the two is carried out simultaneously during this process. Therefore, before the processing process , the cumulative learning time of worker r is divided into the cumulative time for operating the machine and the cumulative time for processing the workpiece , , and that is, they respectively represent the actual efficiency of worker r facing the machine and the workpiece .

[0067] (3) Take the average of the calculated actual machine operation efficiency and actual workpiece operation efficiency to obtain the final actual operation efficiency.

[0068] It should be noted that in fact, it is very common in production to have different degrees of familiarity with machines and workpieces. Therefore, the actual efficiency needs to consider both of them at the same time, and it is set to be determined by the average of the two.

[0069] Furthermore, in some processes, once the machine is adjusted in the initial situation, there is no need to repeatedly adjust and use it. In order to improve the accuracy of the actual operation efficiency, determine the adjustment frequency of the machine and the parameter adjustment frequency of the workpiece according to the production information. Preferably, the ratio of the number of times the machine is adjusted to the total number of steps of different actions of the machine can be used as the adjustment frequency of the machine. Similarly, the ratio of the number of times the workpiece parameters are adjusted to the upper limit of the total number of parameter adjustments of the workpiece during the processing can be used as the parameter adjustment frequency of the workpiece. Take the adjustment frequency of the machine and the parameter adjustment frequency of the workpiece as the weights of the machine operation efficiency and the workpiece operation efficiency respectively, and calculate the weighted sum as the final actual operation efficiency.

[0070] (4) Determine the maximum efficiency constraint and calculate the learning ability of the worker based on the final actual operation efficiency.

[0071] It should be noted that since the processing efficiency of each machine cannot be infinitely improved with the increase of the cumulative time, because there is an upper limit efficiency for the machine itself (the upper limit efficiency value is the smallest, because the larger the value of "efficiency" here, the smaller the actual processing time), and finally the learning ability of the worker is obtained as .

[0072] It should also be noted that after establishing a worker learning ability evaluation model and evaluating the learning ability of each worker, it also includes establishing a worker fatigue model. The establishment process of the worker fatigue model includes: establishing a worker fatigue-recovery index model; obtaining the fatigue degree information of the worker at different times and the current working state information, and updating the index model based on the fatigue degree information and the working state information; receiving the processing task and judging the remaining processable time of the worker to form worker allocation information; establishing a worker fatigue model based on the updated index model and the worker allocation information.

[0073] It should be noted that the worker fatigue-recovery index model is the basis of the entire worker fatigue model. By introducing specific parameters and mathematical relationships, it quantifies the fatigue accumulation of workers during work and the fatigue recovery during rest. For example, the worker fatigue-recovery index model may include key parameters such as the fatigue growth coefficient and the recovery coefficient. The setting of these parameters is based on a large amount of experimental data and research on the physiological characteristics of workers' work. By reasonably setting these parameters, it can more accurately reflect the fatigue change trend of workers under different work intensities and rest durations. Further, obtain the fatigue degree information of workers at different times and the current work status information. Among them, the fatigue degree information can be collected in various ways, such as indirectly reflecting the fatigue degree by monitoring the physiological indicators of workers (such as heart rate variability, electromyogram signals, etc.) with the help of wearable devices, or obtaining it through the subjective evaluation of workers themselves. The work status information includes the type of task currently executed by the worker, the task difficulty, the continuous working time, etc. Based on this rich fatigue degree information and work status information, update the previously established index model. For example, if the worker is currently performing a high-intensity task and has been working continuously for a long time, and the fatigue degree has increased significantly, the model will adjust the fatigue growth coefficient accordingly according to these actual data, so that the model can better fit the actual fatigue change of workers.

[0074] After receiving the processing task, judge the workload that each worker can undertake based on the current fatigue state of the worker and the remaining available processing time. The calculation of the remaining available processing time needs to comprehensively consider factors such as the current fatigue degree of the worker, the expected fatigue recovery time, and the urgency of the task. For example, for workers with a higher fatigue degree, their remaining available processing time is correspondingly reduced, and their workload will be appropriately reduced when assigning tasks. Through such judgment, detailed worker allocation information is formed to ensure reasonable task allocation and avoid over-fatigue operation of workers. Finally, based on the updated index model and the above worker allocation information, a complete worker fatigue model is established. This model will comprehensively reflect the fatigue state of workers under different working conditions and its impact on task allocation, providing a key basis for workshop scheduling. For example, when scheduling tasks, the system can, according to the worker fatigue model, preferentially assign tasks to workers with a lower fatigue degree and sufficient remaining available processing time, so as to ensure the efficient and stable progress of production.

[0075] S103. Take the minimum of the total machine energy consumption information and the machine processing completion time as the objective function, calculate the actual processing time for processing a workpiece on one machine according to the learning ability of each worker, and generate constraint conditions based on the actual processing time and the workshop processing time sequence rules.

[0076] Specifically, the objective function is:

[0077] ; ;

[0078] It should be noted that constraint conditions are generated based on the actual processing time and the workshop processing time sequence rules, including:

[0079] Calculate the actual processing time for processing a workpiece on a machine according to the learning ability of each worker, and generate constraint conditions based on the actual processing time and the workshop processing time sequence rules;

[0080] Determine the first constraint on the process processing time and process processing order of a workpiece;

[0081] Determine the second constraint on the processing time of a workpiece and the processing time of adjacent workpieces on a machine;

[0082] Determine the third constraint on the allocation relationship among workers, machines and processes;

[0083] Generate constraint conditions based on the first constraint, the second constraint and the third constraint.

[0084] Specifically, the generated constraint conditions are: ; Among them, ; ; ; Among them, ; ; Among them, ; ; Among them, ; ; Among them, ; = 1,..., ; ; Among them, ; ; Among them, ;

[0085] It should be noted that the above formula represents that the completion time of process is equal to the start time plus the actual processing time. Formula The upper limit of the worker's fatigue level is restricted, and both are within the range of [0, 1]. The formula ensures that the subsequent process in the workpiece will not be processed in advance beyond the previous process. The formula guarantees that on one machine, subsequent work cannot start until the previous work is completed. The formula ensures that there is always sufficient processing time for the subsequent process between two processes processed on the same machine, without time overlap. The formula ensures that a worker can only change the operating machine after completing the processing task on one machine. The formula ensures that each process is assigned and only assigned to one processing machine. The formula ensures that each process on the machine has and only has one worker responsible for processing.

[0086] S104. Solve the objective function based on the above constraints, and output the optimal scheduling plan of the energy-saving flexible workshop.

[0087] Figure 3 The schematic diagram for solving the objective function provided by this application is shown in Figure 3 , solving the objective function based on the above constraints includes:

[0088] (1) The agent obtains the current state of the workshop, processes it using a neural network based on the current state, and outputs the Q-values of each action; the Q-value represents the expected return of performing a certain action in the current state.

[0089] It should be noted that the agent obtains the current state of the workshop. These state information includes the utilization rate of the machine, the process progress, the state of the worker, and the energy consumption, etc. Through these information, the agent can comprehensively understand the current production status of the workshop.

[0090] (2) Execute the selected action, change the workshop state, and record this process as an experience by the agent and store it in the experience pool.

[0091] The agent processes the current state using a neural network and outputs the Q-values of each possible action. The Q-value represents the expected return of performing a certain action (for example, arranging a certain process) in the current state. The agent selects the action with the largest Q-value, which means selecting the next process arrangement that is most beneficial to the workshop production, so that the workshop enters the next state.

[0092] (3) When the number of experience entries in the experience pool reaches the set minimum batch size, the agent extracts a preset number of experiences from the experience pool for replay.

[0093] After the selected action is executed, the state of the workshop changes, and the agent records this process as an experience, including the current state , the executed action , the obtained reward (for example, reduced energy consumption or increased production efficiency), and the new state after the action is executed . These experience entries are stored in the experience pool

[0094] (4) Calculate the difference between the target Q-value and the predicted Q-value based on the experience replay process, and update the main network

[0095] When the number of experience entries in the experience pool reaches the set minimum batch size, the agent will draw a batch of experiences from the experience pool for replay. Experience replay is to enable the agent to make full use of past experiences and optimize its decision-making strategy

[0096] For each selected experience, the agent first uses the target network to calculate the target Q-value of the action in the current state. Then, the main network (main neural network) calculates the predicted Q-value based on the current state

[0097] It should be noted that the difference between the calculated target Q-value and the predicted Q-value is calculated through the loss function and used to adjust the weights of the main network. Through the gradient descent method, the agent gradually optimizes the performance of the main network, making the action selection of the main network more accurate in different states. When the training reaches a certain number of steps, the agent copies the weights of the main network to the target network. This step ensures the update of the target network, enabling it to better guide the subsequent experience replay process and further improve the scheduling ability of the agent

[0098] (5) Continuously loop and iterate the above operations until the optimal scheduling plan is output

[0099] The agent continuously obtains the workshop state, selects actions, stores and replays experiences, and updates the network weights. Over time, the agent gradually learns to select the optimal scheduling strategy in different production states and finally outputs an efficient and energy-saving workshop scheduling plan

[0100] Through the continuous learning and optimization of the agent, the workshop scheduling solution method based on deep reinforcement learning can effectively search for the optimal scheduling plan in the complex state space. This method can dynamically adapt to the real-time production situation of the workshop, improve the overall production efficiency, and reduce energy consumption

[0101] It should be noted that the state design in the field of workshop scheduling is mainly used to characterize the workshop environment. Specifically, solving the objective function based on the above-mentioned constraint conditions further includes: calculating the machine efficiency state characteristics based on the historical production data of the workshop and the equipment characteristics; the machine efficiency state characteristics include the average machine utilization rate, the standard deviation of the machine utilization rate, the average process completion rate, and the standard deviation of the process completion rate; calculating the energy consumption state characteristics based on the current workshop scheduling task requirements; the energy consumption state characteristics include the expected total energy consumption; calculating the worker state characteristics based on the worker fatigue data; the worker state characteristics include the standard deviation of worker fatigue, the standard deviation of the worker's learning preference for machines, and the standard deviation of the worker's learning preference for workpieces; determining the workshop state based on the machine efficiency state characteristics, the energy consumption state characteristics, and the worker state characteristics.

[0102] It should be noted that for the specific calculation methods of the machine efficiency state characteristics, the energy consumption state characteristics, and the worker state characteristics, please refer to the descriptions of related technologies and will not be elaborated here.

[0103] It should be noted that the reward function is used to evaluate the quality of the agent's action selection and to guide the agent to act in the direction of a better solution during the interactive learning process. The rationality of its setting is crucial for the solution effect of the entire algorithm. The reward function needs to reflect the optimization goal and also consider whether the agent can fully receive feedback.

[0104] Specifically, solving the objective function based on the above-mentioned constraint conditions further includes setting the reward function, which specifically includes:

[0105] Comparing the change degree before and after the target value to obtain the change rate information;

[0106] Determining the reward value based on the change rate information;

[0107] Specifically, the reward function is as shown in the formula:

[0108] ;

[0109] Among them, are the weights of the two objectives respectively, and there is ; is the makespan, is the makespan calculated at time t, SE(t) is the total energy consumption at time t, and SE(t - 1) is the maximum total energy consumption calculated at time t - 1. Specifically, reasonable weights can be set according to the judgment of the importance of the two.

[0110] In this embodiment, by assigning weights to two objective values, it is transformed into a single-objective problem. Since both objectives in this embodiment change non-decreasingly over time, and the reward function needs to combine rewards and punishments (otherwise the agent cannot learn effectively), the reward cannot be directly given by comparing the objective value at a certain state with that at the previous state. Instead, the degree of change between the front and back states is compared. When the increase in the objective value is large, a lower reward value is given; otherwise. Since the dimensions of the total energy consumption and the total completion time are different, the change rates of the two are used as the basis for the reward here, and their negative values are used as the reward value to give a greater reward when the objective value decreases.

[0111] It should be noted that by setting the reward function, the scheduling strategy can be dynamically adjusted during the workshop scheduling process to maximize certain key indicators (such as the lowest energy consumption and the shortest completion time). This mechanism helps to continuously optimize the production process and improve the overall efficiency.

[0112] The energy-saving flexible workshop scheduling method considering human factors provided in this embodiment, on the one hand, by accurately calculating the energy consumption information of each processing operation and synthesizing this information into the total energy consumption of the machine, it can clearly identify which steps or equipment are the main energy consumption sources during the workshop production process. Taking the total energy consumption information of the machine as one of the objective functions, it can guide the scheduling algorithm to preferentially select processing routes or scheduling strategies with lower energy consumption, thereby reducing the overall energy consumption of the workshop. In addition, since the total energy consumption information of the machine is dynamically adjusted with the change of production information, and this change is affected by the energy consumption information of equipment and facilities, the scheduling plan can adapt to the dynamic changes in the production process in real time, ensuring that the workshop can maintain efficient operation under different production conditions. On the other hand, first, by combining the machine operation efficiency and the workpiece operation efficiency to form a worker learning ability evaluation model, it can more accurately evaluate the performance of workers in different tasks, and by adjusting the scheduling strategy, workers can give full play to their potential. As workers become more familiar with the operation process, their operation efficiency will gradually increase, thereby shortening the processing time. Secondly, minimizing the total energy consumption of the machine and the processing completion time as the objective function not only takes into account the energy consumption problem but also ensures production efficiency. In addition, by calculating the actual processing time through the worker learning ability evaluation model and generating constraint conditions based on this, it can more accurately reflect various limitations in the actual production process. Considering the processing time sequence rules of the workshop, this kind of constraint condition based on the actual worker ability is closer to reality and can effectively avoid the scheduling failure problem caused by the inconsistency between the scheduling plan and the actual production situation. Specifically, the fatigue degree of workers is also considered, and a worker fatigue degree model is established, which helps to avoid the reduction of efficiency and operation errors caused by worker fatigue, ensuring the continuity and stability of the production process. The worker fatigue degree model can also help to reasonably arrange the working hours of workers and extend the effective working hours of workers. In addition, using neural network and experience replay technology, the scheduling plan is optimized through continuous learning and iteration to ensure the output of the optimal scheduling result. This deep learning-based scheduling method can adaptively adjust to different production environments and requirements, greatly improving the flexibility and intelligence of scheduling.

[0113] Corresponding to the foregoing embodiment of the energy-saving flexible workshop scheduling method considering human factors, this application also provides an embodiment of an energy-saving flexible workshop scheduling device considering human factors.

[0114] Figure 4 It is a schematic structural diagram of the first embodiment of the energy-saving flexible workshop scheduling device considering human factors provided in this application. Please refer to Figure 4 The device provided in this embodiment includes a calculation module 410, a establishment module 420, a generation module 430, and a solution module 440; among them,

[0115] The calculation module 410 is configured to calculate the energy consumption information in each processing operation based on the production information of the workshop and the energy consumption information of the equipment and facilities, and sum up the energy consumption information in each processing operation to obtain the total energy consumption information of the machine. Wherein, the total energy consumption information of the machine changes with the production information, and the degree of the change is affected by the energy consumption information of the equipment and facilities;

[0116] The establishment module 420 is configured to establish a worker learning ability evaluation model and evaluate the learning ability of each worker;

[0117] Wherein, the worker learning ability evaluation model includes a machine operation efficiency sub-model and a workpiece operation efficiency sub-model. The machine operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the machine, and the workpiece operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the workpiece. The learning ability is affected by the sum of the machine operation efficiency and the workpiece operation efficiency;

[0118] The generation module 430 is configured to take the total energy consumption information of the machine and the minimum machine processing completion time as the objective function, calculate the actual processing time for processing a workpiece on one machine according to the learning ability of each worker, and generate constraint conditions based on the actual processing time and the workshop processing time sequence rules;

[0119] The solution module 440 is configured to solve the objective function based on the constraint conditions and output the optimal scheduling scheme of the energy-saving flexible workshop.

[0120] The device of this embodiment can be used to execute Figure 1 the steps of the method embodiment shown. The specific implementation principle and process are similar and will not be elaborated here.

[0121] For the implementation process of the functions and roles of each unit in the above device, please refer to the implementation process of the corresponding steps in the above method for details, which will not be elaborated here.

[0122] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this application. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0123] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the scope of protection of the present application.

Claims

1. An energy-saving flexible workshop scheduling method considering human factors, characterized in that: The method comprises: Calculate the energy consumption information in each processing step based on the production information of the workshop and the energy consumption information of the equipment and facilities, and sum the energy consumption information in each processing step to obtain the total energy consumption information of the machine, wherein the total energy consumption information of the machine changes with the production information, and the degree of the change is affected by the energy consumption information of the equipment and facilities; Establish a worker learning ability assessment model to assess each worker’s learning ability; The worker learning ability evaluation model includes a machine operation efficiency sub-model and a workpiece operation efficiency sub-model. The machine operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the machine, and the workpiece operation efficiency sub-model is used to evaluate the workpiece operation efficiency of each worker. The learning ability is affected by the sum of the machine operation efficiency and the workpiece operation efficiency. Taking the total energy consumption information of the machine and the minimum processing completion time of the machine as the objective function, the actual processing time of processing a workpiece on a machine is calculated according to the learning ability of each worker, and the constraint conditions are generated based on the actual processing time and the workshop processing timing rules; The objective function is solved based on the constraint conditions, and an optimal scheduling solution for the energy-saving flexible workshop is output.

2. The method according to claim 1, characterized in that The assessment of each worker's learning ability includes: Determining basic operation efficiency, wherein the basic operation efficiency includes basic machine operation efficiency and basic workpiece operation efficiency; Calculating the actual operation efficiency in the processing process based on a learning rate parameter and the basic operation efficiency, wherein the learning rate parameter represents the influence of the worker's accumulated learning experience over time on the operation efficiency in the processing process; The calculated actual machine operation efficiency and the actual workpiece operation efficiency are averaged to obtain the final actual operation efficiency; A maximum efficiency constraint is determined, and a worker's learning ability is calculated based on the final actual operating efficiency.

3. The method according to claim 1, characterized in that The energy consumption information in each processing step is calculated based on the production information of the workshop and the energy consumption information of equipment and facilities, and the total energy consumption information of the machine is obtained by summing up the energy consumption information in each processing step, including: Determine multiple working states of each machine according to the production information of the workshop, divide the processing procedures according to the continuous production tasks performed by each machine in different working states, calculate the process energy consumption of different processing procedures of each machine, and the process energy consumption of each process is affected by the energy consumption information of the machine facilities corresponding to the process; For each machine, the total energy consumption information of the machine is calculated based on the sum of the process energy consumptions of different processing processes of the machine.

4. The method according to claim 1, characterized in that: The establishment of the worker learning ability evaluation model further includes establishing a worker fatigue model after evaluating the learning ability of each worker. The establishment process of the worker fatigue model includes: Establish a worker fatigue-recovery index model; Obtain workers’ fatigue information at different times and their current work status information; updating the index model based on the fatigue information and the work status information; Receive processing tasks and determine the remaining processing time of workers to form worker allocation information; A worker fatigue model is established based on the updated index model and the worker allocation information.

5. The method according to claim 1, characterized in that The generation of constraint conditions based on the actual processing time and the workshop processing timing rules includes: Calculate the actual processing time of processing a workpiece on a machine according to the learning ability of each worker, and generate constraint conditions based on the actual processing time and the workshop processing timing rules; Determine the first constraint of the process processing time and process processing sequence of a workpiece; A second constraint for determining a processing time of a workpiece and a processing time of adjacent workpieces on a machine; The third constraint that determines the distribution relationship between workers, machines and processes; A constraint condition is generated based on the first constraint, the second constraint, and the third constraint.

6. The method according to claim 1, characterized in that The solving the objective function based on the constraint condition comprises: The agent obtains the current state of the workshop, processes it using a neural network based on the current state, and outputs the Q value of each action; the Q value represents the expected return of performing an action in the current state; Execute the selected action, change the state of the workshop, and record the process as an experience and store it in the experience pool through the intelligent agent; When the number of experience bars in the experience pool reaches a set minimum batch, the agent extracts a preset number of experiences from the experience pool for playback; The experience replay process calculates the difference between the target Q value and the predicted Q value and updates the main network; The above operations are iterated repeatedly until the optimal scheduling solution is output.

7. The method according to claim 1, characterized in that The solving the objective function based on the constraint condition further includes: Calculate machine efficiency status characteristics based on workshop historical production data and equipment characteristics; the machine efficiency status characteristics include average machine utilization, machine utilization standard deviation, average process completion rate and process completion rate standard deviation; Calculate energy consumption state characteristics based on current workshop scheduling task requirements; the energy consumption state characteristics include expected total energy consumption; Calculating worker status characteristics based on worker fatigue data; the worker status characteristics include worker fatigue standard deviation, worker learning bias standard deviation of the machine and worker learning bias standard deviation of the workpiece; The workshop status is determined based on the machine efficiency status characteristics, the energy consumption status characteristics and the worker status characteristics.

8. The method according to claim 1, characterized in that Solving the objective function based on the constraint conditions also includes setting a reward function, specifically including: Compare the degree of change before and after the target value to obtain the change rate information; determining a reward value based on the rate of change information; Specifically, the reward function is shown as follows: ； in, are the weights of the two objectives respectively, ; is the maximum completion time, is the maximum completion time calculated at time t, SE(t) is the total energy consumption at time t, and SE(t-1) is the maximum total energy consumption calculated at time t-1.

9. The method according to claim 1, characterized in that: The energy consumption information in each processing step is calculated based on the production information of the workshop and the energy consumption information of equipment and facilities, including: Determine, according to the production information, a machine used in the workshop as a target machine for calculating energy consumption information; Establishing a work schedule for each of the target machines according to the production information; Dividing the processing procedures of the target machine according to the work schedule to obtain a plurality of initial states to be calculated; Merging the same processing steps in the initial state to be calculated to obtain a target state to be calculated; Calculate the energy consumption of each target in the pending state.

10. An energy-saving flexible workshop scheduling device considering human factors, characterized in that: The device includes a calculation module, a building module, a generation module and a solution module; wherein, The calculation module is used to calculate the energy consumption information in each processing step based on the production information of the workshop and the energy consumption information of the equipment and facilities, and to sum the energy consumption information in each processing step to obtain the total energy consumption information of the machine, wherein the total energy consumption information of the machine changes with the production information, and the degree of the change is affected by the energy consumption information of the equipment and facilities; The establishment module is used to establish a worker learning ability assessment model to assess the learning ability of each worker; The worker learning ability evaluation model includes a machine operation efficiency sub-model and a workpiece operation efficiency sub-model. The machine operation efficiency sub-model is used to evaluate the operation efficiency of each worker on the machine, and the workpiece operation efficiency sub-model is used to evaluate the workpiece operation efficiency of each worker. The learning ability is affected by the sum of the machine operation efficiency and the workpiece operation efficiency. The generation module is used to calculate the actual processing time of processing a workpiece on a machine based on the learning ability of each worker, taking the total energy consumption information of the machine and the minimum processing completion time of the machine as the objective function, and generating constraint conditions based on the actual processing time and the workshop processing timing rules; The solution module is used to solve the objective function based on the constraint conditions and output the optimal scheduling solution of the energy-saving flexible workshop.

Images