Multi-target production plan optimization method and system for digital factory

Abstract

The invention relates to the technical field of digital factory production management, and discloses a digital factory-oriented multi-target production plan optimization method and system, and the method comprises the steps: decomposing an enterprise order into a plurality of subtasks through a production task decomposition module, and extracting related information; the resource dynamic evaluation module collects data in real time to generate a resource state matrix; the multi-objective optimization algorithm module constructs a model and adopts an improved non-dominated sorting genetic algorithm to solve; the dynamic priority distribution module adjusts the priorities of the sub-tasks according to the real-time data; and the conflict resolution unit is used for solving resource allocation conflicts. The method can accurately decompose tasks, evaluate resources in real time, collaboratively optimize multiple targets, flexibly deal with abnormities, efficiently resolve conflicts, comprehensively improve the production efficiency and benefits of the digital factory, and effectively solve the complex problems in the production plan making and optimizing process of the digital factory.

Description

Technical Field

[0001] The present invention relates to the technical field of digital factory production management, and specifically to a multi-objective production plan optimization method and system for digital factories. Background Art

[0002] With the digital transformation of the manufacturing industry, digital factories have shown great advantages in aspects such as production efficiency and quality control, but at the same time, they also face complex production plan optimization problems. In the traditional production mode, the formulation of production plans often relies on experience or simple calculation methods, making it difficult to adapt to the complex and changeable production environment in digital factories. First, the diversity and complexity of production tasks increase. The types of enterprise orders are becoming increasingly rich, and the demand for personalized customization of products is increasing, resulting in extremely complex decomposition and arrangement of production tasks. Taking automobile manufacturing as an example, different vehicle models have diverse configurations, and the production tasks of components are intertwined. If tasks cannot be accurately decomposed and process constraints are not considered, it is easy to cause production chaos and extend the production cycle.

[0003] Secondly, the difficulty of resource management increases. In digital factories, there are a wide variety of production equipment, dynamic changes in material inventory, and complex human resource structures. The operating status of equipment is unstable and may malfunction at any time; material supply is affected by market fluctuations, making it difficult to accurately control inventory; there are significant differences in the skill levels and work efficiencies of human resources, and there are also working hour restrictions. For example, in the production process of an electronic manufacturing enterprise, if the production plan is not adjusted in time due to a sudden equipment failure, it may lead to material backlogs and order delays; at the same time, if human resources cannot be reasonably allocated, it will cause some employees to be overloaded while some employees are idle, reducing the overall production efficiency.

[0004] Furthermore, the problem of multi-objective conflicts in the production process is prominent. Enterprises usually pursue multiple production goals, such as shortening the production cycle, reducing energy consumption costs, and improving equipment utilization rate, but these goals often restrict each other. In chemical production, increasing the equipment operating speed to shorten the production cycle may lead to a significant increase in energy consumption, and at the same time, the equipment load is unbalanced, accelerating equipment wear.

[0005] In addition, traditional production plan optimization methods are difficult to cope with real-time changes in the production process. When production abnormal events occur, such as raw material quality problems and urgent order changes, traditional methods cannot quickly adjust the production plan and lack dynamic adaptability. In the garment manufacturing industry, if the fabric quality is unqualified, traditional production plans are difficult to make effective adjustments quickly, resulting in production stagnation and cost increase.

[0006] Existing production plan optimization technologies have many deficiencies when facing these challenges in digital factories. Some methods only consider single-objective optimization and cannot comprehensively balance multiple production objectives; some algorithms have high computational complexity and low solution efficiency, making it difficult to meet the real-time requirements of actual production; and some methods do not monitor and respond to the dynamic changes of resources in a timely manner, resulting in the disconnection between production plans and actual situations. Summary of the Invention

[0007] The purpose of the present invention is to provide a multi-objective production plan optimization method and system for digital factories to solve the problems raised in the above background technology.

[0008] To achieve the above purpose, the present invention provides the following technical solution: A multi-objective production plan optimization system for digital factories, the system includes:

[0009] A production task decomposition module, which is used to receive enterprise order data and decompose it into multiple independent production subtasks. Each subtask contains a task number i, , where k is a positive integer greater than 2, and extract the process constraint conditions and resource requirement type sets of each subtask;

[0010] A resource dynamic evaluation module, which is used to collect the operation status data of production equipment, the material inventory characteristic values, and the human resource availability parameters in the digital factory in real time, and generate a real-time resource status matrix;

[0011] A multi-objective optimization algorithm module, which is used to construct a multi-objective optimization model based on the decomposed subtask set and the resource status matrix. The multi-objectives include minimizing the production cycle, minimizing the energy consumption cost, and maximizing the equipment load balance degree;

[0012] A dynamic priority allocation module, which is used to calculate the dynamic priority weights of each subtask according to the real-time updated production exception event data and the equipment failure rate prediction value, and adjust the subtask execution sequence;

[0013] A conflict resolution unit, which is used to detect resource allocation conflicts and reset the time window of conflicting tasks by using a relaxation constraint algorithm;

[0014] The construction method of the multi-objective optimization model is as follows:

[0015] Define the decision variables as the equipment allocation plan and time scheduling sequence of the subtasks;

[0016] Establish a set of objective functions, including the total production cycle function, the total energy consumption function, and the equipment load variance function;

[0017] Introduce process constraints, resource capacity limitations, and task dependencies as constraint conditions;

[0018] The improved non - dominated sorting genetic algorithm is used to iteratively solve the multi - objective optimization model to generate a non - dominated solution set.

[0019] Preferably, the specific method for generating the real - time resource status matrix in the resource dynamic evaluation module is as follows:

[0020] Collect the real - time idle time window, remaining maintenance cycle and energy consumption efficiency parameters of production equipment to construct an equipment resource vector;

[0021] Collect the inventory, replenishment cycle and quality decay coefficient of materials to construct a material resource vector;

[0022] Collect the matching degree of personnel skills, working - hour limit and task - handling efficiency to construct a human - resource vector;

[0023] Aggregate the equipment, material and human - resource vectors by time slices to form a three - dimensional resource status matrix.

[0024] Preferably, the specific optimization steps of the improved non - dominated sorting genetic algorithm include:

[0025] When initializing the population, a heuristic rule based on process constraints is used to generate feasible solutions;

[0026] In the crossover operation, an equipment affinity factor is introduced, and a distribution plan with a high historical cooperation frequency of equipment is preferentially selected for gene exchange;

[0027] In the mutation operation, a dynamic mutation probability is adopted, and the mutation intensity is adaptively adjusted according to the convergence degree of the population;

[0028] The entropy weight method is used to assign weights to the non - dominated solution set to screen the Pareto - front optimal solutions.

[0029] Preferably, the specific method for calculating the dynamic priority weight in the dynamic priority allocation module is as follows:

[0030] Obtain the basic priority coefficient of each sub - task from the local database ;

[0031] Real - time collect the equipment failure prediction value and the order urgency parameter , where j represents the equipment number;

[0032] Calculate the real - time priority weight of sub - task i :

[0033]

[0034] Where represents the dependence degree of sub - task i on equipment j, and m represents the total number of production equipment in the digital factory.

[0035] Preferably, the specific steps of using the relaxation constraint algorithm in the conflict resolution unit include:

[0036] Identify the conflict task groups with overlapping resource occupancy, and extract the process paths and time windows of the conflict tasks;

[0037] Calculate the upper limit of the delay time and the cost penalty coefficient for each task;

[0038] Construct a conflict resolution objective function, with the optimization direction of minimizing the total delay penalty value, and use the branch and bound algorithm to solve the optimal time window adjustment scheme.

[0039] Preferably, the calculation method for maximizing the equipment load balance is:

[0040] Statistical total working hours of each device within the scheduling period and idle time ;

[0041] Calculate the equipment load balance :

[0042]

[0043] Wherein, is the average working hours of the device, is the average idle time of the device, and the denominator normalizes the load difference value through the total time dimension.

[0044] Preferably, the method for extracting the material inventory characteristic value includes:

[0045] Obtain the batch code and the warehousing timestamp of the material through the RFID sensor;

[0046] Combined with the environmental temperature and humidity monitoring data, calculate the real-time quality decay rate of the material :

[0047]

[0048] Where is the reference decay coefficient, is the real-time humidity, is the ideal humidity, is the adjustment parameter.

[0049] Preferably, the digital method for the process constraint conditions is:

[0050] Analyze the operation sequence and equipment dependency relationship in the CAD process drawing to generate a directed acyclic graph;

[0051] Map the nodes in the graph to subtasks, and the edge weight represents the task switching time cost;

[0052] Generate an initial feasible scheduling sequence through a topological sorting algorithm.

[0053] Preferably, the optimization method for minimizing the energy consumption cost includes:

[0054] Collect the power curves of the device at different load rates and fit a piecewise linear energy consumption function.

[0055] Introduce time-of-use electricity price parameters and convert the energy consumption cost into a time-dependent weighted function.

[0056] Embed Lagrange multipliers in the improved non-dominated sorting genetic algorithm to dynamically relax the energy consumption constraints.

[0057] The present invention also includes a multi-objective production plan optimization method for a digital factory, including the following steps:

[0058] Step 1: Use the production task decomposition module to receive enterprise order data, decompose it into multiple independent production sub-tasks, each sub-task includes a task number i, , where k is a positive integer greater than 2, and extract the process constraint conditions and resource requirement type sets of each sub-task.

[0059] Step 2: Through the resource dynamic evaluation module, collect the operation status data of production equipment, the material inventory characteristic values, and the human resource availability parameters in the digital factory in real time, and generate a real-time resource status matrix.

[0060] Step 3: Use the multi-objective optimization algorithm module to construct a multi-objective optimization model based on the decomposed sub-task set and the resource status matrix. Among them, define the decision variables as the equipment allocation plan and time scheduling sequence of the sub-tasks, establish a set of objective functions, including the total production cycle function, the total energy consumption function, and the equipment load variance function, introduce process constraints, resource capacity limitations, and task dependencies as constraint conditions, and use the improved non-dominated sorting genetic algorithm to iteratively solve the multi-objective optimization model to generate a non-dominated solution set. The optimization objectives include minimizing the production cycle, minimizing the energy consumption cost, and maximizing the equipment load balance degree.

[0061] Step 4: Use the dynamic priority allocation module to calculate the dynamic priority weights of each sub-task according to the real-time updated production exception event data and the equipment failure rate prediction value, and adjust the sub-task execution sequence.

[0062] Step 5: Use the conflict resolution unit to detect resource allocation conflicts and use the relaxation constraint algorithm to reset the time window of the conflicting tasks.

[0063] Compared with the prior art, the beneficial effects of the present invention are:

[0064] The production task decomposition module of the present invention can accurately decompose enterprise order data into multiple independent production sub-tasks, and extract the corresponding process constraint conditions and resource requirement type sets. This makes the production task arrangement more organized, and can carry out targeted planning according to the characteristics and requirements of different sub-tasks. Taking a machinery manufacturing enterprise as an example, a large equipment manufacturing order can be carefully decomposed into sub-tasks such as component processing, component assembly, and overall commissioning. The process path and resource requirements of each sub-task are clear and definite, avoiding the chaos of task arrangement, improving the accuracy and efficiency of production, and ensuring the orderly progress of the entire production process.

[0065] The resource dynamic assessment module generates a real-time resource status matrix by collecting the operation status data of production equipment, the characteristic values of material inventory, and the availability parameters of human resources in real time. This function enables the enterprise to grasp the dynamic changes of resources in real time and adjust the production plan in a timely manner. For example, in the process of electronic product manufacturing, when the inventory of a certain key material is lower than the safety threshold, the system can quickly perceive and adjust the production sequence, and preferentially arrange other tasks with low dependence on this material, avoiding production stagnation caused by material shortage, improving the utilization efficiency of resources, and reducing inventory backlog and waste.

[0066] The multi-objective optimization model constructed by the multi-objective optimization algorithm module comprehensively considers the minimization of production cycle, the minimization of energy consumption cost, and the maximization of equipment load balance. In actual production, taking a steel production enterprise as an example, by reasonably arranging the equipment operation time and production task allocation through the optimization model, not only the production cycle of the product is shortened, the order delivery time is reduced, but also the energy consumption cost is reduced. At the same time, the load of each equipment is more balanced, the service life of the equipment is extended, and the equipment maintenance cost is reduced. This way of multi-objective collaborative optimization effectively balances the contradictions between different production objectives of the enterprise and enhances the comprehensive competitiveness of the enterprise.

[0067] The dynamic priority allocation module calculates the dynamic priority weights of each sub-task according to the real-time updated production exception event data and the predicted values of equipment failure rates, and adjusts the execution sequence of sub-tasks. During the production process, if there are potential faults in an important equipment, the system will automatically increase the priority of the key sub-tasks that depend on this equipment, arrange production in advance, and at the same time adjust the order of other tasks to ensure the continuity and stability of production. This enables the enterprise to quickly respond in the face of emergencies, ensure the smooth progress of production, and reduce the losses caused by production exceptions.

[0068] The conflict resolution unit uses a relaxation constraint algorithm to detect resource allocation conflicts and reset the time windows of conflicting tasks. In a complex production environment, resource competition is inevitable. For example, multiple tasks may require a certain special device simultaneously. This unit can quickly identify conflicts, construct a conflict resolution objective function by calculating the upper limit of the delay time and the cost penalty coefficient of each task, and use the branch and bound algorithm to find the optimal time window adjustment plan, solve resource allocation conflicts, ensure the orderly execution of the production plan, and improve the feasibility and stability of the production plan.

[0069] From the precise extraction of the characteristic values of material inventory, to the digital processing of process constraint conditions, to the optimization method of energy consumption costs, and to the reasonable construction of the human resource vector, the present invention has been comprehensively optimized in the details of each production link. For example, by accurately calculating the real-time quality decay rate of materials and reasonably arranging the usage order, the product quality is guaranteed; by using the optimization steps of the improved non-dominated sorting genetic algorithm, the solution efficiency and accuracy of the multi-objective optimization model are improved; by collecting the physiological index data of employees in real time to adjust the human resource efficiency parameters, the efficient utilization of human resources is realized. These detailed optimizations comprehensively improve the overall production efficiency and benefits of the digital factory, creating greater value for the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is the working principle diagram of the multi-objective production plan optimization system described in the present invention;

[0071] Figure 2 is the optimization flow chart of the improved non-dominated sorting genetic algorithm;

[0072] Figure 3 is the flow chart of the conflict resolution unit using the relaxation constraint algorithm. DETAILED DESCRIPTION OF THE INVENTION

[0073] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0074] Please refer to Figures 1-3 , the present invention provides a multi-objective production plan optimization system for a digital factory, and the system includes:

[0075] A production task decomposition module: After receiving enterprise order data, it decomposes the data into multiple independent production sub-tasks, and assigns a unique task number i to each sub-task. , where k is a positive integer greater than 2. Meanwhile, the process constraint conditions and the set of resource requirement types for each subtask are extracted. For example, when an enterprise receives an order to produce a batch of auto parts, this module will decompose the order into multiple subtasks such as casting, machining, and assembly according to the production process of the parts, and clarify the sequential requirements in terms of technology (process constraint conditions) for each subtask, as well as the types of equipment, materials, and human resources required (set of resource requirement types).

[0076] Resource dynamic assessment module: Real-time collect the operation status data of production equipment, the characteristic values of material inventory, and the availability parameters of human resources in the digital factory, and generate a real-time resource status matrix. This process provides an accurate basis of resource information for the subsequent formulation of production plans, ensuring the feasibility of the plans.

[0077] Multi-objective optimization algorithm module: Based on the decomposed set of subtasks and the resource status matrix, construct a multi-objective optimization model. Among them, the multi-objectives include minimizing the production cycle, minimizing the energy consumption cost, and maximizing the equipment load balance. By constructing a reasonable model and selecting an effective algorithm, find the optimal production plan that meets multiple objectives. The specific construction method is as follows: Define the decision variables as the equipment allocation plan and time scheduling sequence of subtasks; establish a set of objective functions, including the total production cycle function, the total energy consumption function, and the equipment load variance function; introduce process constraints, resource capacity limitations, and task dependencies as constraint conditions; use an improved non-dominated sorting genetic algorithm to iteratively solve the multi-objective optimization model and generate a non-dominated solution set.

[0078] Dynamic priority allocation module: According to the real-time updated production exception event data and the predicted value of equipment failure rate, calculate the dynamic priority weights of each subtask and adjust the execution sequence of subtasks. In this way, in the face of unexpected situations during the production process, important or urgent tasks can be arranged first to ensure the smooth progress of production.

[0079] Conflict resolution unit: Detect resource allocation conflicts and use a relaxation constraint algorithm to reset the time window for conflicting tasks. When multiple subtasks compete for the same resource and cause conflicts, this unit can resolve the conflicts through a reasonable algorithm to ensure the orderly execution of the production plan.

[0080] The following further illustrates the implementation of the present invention in combination with Embodiments 1 to 5.

[0081] Embodiment 1:

[0082] This embodiment elaborates in detail the specific implementation method for the resource dynamic assessment module to generate a real-time resource status matrix, and its role is to accurately obtain and integrate the real-time status information of various resources in the digital factory.

[0083] In terms of production equipment, collect the real-time idle time window, remaining maintenance cycle, and energy consumption efficiency parameters of the production equipment. For example, for a numerically controlled machine tool, obtain the current unoccupied time interval (real-time idle time window) in real time through equipment sensors, and at the same time obtain the remaining duration until the next maintenance (remaining maintenance cycle) from the equipment management system, as well as the energy consumption efficiency data (energy consumption efficiency parameters) of the equipment under different working modes. Use these data to construct an equipment resource vector, which can comprehensively reflect information such as the available time, maintenance requirements, and energy consumption characteristics of the equipment.

[0084] For material resources, collect the inventory quantity, replenishment cycle, and quality decay coefficient of the materials. Obtain the current inventory quantity of the materials (inventory quantity) through the warehouse management system, and obtain the replenishment cycle information of the materials from the supplier. At the same time, combine the environmental temperature and humidity monitoring data to calculate the real-time quality decay rate , and the calculation formula is:

[0085]

[0086] where is the reference decay coefficient, is the real-time humidity, is the ideal humidity, is the adjustment parameter. Construct a material resource vector based on these data to accurately reflect the quantity, supply timeliness, and quality changes of the materials.

[0087] In terms of human resources, collect the matching degree of personnel skills, working hour limits, and task processing efficiency. Obtain the matching degree between the employee skills and the required skills of each subtask (matching degree of personnel skills) through the employee skills management system, determine the working hour limits of employees according to labor laws and enterprise regulations, and at the same time record the actual processing efficiency data of employees in previous similar tasks. Construct a human resource vector, which comprehensively reflects the ability, time limit, and efficiency level of employees to complete tasks.

[0088] Finally, aggregate the equipment, material, and human resource vectors by time slice to form a three-dimensional resource status matrix. For example, taking one hour as a time slice, integrate the equipment, material, and human resource vector information collected within this time slice to generate a three-dimensional matrix. This matrix is dynamically updated in the time dimension, providing real-time and comprehensive resource status information for the multi-objective optimization algorithm module, enabling the production plan to be accurately formulated according to the actual situation of the resources.

[0089] Example 2:

[0090] When initializing the population, heuristic rules based on process constraints are used to generate feasible solutions. Taking the production of automotive parts as an example, when formulating the initial population of the production plan, according to the operation sequence specified in the process drawings, such as casting first, then machining, and finally assembly, task combinations that conform to this sequence are preferentially arranged. At the same time, considering the process requirements of the equipment, such as certain high-precision machining tasks can only be carried out on specific high-precision equipment, it is ensured that the generated initial solutions meet the process constraint conditions, avoiding the emergence of infeasible production plan schemes, and improving the quality of the initial solutions and the convergence speed of the algorithm.

[0091] In the crossover operation, an equipment affinity factor is introduced, and the allocation scheme with a high historical cooperation frequency of equipment is preferentially selected for gene exchange. For example, in a certain digital factory, equipment A and equipment B have jointly completed specific types of production tasks many times in the past, and their historical cooperation frequency is relatively high. When performing the crossover operation, for the gene segment related to equipment allocation, if it contains the allocation information of equipment A and equipment B, and compared with the gene segments of other equipment combinations, the equipment affinity factor corresponding to this combination is higher, then this combination is preferentially selected for gene exchange. This can make full use of the cooperation experience between equipment, improve the quality of the generated new solutions, make the production plan more reasonable in terms of equipment collaborative work, and reduce the time and cost losses caused by equipment switching.

[0092] In the mutation operation, a dynamic mutation probability is adopted, and the mutation intensity is adaptively adjusted according to the convergence degree of the population. When the algorithm is in the initial stage of iteration, the diversity of the population is relatively high and the convergence speed is slow, the mutation probability is appropriately increased to encourage the algorithm to explore a wider solution space and avoid falling into local optimal solutions. As the iteration progresses, if it is found that the convergence speed of the population is too fast and premature convergence may occur, the mutation probability is reduced at this time, so that the algorithm performs a fine search near the current better solution to further optimize the quality of the solution. For example, by setting a convergence index, such as when the change range of the optimal solution in the population for several consecutive generations is less than a certain threshold, it is determined that the population tends to converge, and thus the mutation probability is adjusted.

[0093] The entropy weight method is used to assign weights to the non-dominated solution set and screen the Pareto front optimal solutions. The entropy weight method is an objective weight assignment method. By calculating the difference degree of each objective in the non-dominated solution set among different solutions, the weight of each objective is determined. For example, for the three objectives of production cycle, energy consumption cost, and equipment load balance degree, the entropy weight method will calculate their respective weights according to their distribution in the non-dominated solution set. Then, each solution in the non-dominated solution set is comprehensively evaluated according to these weights, and the solution with the optimal comprehensive score is selected as the Pareto front optimal solution, so that the finally obtained production plan scheme can achieve a better balance among multiple objectives.

[0094] Example 3:

[0095] This embodiment details the specific process of the dynamic priority allocation module calculating the dynamic priority weights. Its function is to reasonably adjust the priorities of each subtask according to the real-time situation in the production process, ensure that the production plan can respond promptly to abnormal events and equipment failure risks, and guarantee the efficient progress of production.

[0096] First, obtain the basic priority coefficients of each subtask from the local database. . For example, in an electronic product production project, for the production subtasks of key components, since they have a greater impact on the quality and performance of the entire product, relatively high basic priority coefficients are preset in the database; while for some auxiliary subtasks, such as the task of preparing packaging materials, the basic priority coefficients are relatively low. These basic priority coefficients are preset according to factors such as the importance of the subtask and its contribution to the overall value of the product.

[0097] Next, collect the equipment failure prediction values in real time and the order urgency parameters , where j represents the equipment number. Through the equipment failure monitoring system and data analysis model, obtain the predicted probability of each equipment failing in the next period of time (equipment failure prediction value ). At the same time, calculate the order urgency parameters according to factors such as the delivery time requirement of the customer order and the order amount. . For example, if the delivery time of a certain order is approaching and the order amount is large, then the corresponding order urgency parameter

[0098] Then, calculate the real-time priority weight of subtask i , and the calculation formula is:

[0099] where represents the dependence degree of subtask i on equipment j, and m represents the total number of production equipment in the digital factory. For example, in a certain machinery manufacturing factory, subtask i is the processing task of a certain key component, and this task highly depends on equipment j for high-precision processing. Therefore, the dependence degree of subtask i on equipment j is relatively high. Assume that the calculated equipment failure prediction value of equipment j is 0.2, the basic priority coefficient of subtask i is 0.8, the order urgency parameter is 1.5, and subtask i only depends on equipment j. Then the real-time priority weight of subtask i. According to the calculated real-time priority weights, dynamically adjust the execution sequence of subtasks, and give priority to arranging subtasks with high priority weights for production to improve the flexibility of the production plan and the ability to respond to emergencies.

[0100] Example 4:

[0101] In the stage of identifying conflicting task groups, by monitoring the resource allocation situation in real time, identify the conflicting task groups with overlapping resource occupancy, and extract the process paths and time windows of the conflicting tasks. For example, in a clothing production workshop, both Task A and Task B need to use the same sewing machine during the same time period, which constitutes a conflicting task group. At the same time, obtain the entire process paths of Task A and Task B from raw material preparation to finished product completion, as well as the time window information of their respective planned start and end times.

[0102] Calculate the upper limit of the delay time and the cost penalty coefficient for each task. For each conflicting task, determine the upper limit of the delay time according to its process characteristics and the time requirements of subsequent tasks. For example, if Task A is delayed by more than 3 hours, it will affect the timely progress of multiple subsequent tasks, then the upper limit of the delay time of Task A is 3 hours. The cost penalty coefficient is determined according to the impact degree of task delay on production efficiency, cost increase, etc. For example, the delay in production of Task A will lead to an increase in the cost of raw material backlog and the cost of equipment idleness. After comprehensively considering these factors, the cost penalty coefficient of Task A is determined to be a fine of 500 yuan per hour of delay.

[0103] Construct a conflict resolution objective function, with the optimization direction of minimizing the total delay penalty value, and use the branch and bound algorithm to solve the optimal time window adjustment scheme. The branch and bound algorithm will search all possible time window adjustment schemes, continuously expand the solution space by branching, and prune according to the objective function value, discarding those branches that cannot obtain the optimal solution, and finally find the time window adjustment scheme that minimizes the total delay penalty value. For example, assume that there are Task A, Task B, and Task C in the conflicting task group. After algorithm calculation, it is found that delaying the start time of Task A by 1 hour, keeping the time window of Task B unchanged, and advancing the start time of Task C by 0.5 hour can minimize the total delay penalty value. Then reset the time windows of the conflicting tasks according to this scheme to resolve the resource allocation conflict.

[0104] Example 5:

[0105] In terms of maximizing the equipment load balance degree, count the total working hours of each equipment within the scheduling cycle and the idle hours . For example, within a one-month scheduling cycle, count that the total working hours of Equipment 1 is 150 hours and the idle hours is 50 hours; the total working hours of Equipment 2 is 130 hours and the idle hours is 70 hours, etc. Then calculate the equipment load balance degree B, and the calculation formula is:

[0106] where, is the average working hours of the equipment, is the average idle duration of the equipment, and the denominator normalizes the load difference value through the total time dimension. Suppose there are 3 pieces of equipment in a digital factory. The working hours of Equipment 1 hours, the working hours of Equipment 2 hours, the working hours of Equipment 3 hours, and the average working hours hours; the idle duration of Equipment 1 hours, the idle duration of Equipment 2 hours, the idle duration of Equipment 3 hours, and the average idle duration hours.

[0107] Then the equipment load balance degree . By improving the equipment load balance degree, overuse or idleness of the equipment can be reduced, the service life of the equipment can be extended, and the overall utilization rate of the equipment can be improved.

[0108] In terms of extracting the characteristic values of material inventory, the batch code and warehousing timestamp of the material are obtained through RFID sensors. For example, in a food processing factory, RFID sensors can be used to quickly and accurately obtain the coding information of each batch of raw materials and the warehousing time of this batch of raw materials. Combining with the environmental temperature and humidity monitoring data, calculate the real-time quality decay rate of the material . By accurately obtaining the characteristic values of material inventory, the quality change of the material can be better grasped, production can be reasonably arranged, and the product quality can be prevented from being affected by using deteriorated materials.

[0109] For the digitization of process constraint conditions, analyze the process sequence and equipment dependency relationship in the CAD process drawings to generate a directed acyclic graph. For example, in the process drawings of machining mechanical parts, the sequence of rough machining, semi-finishing, finishing and other processes, as well as the equipment required for each process, are clearly specified. Convert this information into a directed acyclic graph, where the nodes in the graph represent each subtask, and the edge weights represent the task switching time cost. Then generate an initial feasible scheduling sequence through the topological sorting algorithm to ensure that the production tasks are arranged in a reasonable process sequence and improve the fluency of production.

[0110] In terms of minimizing energy consumption costs, the power curves of the acquisition device at different load rates are collected to fit a piecewise linear energy consumption function. For example, for an injection molding machine, the power data at different injection volumes (corresponding to different load rates) are collected, and a piecewise linear energy consumption function is fitted through mathematical methods. The time-of-use electricity price parameter is introduced to convert the energy consumption cost into a time-dependent weighted function. Assuming that the electricity price is high during the peak electricity consumption period during the day and low during the low valley period at night, the energy consumption cost is weighted according to the electricity prices at different times. The Lagrange multiplier is embedded in the improved non-dominated sorting genetic algorithm to dynamically relax the energy consumption constraint, so as to minimize the energy consumption cost as much as possible while meeting the production requirements.

[0111] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "including", "comprising" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0112] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-objective production planning optimization system for a digital factory, characterized in that, The system includes: The production task decomposition module is used to receive enterprise order data and decompose it into multiple independent production subtasks. Each subtask contains a task number i, , where k is a positive integer greater than 2, and extract the process constraint conditions and the set of resource requirement types for each subtask; A resource dynamic evaluation module, which is used to collect the operation status data of production equipment, the material inventory characteristic values, and the human resource availability parameters in a digital factory in real time, and generate a real-time resource status matrix; A multi-objective optimization algorithm module, which is used to construct a multi-objective optimization model based on the decomposed sub-task set and the resource status matrix. The multi-objectives include minimizing the production cycle, minimizing the energy consumption cost, and maximizing the equipment load balance degree; A dynamic priority allocation module, which is used to calculate the dynamic priority weights of each sub-task according to the real-time updated production exception event data and the equipment failure rate prediction value, and adjust the sub-task execution sequence; A conflict resolution unit, which is used to detect resource allocation conflicts and reset the time window of conflict tasks by using a relaxation constraint algorithm; The construction method of the multi-objective optimization model is as follows: Define the decision variables as the equipment allocation scheme and time scheduling sequence of sub-tasks; Establish a set of objective functions, including the total production cycle function, the total energy consumption function, and the equipment load variance function; Introduce process constraints, resource capacity limitations, and task dependencies as constraint conditions; Use an improved non-dominated sorting genetic algorithm to iteratively solve the multi-objective optimization model and generate a non-dominated solution set.

2. The multi-objective production plan optimization system for a digital factory according to claim 1, characterized in that, The specific method for generating the real-time resource status matrix in the resource dynamic evaluation module is as follows: Collect the real-time idle time window, remaining maintenance cycle, and energy consumption efficiency parameters of production equipment, and construct an equipment resource vector; Collect the inventory quantity, replenishment cycle, and quality decay coefficient of materials, and construct a material resource vector; Collect the matching degree of personnel skills, working hour limitations, and task processing efficiency, and construct a human resource vector; Aggregate the equipment, material, and human resource vectors by time slices to form a three-dimensional resource status matrix.

3. The multi-objective production planning optimization system for a digital factory according to claim 2, wherein The specific optimization steps of the improved non-dominated sorting genetic algorithm include: When initializing the population, use a heuristic rule based on process constraints to generate feasible solutions; Introduce an equipment affinity factor in the crossover operation, and preferentially select the allocation scheme with a high historical cooperation frequency of equipment for gene exchange; Adopt a dynamic mutation probability in the mutation operation, and adaptively adjust the mutation intensity according to the population convergence degree; Use the entropy weight method to assign weights to the non-dominated solution set and screen the Pareto front optimal solutions.

4. The multi-objective production planning optimization system for a digital factory according to claim 3, wherein The specific method for calculating the dynamic priority weights in the dynamic priority allocation module is as follows: Obtain the basic priority coefficients of each subtask from the local database ; Real-time acquisition of equipment failure prediction values and order urgency parameters , where j represents the equipment number; Calculate the real-time priority weight of subtask i : Among them represents the degree of dependence of subtask i on device j, and m represents the total number of production devices in the digital factory.

5. The multi-objective production plan optimization system for a digital factory according to claim 4, wherein The specific steps of using the relaxation constraint algorithm in the conflict resolution unit include: Identify the conflict task group with overlapping resource occupancy, and extract the process path and time window of the conflict tasks; Calculate the upper limit of the delay time and the cost penalty coefficient of each task; Construct a conflict resolution objective function, with the direction of minimizing the total delay penalty value as the optimization direction, and use the branch and bound algorithm to solve the optimal time window adjustment scheme.

6. The multi-objective production plan optimization system for a digital factory according to claim 5, characterized in that The calculation method for maximizing the equipment load balance degree is as follows: Statistically calculate the total working hours of each device within the scheduling period and the idle hours ; Calculate the load balance of computing devices : Among them, is the average working duration of the device, is the average idle duration of the device, and the denominator is the normalized load difference value through the total time dimension.

7. The multi-objective production plan optimization system for a digital factory according to claim 6, characterized in that The extraction method of the material inventory characteristic values includes: Obtain the batch code and warehousing timestamp of materials through RFID sensors; Calculate the real-time quality decay rate of the material by combining the environmental temperature and humidity monitoring data : wherein is the reference attenuation coefficient, is the real-time humidity, is the ideal humidity, is the adjustment parameter.

8. The multi-objective production planning optimization system for a digital factory according to claim 7, characterized in that The digitalization method of the process constraint conditions is as follows: Analyze the process sequence and equipment dependencies in the CAD process drawings to generate a directed acyclic graph; Map the nodes in the graph to sub-tasks, and the edge weights represent the task switching time cost; Generate an initial feasible scheduling sequence through a topological sorting algorithm.

9. The multi-objective production plan optimization system for a digital factory according to claim 8, characterized in that The optimization method for minimizing the energy consumption cost includes: Collect the power curves of the device at different load rates and fit a piecewise linear performance energy consumption function; Introduce time-of-use electricity price parameters and convert the energy consumption cost into a time-dependent weighted function; Embed Lagrange multipliers in the improved non-dominated sorting genetic algorithm to perform dynamic relaxation processing on the energy consumption constraints.

10. A multi-objective production plan optimization method for a digital factory using the system described in claim 9, characterized in that, It includes the following steps: Step 1: Use the production task decomposition module to receive enterprise order data, decompose it into multiple independent production subtasks, each subtask contains a task number i, , where k is a positive integer greater than 2, and extract the process constraint conditions and resource requirement type sets of each subtask; Step 2: Use the resource dynamic evaluation module to collect the operation status data of production equipment, material inventory characteristic values, and human resource availability parameters in the digital factory in real time, and generate a real-time resource status matrix; Step 3: Use the multi-objective optimization algorithm module to construct a multi-objective optimization model based on the decomposed sub-task set and the resource status matrix. Among them, define the decision variables as the device allocation scheme and time scheduling sequence of the sub-tasks, establish a set of objective functions, including the total production cycle function, the total energy consumption function, and the device load variance function, introduce process constraints, resource capacity limits, and task dependencies as constraint conditions, and use the improved non-dominated sorting genetic algorithm to iteratively solve the multi-objective optimization model to generate a non-dominated solution set. The optimization objectives include minimizing the production cycle, minimizing the energy consumption cost, and maximizing the device load balance; Step 4: Use the dynamic priority allocation module to calculate the dynamic priority weights of each sub-task according to the real-time updated production anomaly event data and the predicted device failure rate, and adjust the sub-task execution sequence; Step 5: Use the conflict resolution unit to detect resource allocation conflicts and use the relaxation constraint algorithm to reset the time window of the conflicting tasks.

Images