A production scheduling method, apparatus and device

Abstract

The application provides a production scheduling method, device and equipment. The method comprises the following steps: acquiring real-time production data through an intelligent manufacturing system; determining constraint conditions of a target optimization model according to the real-time production data; acquiring multiple dimension production targets; determining a composite target function according to the multiple dimension production targets; determining the target optimization model according to the constraint conditions and the composite target function; iteratively solving the target optimization model through at least one optimization model to obtain a target solution set; and controlling a production scheduling module to output a target production schedule according to the target solution set. The application improves the quality and efficiency of production line scheduling and enhances the adaptability to dynamic production environment.

Description

Technical Field

[0001] This invention relates to the field of industrial production technology, and also to a production scheduling method, apparatus and equipment. Background Technology

[0002] In modern manufacturing, production planning and scheduling (the process of optimizing the allocation of time, equipment, and personnel for production tasks under limited resource constraints, aiming to maximize efficiency, minimize costs, and ensure on-time delivery) is a crucial link affecting production efficiency, resource utilization, and cost control. Traditional scheduling methods do not consider the capacity loss caused by outsourced order production and circulation, and cannot provide strong technical support for scheduling planning that involves collaboration between multiple processes inside and outside the enterprise. Furthermore, they use single constraint algorithms or low-heuristic algorithms, which generally suffer from problems such as local convergence, high computational complexity, insufficient adaptability, and difficulty in multi-objective optimization, making it difficult to achieve optimal results in real industrial environments. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a production scheduling method, apparatus and equipment to improve the efficiency of production planning and scheduling.

[0004] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0005] A first aspect of the present invention provides a production scheduling method, comprising:

[0006] Obtain real-time production data through intelligent manufacturing systems;

[0007] Based on the real-time production data, determine the constraints of the target optimization model;

[0008] Obtain production targets from multiple dimensions;

[0009] Based on the production objectives across multiple dimensions, a composite objective function is determined;

[0010] Based on the constraints and the composite objective function, determine the objective optimization model;

[0011] The target optimization model is iteratively solved using at least one optimization model to obtain the target solution set;

[0012] Based on the target solution set, the production scheduling module is controlled to output the target production schedule.

[0013] Optionally, based on the real-time production data, the constraints of the target optimization model are determined, including:

[0014] Based on the real-time production data, the following data are obtained: process time data for each piece of equipment, available machine time data for each piece of equipment, start-up time data for each process, material availability time data and safety buffer period data, sequence classification data for preset key processes and allowable adjustment range data, skill level data for personnel and skill requirement level data for each process, historical utilization rate data and fluctuation range limit data for each piece of equipment, historical energy consumption data per unit of output value and energy consumption reduction rate target data.

[0015] Based on the process time data and available machine time data of each piece of equipment, the first constraint condition is determined;

[0016] Based on the start time data, material availability time data, and safety buffer period data for each process, the second constraint condition is determined.

[0017] Based on the preset key process sequence classification data and the allowable adjustment range data, the third constraint condition is determined;

[0018] Based on the skill level data of the personnel and the skill requirement level data of each process, the fourth constraint condition is determined;

[0019] Based on the historical utilization data and fluctuation range limit data of each device, the fifth constraint condition is determined;

[0020] Based on the historical energy consumption data per unit of output and the target data for energy consumption reduction rate, the sixth constraint condition is determined.

[0021] Based on the first constraint, the second constraint, the third constraint, the fourth constraint, the fifth constraint, and the sixth constraint, the constraints of the target optimization model are determined.

[0022] Optionally, based on the multiple dimensions of the production objective, a composite objective function is determined, including:

[0023] Based on the production objectives across multiple dimensions, determine the core objective function;

[0024] A composite objective function is determined based on the preset auxiliary objective function and the core objective function.

[0025] Optionally, a composite objective function is determined based on a preset auxiliary objective function and the core objective function, including:

[0026] Obtain a preset auxiliary objective function; the preset auxiliary objective function includes maximizing equipment utilization.

[0027] Based on the preset auxiliary objective function and the core objective function, a composite objective function is determined; the composite objective function is min(C-λU);

[0028] Where C = ∑(P + S + E);

[0029] Wherein, min(C-λU) is the composite objective function, C is the total cost in the core objective function, λ is the weighting coefficient, U is the maximum equipment utilization rate, P is the delivery deviation penalty cost, S is the changeover cost, and E is the energy consumption cost.

[0030] Optionally, the objective optimization model is iteratively solved using at least one optimization model to obtain the objective solution set, including:

[0031] Based on real-time monitoring data of the production equipment, determine the equipment disturbance events;

[0032] Identify order-level disturbance events based on real-time order data;

[0033] Based on real-time personnel and material supply data, identify resource-level disturbance events;

[0034] Based on the equipment disturbance event, the order-level disturbance event, the resource-level disturbance event, and the multi-dimensional preset performance indicators, determine at least one combination of optimization models to obtain an optimized combination model;

[0035] The objective optimization model is iteratively solved based on the aforementioned optimization combination model to obtain the objective solution set.

[0036] Optionally, when the optimization combination model includes a first optimization sub-model, a second optimization sub-model, and a third optimization sub-model, the objective optimization model is iteratively solved according to the optimization combination model to obtain the objective solution set, including:

[0037] Obtain a preset random initial solution set;

[0038] A global search is performed based on the first optimization sub-model and the preset random initial solution set to obtain the first optimal solution;

[0039] A local search is performed based on the second optimization sub-model and the first optimal solution to obtain the second optimal solution;

[0040] The target solution set is obtained by perturbation processing based on the third optimization sub-model and the second optimal solution.

[0041] Optionally, based on the target solution set, the production scheduling module is controlled to output a target production schedule, including:

[0042] Based on the target solution set, a production task list is determined; the production task list includes the quantity, priority, and delivery date of each production task.

[0043] Based on the production task list, resource allocation is determined; the resource allocation includes equipment resources, human resources, and material resources.

[0044] Based on the production task list and the resource allocation, determine the target production schedule;

[0045] The production scheduling module controls the output of the target production schedule.

[0046] A second aspect of the present invention provides a production scheduling apparatus, comprising:

[0047] The first acquisition module is used to acquire real-time production data through the intelligent manufacturing system;

[0048] The first determining module is used to determine the constraints of the target optimization model based on the real-time production data.

[0049] The second acquisition module is used to acquire production targets from multiple dimensions;

[0050] The second determining module is used to determine the composite objective function based on the multiple dimensions of the production target;

[0051] The third determining module is used to determine the target optimization model based on the constraints and the composite objective function;

[0052] The processing module is used to iteratively solve the target optimization model using at least one optimization model to obtain the target solution set;

[0053] The control module is used to control the production scheduling module to output the target production schedule based on the target solution set.

[0054] A third aspect of the present invention provides a computing device, comprising: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described in the first aspect.

[0055] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method as described in the first aspect.

[0056] The above-described solution of the present invention has at least the following beneficial effects:

[0057] The above-mentioned solution of the present invention acquires real-time production data through an intelligent manufacturing system and uses it to determine the constraints of the target optimization model; it acquires production targets in multiple dimensions and uses them to determine a composite objective function; then, based on the constraints and the composite objective function, it determines the target optimization model, and iteratively solves the target optimization model using at least one optimization model to obtain the target solution set; finally, based on the target solution set, it controls the production scheduling module to output the target production schedule, thereby improving the quality and efficiency of production line scheduling and enhancing the adaptability to dynamic environments. Attached Figure Description

[0058] Figure 1 This is a flowchart illustrating the production scheduling method in an embodiment of the present invention;

[0059] Figure 2 This is a schematic diagram of the process of solving the target optimization model in an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of the production scheduling device in an embodiment of the present invention. Detailed Implementation

[0061] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the invention and to fully convey the scope of the invention to those skilled in the art.

[0062] like Figure 1 As shown, an embodiment of the present invention proposes a production scheduling method, including the following steps:

[0063] Step 101: Obtain real-time production data through the intelligent manufacturing system;

[0064] Step 102: Determine the constraints of the target optimization model based on the real-time production data;

[0065] Step 103: Obtain production targets across multiple dimensions;

[0066] Step 104: Determine the composite objective function based on the multiple dimensions of the production objective;

[0067] Step 105: Determine the target optimization model based on the constraints and the composite objective function;

[0068] Step 106: Iteratively solve the target optimization model using at least one optimization model to obtain the target solution set;

[0069] Step 107: Based on the target solution set, control the production scheduling module to output the target production schedule.

[0070] The production scheduling method proposed in this invention acquires real-time production data through an intelligent manufacturing system and uses this data to determine the constraints of the target optimization model. It acquires production targets across multiple dimensions and uses this data to determine a composite objective function. Based on the constraints and the composite objective function, it determines the target optimization model and iteratively solves the target optimization model using at least one optimization model to obtain the target solution set. Finally, based on the target solution set, it controls the production scheduling module to output the target production schedule, thereby improving the quality and efficiency of production line scheduling and enhancing its adaptability to dynamic environments.

[0071] In an optional embodiment of the present invention, step 101 includes:

[0072] The first data is obtained in real time from the manufacturing execution system; the first data includes work order data, production progress data, and equipment utilization data.

[0073] Obtain real-time secondary data from the Enterprise Resource Planning (ERP) system; the secondary data includes bill of materials data, inventory data, and production plan data;

[0074] Real-time third-party data is acquired from Internet of Things (IoT) devices via sensors; the third-party data includes device status data and production environment parameter data.

[0075] The first data, the second data, and the third data are preprocessed to obtain real-time production data.

[0076] Specifically, a smart manufacturing system can include a Manufacturing Execution System (MES), an Enterprise Resource Planning (ERP) system, and Internet of Things (IoT) devices. The MES system collects production data in real time, including work order data, production progress data, and equipment utilization data. Through its equipment data acquisition functions, barcode scanning capabilities, and data interfaces with production equipment, the MES system can obtain real-time information on production line operation status, output, and fault information. The ERP system stores key information such as production plans, material requirements, and inventory status. Through integration with the MES system, the ERP system can share this data with the MES system in real time, while the MES system also feeds back real-time production data to the ERP system, achieving two-way data exchange. IoT devices, such as sensors and wireless monitoring equipment, can monitor various parameters in the production process in real time, such as temperature, pressure, humidity, and vibration, and transmit this data to the MES system via wireless or wired networks, providing crucial information for production decisions. Real-time production data forms the basis for subsequently determining target optimization models and target production scheduling.

[0077] It should be noted that data acquisition methods such as Ethernet mode, ordinary Ethernet mode, data acquisition card, configuration software acquisition, and manual assistance can cover data from various aspects such as production equipment, personnel, raw materials and materials, production processes and procedures. Therefore, real-time production data can also be obtained using Ethernet mode, ordinary Ethernet mode, data acquisition card, configuration software acquisition, and manual assistance.

[0078] Here, preprocessing methods can include integration and cleaning, or data format conversion, outlier handling, and data fusion to ensure data accuracy and consistency.

[0079] In an optional embodiment of the present invention, step 102 includes:

[0080] Step 1021: Based on the real-time production data, obtain the process processing time data and available machine time data for each piece of equipment, start-up time data for each process, material availability time data and safety buffer period data, sequence classification data and allowable adjustment range data for preset key processes, skill level data for personnel and skill requirement level data for each process, historical utilization rate data and fluctuation range limit data for each piece of equipment, historical energy consumption data per unit output value and energy consumption reduction rate target data.

[0081] Specifically, the real-time production data includes or is calculated from the aforementioned data, providing a data foundation for subsequently determining various constraints.

[0082] Step 1022: Determine the first constraint condition based on the process time data of each machine and the available machine time data of the machine;

[0083] Specifically, for each piece of equipment, the sum of the processing times of all its processes should not exceed the available machine hours of that equipment. Therefore, the first constraint is the equipment capacity constraint: ∑process processing time ≤ available machine hours (including maintenance window reservation).

[0084] Step 1023: Determine the second constraint condition based on the start time data, material availability time data, and safety buffer period data for each process.

[0085] Specifically, the start time of each process should not be earlier than the material preparation time plus the safety buffer period. Therefore, the second constraint is a material hard constraint: the start time of the process ≥ the material preparation time + the safety buffer period.

[0086] Step 1024: Determine the third constraint condition based on the preset key process sequence classification data and the allowable adjustment range data;

[0087] Specifically, the third constraint is the process sequence constraint: a strict hierarchy of critical process sequences (allowing limited flexibility). The sequence of critical processes should follow a strict hierarchy, but flexible adjustments are allowed within a limited range. This typically requires setting a priority matrix for the process sequence and determining it in conjunction with the actual adjustment range. In a specific embodiment, a production process includes the following five processes: A: Raw material preparation, B: Preliminary processing, C: Critical heat treatment, D: Fine processing, and E: Quality inspection. The priorities of these processes are set according to their importance and impact on the overall production flow. Simultaneously, the strictness of the critical process sequence is graded, allowing limited adjustments to non-critical processes. Table 1 shows the priority matrix.

[0088] Table 1 Priority Matrix

[0089] Process Priority Keyness classification Allowable adjustment range A 2 Low Larger B 3 middle medium C 5 high No (strict) D 4 middle medium E 1 Low Larger

[0090] Priority: Higher values ​​indicate higher priority. In Table 1, quality inspection (E), although having the lowest priority, may be a necessary final step in actual production to ensure product quality. Critical heat treatment (C) has the highest priority because it is a key factor affecting product quality and performance.

[0091] Criticality classification: Divided into three levels: high, medium, and low. The sequence of high-critical processes must be strictly adhered to and cannot be adjusted. Medium-critical processes allow for adjustments within a limited range to adapt to changes in production. Low-critical processes have a wider range of adjustments.

[0092] Permissible adjustment range: Determined based on criticality level. Highly critical processes (C) are not allowed any adjustments and must be performed in the predetermined sequence. Medium-critical processes (B and D) are allowed adjustments within a moderate range to accommodate minor changes in production. Low-critical processes (A and E) have a larger adjustment range and can be flexibly adjusted according to production needs.

[0093] Based on the priority matrix above, the following third constraint condition can be determined:

[0094] The critical heat treatment (C) must be carried out strictly in the predetermined order, and no adjustments are allowed.

[0095] The order of preliminary processing (B) and fine processing (D) can be adjusted within a moderate range to accommodate changes in production, but it should be ensured that they do not interfere with the critical heat treatment (C).

[0096] Raw material preparation (A) and quality inspection (E) have a wide range of adjustments and can be flexibly arranged according to production needs.

[0097] Step 1025: Determine the fourth constraint condition based on the skill level data of the personnel and the skill requirement level data of each process.

[0098] Specifically, the skill level of personnel should not be lower than the skill requirement level of the process they are performing. Therefore, the fourth constraint is the personnel skill matching threshold (skill level ≥ process requirement level).

[0099] Step 1026: Determine the fifth constraint condition based on the historical utilization data and fluctuation range limit data of each device;

[0100] Specifically, the utilization rate fluctuation of a single device should be controlled within a certain range, such as not exceeding 15%. Therefore, the fifth constraint is the load balancing limit (single device utilization rate fluctuation ≤ 15%).

[0101] Step 1027: Determine the sixth constraint condition based on the historical energy consumption data per unit of output and the target data for energy consumption reduction rate.

[0102] Specifically, the energy consumption reduction rate per unit of output should reach or exceed a given target value, such as 5%, which can be achieved by calculating the actual energy consumption reduction rate and comparing it with the target value. Therefore, the sixth constraint is the green production indicator (energy consumption reduction rate per unit of output ≥ 5%).

[0103] Step 1028: Determine the constraints of the target optimization model based on the first constraint, the second constraint, the third constraint, the fourth constraint, the fifth constraint, and the sixth constraint.

[0104] Specifically, the constraints of the objective optimization model include the first, second, third, fourth, fifth, and sixth constraints. These constraints limit the range of values ​​for decision variables (including the allocation of production tasks, the arrangement of production time, and the selection of equipment), restrict the feasible solution space of the objective production schedule, ensure the feasibility and compliance of the final objective production schedule, and improve its effectiveness.

[0105] In an optional embodiment of the present invention, step 103 includes:

[0106] Step 1031: Obtain the production target that minimizes the production time dimension;

[0107] Specifically, minimizing production time is a production goal that requires optimizing production scheduling to reduce unnecessary waiting and idle time, thereby improving production efficiency.

[0108] Step 1032: Obtain the production target that minimizes the cost dimension;

[0109] Specifically, minimizing production costs requires optimizing production plans and resource allocation to reduce waste and lower production costs.

[0110] Step 1033: Obtain the production target that maximizes resource utilization;

[0111] Specifically, the production goal of maximizing resource utilization requires improving the utilization rate of equipment and materials by rationally arranging production tasks and equipment use.

[0112] Step 1034: Determine multiple production objectives based on the production objective of minimizing production time, the production objective of minimizing cost, and the production objective of maximizing resource utilization.

[0113] Specifically, the multi-dimensional production objectives include minimizing production time, minimizing costs, and maximizing resource utilization. Obtaining these three dimensions of production objectives helps to determine the objective optimization model and improve the effectiveness of objective production scheduling.

[0114] In an optional embodiment of the present invention, step 104 includes:

[0115] Step 1041: Determine the core objective function based on the multiple dimensions of production objectives;

[0116] Specifically, to achieve the production objective of minimizing production time, it is necessary to minimize the delivery deviation penalty cost P; to achieve the production objective of minimizing cost, it is necessary to minimize changeover cost S and energy consumption cost E; to achieve the production objective of maximizing resource utilization, it is necessary to minimize the delivery deviation penalty cost P, changeover cost S, and energy consumption cost E. The delivery deviation penalty cost P is quantified by calculating the penalty cost based on the number of days or proportion of delayed delivery; the changeover cost S is quantified, including costs incurred for equipment adjustments, production line switching, and material replacement; the energy consumption cost E is quantified by measuring energy consumption during the production process and converting it into cost. After quantifying the delivery deviation penalty cost P, changeover cost S, and energy consumption cost E based on real-time production data, the core objective function can be determined as min(∑(P+S+E)), where P is the delivery deviation penalty cost, S is the changeover cost, and E is the energy consumption cost.

[0117] Step 1042: Determine the composite objective function based on the preset auxiliary objective function and the core objective function.

[0118] Specifically, the pre-set auxiliary objective function can be to maximize equipment utilization and / or minimize inventory costs. In multi-objective optimization problems, there may be conflicts between the various objectives. For example, in production scheduling problems, improving production efficiency may increase energy costs, while reducing changeover costs may extend delivery time. By introducing a pre-set auxiliary objective function and combining it with the core objective function to form a composite objective function, the advantages and disadvantages of these objectives can be better balanced, and a solution that satisfies multiple objectives can be found. This makes the final determined target production schedule more consistent with actual production conditions and balances the various optimization objectives.

[0119] In an optional embodiment of the present invention, step 1042 includes:

[0120] Step 10421: Obtain a preset auxiliary objective function; the preset auxiliary objective function includes maximizing equipment utilization.

[0121] Specifically, the preset auxiliary objective function can be selected according to the actual production situation. In this embodiment, the preset auxiliary objective function is to maximize equipment utilization, where equipment utilization is a value between 0 and 1, representing the effective utilization degree of the equipment.

[0122] Step 10422: Determine the composite objective function based on the preset auxiliary objective function and the core objective function; the composite objective function is min(C-λU);

[0123] Where C = ∑(P + S + E); min(C - λU) is the composite objective function, where C is the total cost, λ is the weighting coefficient, and U is the maximum equipment utilization rate, where P is the delivery deviation penalty cost, S is the changeover cost, and E is the energy consumption cost.

[0124] Specifically, the total production cost of various products is C, the delivery deviation penalty is P, the changeover cost is S, and the energy consumption cost is E. The core objective function is min(C)=min(∑(P+S+E)). We also need to consider the preset auxiliary objective function, which is to maximize equipment utilization. Based on the preset auxiliary objective function and the core objective function, we can construct the composite objective function min(C-λU).

[0125] Specifically, λ is a positive weight used to adjust the relative importance between the two objectives (total cost and maximizing equipment utilization). If λ is large, then equipment utilization will play a more important role in the optimization process; if λ is small, then total cost will play a more important role. In practical applications, the optimal value of λ can be found through experimentation and adjustment.

[0126] In an optional embodiment of the present invention, the target optimization model in step 105 is:

[0127] Objective function: Composite objective function;

[0128] Constraints: Objective constraints.

[0129] In an optional embodiment of the present invention, step 106 includes:

[0130] Step 1061: Determine the equipment disturbance event based on the real-time monitoring data of the production equipment;

[0131] Specifically, real-time monitoring data of the production equipment, such as fault codes and maintenance time to repair (MTTR) predictions, is acquired through sensors or monitoring systems deployed on the production equipment. If the equipment monitoring data includes fault code identification and / or MTTR predictions exceeding 1 hour, it is identified as an equipment disturbance event.

[0132] Step 1062: Based on real-time order data, determine order-level disturbance events;

[0133] Specifically, order changes are obtained in real time through the order management system or customer relationship management system, such as the emergence of urgent orders (e.g., delivery time is less than the preset service period) and the impact of order insertion. When the real-time order data, including the delivery time compression rate of urgent orders, is greater than or equal to 30%, it is identified as an order-level disturbance event.

[0134] Step 1063: Based on real-time personnel data and material supply data, determine resource-level disturbance events;

[0135] Specifically, the system uses a human resources management system to obtain real-time data on employee absences, such as absence rate warnings; and a materials management system to obtain real-time data on material supply status, such as material delay probability. Specifically, by comparing production plans before and after order insertion, the system calculates the additional production time and delayed delivery dates caused by the order insertion, thus determining the production time delay. It also assesses the increased cost by calculating additional material costs, labor costs, and equipment maintenance costs, thus determining the additional costs caused by the order insertion. Furthermore, it compares resource utilization rates before and after the order insertion to determine the degree of decrease in resource utilization. Finally, it determines the impact of the order insertion based on the production time delay, additional costs, and decrease in resource utilization. A resource-level disturbance event is identified when personnel data, including the absence rate, exceeds the preset absence rate, and / or material supply data includes incomplete material supply, or when the personnel and material availability delay caused by the absence rate warning and / or material delay probability exceeds the current work order buffer period.

[0136] Step 1064: Based on the equipment disturbance event, the order-level disturbance event, the resource-level disturbance event, and the multi-dimensional preset performance indicators, determine at least one combination of optimization models to obtain an optimized combination model;

[0137] Specifically, the multi-dimensional preset performance indicators include:

[0138] Convergence speed metric: By monitoring the fitness improvement rate of each generation, the convergence speed of the optimization sub-model can be evaluated. A faster convergence speed means that the optimization sub-model can find a high-quality solution in a shorter time.

[0139] Solution quality metrics: The quality of the solution is evaluated using the hypervolume metric (HV value) or other similar metrics. The larger the HV value, the wider the coverage of the solution set in the target space, and the higher the quality of the solution.

[0140] Resource consumption metrics: By monitoring CPU (Central Processing Unit) time and memory usage, the resource consumption of the optimization sub-model can be evaluated. This is crucial for optimizing the running efficiency of the optimization sub-model while ensuring solution quality.

[0141] The optimization model is dynamically adjusted based on disturbance events and multi-dimensional preset performance indicators. When the convergence speed indicator decreases, a first or second optimization sub-model is added to the optimization model; when the CPU exceeds its limit, the optimization model is downgraded to the first optimization sub-model; when the solution quality indicator stagnates, a third optimization sub-model is added to the optimization model; when a device disturbance event is identified and the maintenance time is greater than 30 minutes, the optimization model switches to the first or second optimization sub-model; when an order-level disturbance event is identified and the urgency is not greater than 0.9, the optimization model switches to the first optimization sub-model; when an order-level disturbance event is identified and the urgency is greater than 0.9, a second optimization sub-model is added to the optimization model; when a resource-level disturbance event is identified and the gap is greater than 20%, the optimization model switches to the third optimization sub-model.

[0142] Step 1065: Iteratively solve the target optimization model according to the optimization combination model to obtain the target solution set.

[0143] like Figure 2 As shown, in an optional embodiment of the present invention, when the optimization combination model in step 1065 includes a first optimization sub-model, a second optimization sub-model, and a third optimization sub-model, step 1065 includes:

[0144] Step 10651: Obtain a preset random initial solution set;

[0145] Specifically, by initializing the population, a preset random initial solution set containing parameters such as equipment allocation, process sequence, and resource combination is generated, and the fitness of individuals is evaluated based on a multi-objective function. Then, an iterative optimization loop is entered. In one specific embodiment, 100 preset random initial solution sets can be randomly generated.

[0146] Step 10652: Perform a global search based on the first optimized sub-model and the preset random initial solution set to obtain the first optimal solution;

[0147] Specifically, a preset random initial solution set is input into the first optimization sub-model. In each iteration, the first optimization sub-model first performs a global search, generating a new population through selection, crossover, and mutation operations, focusing on exploring potential high-quality regions in the solution space. The first optimization sub-model transforms the problem into a string in chromosome form, such as binary encoding. Then, based on the fitness of individuals, it selects individuals from the current population for reproduction, simulating chromosome crossover in biological genetics, to generate new individuals. Genes in individuals are altered with a preset probability to increase population diversity. The fitness of individuals is calculated according to the fitness function f(x) = C(x) - λU(x) until a preset first iteration number or fitness criterion is reached. Here, f(x) is the fitness function, representing the optimization objective value under solution x; C(x) is the total cost function, representing the total cost under solution x; U(x) is the equipment utilization function, representing the equipment utilization under solution x; C is the total cost; λ is the weighting coefficient; and U is the maximum equipment utilization. By maintaining the diversity of multiple individuals in the population, it is easy to perform a global search in the search space, thus obtaining the first optimal solution.

[0148] Step 10653: Perform a local search based on the second optimization sub-model and the first optimal solution to obtain the second optimal solution;

[0149] Specifically, when the preset number of first iterations is reached, the second optimization sub-model performs a local fine search based on the first optimal solution, and adjusts the process time window and resource allocation scheme by using the dual guidance mechanism of individual historical best and group optimal solutions.

[0150] The second optimization sub-model is achieved through formula v. new =w·v+c1·r1·(p Best -x)+c2·r2·(g Best -x) is used to update the particle's velocity, via x new =x+v new Tracking the individual's historical best p Best and the group optimal solution g Best , where v new Let be the particle's new velocity vector, w be the inertia weight used to control the influence of the particle's current velocity on its subsequent velocities, controlling the global / local search balance, v be the particle's current velocity vector, c1 be the individual learning factor, c2 be the social learning factor, r1 and r2 be random numbers between 0 and 1 used to increase the randomness of the search, and p be... Best Let g be the particle's historical best position, and x be the particle's current position. Here, the particle's initial position is the first optimal solution, and g is the particle's current position.Best For the global optimal position of all particles, x new This is the new position of the particle.

[0151] Step 10654: Perturbation processing is performed based on the third optimization sub-model and the second optimal solution to obtain the target solution set.

[0152] Specifically, when the preset second iteration number is reached, the third optimization sub-model intervenes, applies a controllable perturbation to the second optimal solution, and helps the model escape the local optimum trap by probabilistically accepting inferior solutions.

[0153] Specifically, the third optimization sub-model first sets the initial temperature T (usually relatively high) and the termination temperature T. end (Lower), and cooling strategies (such as exponential decay); the second optimal solution is used as the initial solution of the third optimization sub-model, and a small random perturbation is applied to the initial solution to generate a new solution in its neighborhood. This can be achieved by adding random noise or applying some transformation in the solution space; the objective function value of the new solution is calculated and compared with the objective function value of the current solution. The objective function value of the solution is calculated by f1(x) = C(x) - λU(x), where f1(x) is the objective function value of the solution, C(x) is the total cost function, representing the total cost under solution x, and U(x) is the equipment utilization function. Let represent the equipment utilization rate under solution x, C be the total cost, λ be the weighting coefficient, and U be the maximum equipment utilization rate. If the probability exp(-Δf / T) is less than 0, it means the new solution is better, and the new solution is accepted unconditionally. If the probability exp(-Δf / T) is greater than 0, it means the new solution is worse, and the new solution is accepted with a probability of exp(-Δf / T). Here, Δf is the difference between the objective function value of the new solution and the objective function value of the current solution, and T is the current temperature. The temperature is gradually reduced using an exponential cooling strategy or other cooling strategies. If the condition is met, the calculation stops and the target solution set is output (if the temperature T is reduced to the termination temperature T). end (If the temperature is less than 1, stop the calculation; otherwise, recalculate the objective function value of the current solution.)

[0154] Specifically, during the optimization process of a preset random initial solution set using multiple optimization sub-models, core parameters such as crossover rate, mutation rate, and inertia weight are dynamically adjusted. A strategy of retaining the best individual and a certain proportion of excellent individuals is employed to ensure that high-quality solutions are not lost. When the preset number of iterations or the solution set convergence threshold is reached, the optimization combination model terminates and outputs the Pareto optimal solution set as the target solution set. The entire process, through the complementary advantages of the models, significantly improves the convergence speed while ensuring solution diversity, effectively addressing the optimization challenges of large-scale complex scheduling problems.

[0155] It should be noted that the optimization combination model can also be the first optimization sub-model, or the second optimization sub-model, or the third optimization sub-model, or a combination of the first and second optimization sub-models, or a combination of the first and third optimization sub-models, or a combination of the second and third optimization sub-models, etc. The solution process will not be described in detail.

[0156] In an optional embodiment of the present invention, step 107 includes:

[0157] Step 1071: Determine the production task list based on the target solution set; the production task list includes the quantity, priority, and delivery date of the production tasks.

[0158] Specifically, the target solution set includes: production task allocation: clarifying which equipment, production line, or worker should perform each production task; time scheduling: determining the start and end times of each production task; resource allocation: allocating the required equipment, manpower, tools, and materials; priority management: setting the priority of production tasks based on factors such as order urgency and customer importance; and bottleneck management: identifying and optimizing bottleneck processes in the production flow to ensure smooth production. Therefore, the quantity, priority, and delivery time of production tasks can be extracted from the target solution set to form a production task list.

[0159] Based on the target solution set, determine key information such as the quantity, priority, and delivery date of each production task. Then, based on the target solution set, develop a clear production task list, specifying the detailed requirements for each task, such as product type, quantity, and quality standards.

[0160] Step 1072: Determine resource allocation based on the production task list; the resource allocation includes equipment resources, human resources, and material resources.

[0161] Specifically, based on the production task list and target set, equipment resources, human resources and material resources can be allocated to determine the resource configuration required for each production task, ensuring that each production task has sufficient resource support while avoiding resource waste.

[0162] Step 1073: Determine the target production schedule based on the production task list and the resource allocation;

[0163] Specifically, the details of production tasks, resource allocation, and production sequence can be compiled into a production plan table as a target production schedule, which will facilitate the subsequent manufacturing execution system to carry out production tasks according to the plan table.

[0164] Step 1074: Control the production scheduling module to output the target production schedule.

[0165] Specifically, the target production schedule fully considers the actual needs of the enterprise and the real-time changes in the production environment, controlling the output of the target production schedule from the production scheduling module to the MES system (Manufacturing Execution System) for execution. The MES system allocates, schedules, and monitors production tasks based on this target production schedule to ensure the smooth operation of production activities.

[0166] A specific embodiment of the production scheduling method of the present invention includes:

[0167] First, real-time production data is obtained from Manufacturing Execution System (MES), Enterprise Resource Planning System (ERP), and Internet of Things (IoT) devices.

[0168] Then, a target optimization model is constructed. This model achieves a dynamic trade-off between multiple production objectives through a composite objective function and a multi-dimensional constraint system. The multi-objective optimization model uses min(∑(delivery deviation penalty + changeover cost + energy consumption cost)) as its core objective function, where:

[0169] Delivery deviation penalty: adopts piecewise function design (differentiated penalty coefficient for early / late delivery); changeover cost: includes equipment setup time conversion cost and material changeover loss; energy consumption cost is based on a dynamic electricity cost model based on process parameters and equipment load rate;

[0170] The multi-dimensional constraint system is constructed using a three-level hierarchical architecture:

[0171] 1. Hard constraint layer:

[0172] Equipment capacity constraints: ∑processing time ≤ available machine hours (including reserved maintenance windows);

[0173] Material hard constraints: Process start time ≥ material kitting time + safety buffer period;

[0174] 2. Soft constraint layer:

[0175] Process sequence constraints: Critical process sequence strictness classification (with limited flexibility allowed) (as shown in Table 1);

[0176] Personnel skill matching threshold (skill level ≥ process requirement level);

[0177] 3. Optimize the guidance layer:

[0178] Load balancing limit (single device utilization fluctuation ≤ 15%);

[0179] Green production indicators (energy consumption per unit of output value decrease rate ≥ 5%);

[0180] Then, the objective optimization model is iteratively solved using at least one optimization model. The solution to the objective optimization model dynamically adjusts the objective weights through an adaptive weight allocator, and outputs a non-dominated solution set in conjunction with a Pareto front screening mechanism, supporting decision-makers in selecting the optimal balance scheme based on the actual scenario.

[0181] Here, by constructing a three-layer intelligent decision-making system, dynamic optimization of the combination of the first, second, and third optimization sub-models is achieved:

[0182] 1. Environmental perception layer: By deploying a distributed sensor network, three types of production disturbance events are captured in real time:

[0183] Equipment-level disturbances (fault code identification, MTTR prediction);

[0184] Order-level disturbances (urgent order feature extraction, order insertion impact calculation);

[0185] Resource-level disturbances (employee absenteeism warning, material delay probability);

[0186] 2. Algorithm decision layer: An algorithm performance monitoring matrix is ​​constructed using three indicators:

[0187] Convergence speed metric (fitness improvement rate per generation);

[0188] Solve the quality indicators (excess volume indicators);

[0189] Resource consumption metrics (CPU time / memory usage);

[0190] It can also automatically combine the three first, second, and third optimization sub-models using pre-stored historical data, such as historical scene feature matching (triggering similar case reuse when Euclidean distance ≤ 0.2); or model combination recommendation (such as automatically switching to the second optimization sub-model + third optimization sub-model for rapid convergence in equipment failure scenarios).

[0191] 3. Execution control layer

[0192] When the optimization combination model is a combination of the first, second, and third optimization sub-models, the hybrid algorithm relay mechanism is as follows:

[0193] Initial stage (iterations 1-100): The first optimization sub-model dominates the global exploration (crossover rate 0.8→0.6 linear decay); Mid-stage (iterations 101-300): The second optimization sub-model (inertia weight 0.9→0.4 dynamically adjusted); Late stage (iterations 301-500): The third optimization sub-model escapes local optima (cooling coefficient 0.95).

[0194] The parameters in the three optimization sub-models are adaptively adjusted: the mutation rate is automatically adjusted (0.01→0.1) based on the population diversity index (gene entropy value); or the particle velocity limit value of the second optimization sub-model is dynamically adjusted based on the neighborhood search effect.

[0195] like Figure 2 As shown, the specific process for solving the multi-objective optimization model includes: achieving a dynamic balance between global exploration and local optimization through cross-iteration of the first, second, and third optimization sub-models. First, the population is initialized, generating an initial solution set (random scheduling scheme) containing parameters such as equipment allocation, process sequence, and resource combination, and the fitness of individuals is evaluated based on the multi-objective function. Then, an iterative optimization loop begins: in each iteration, the first optimization sub-model performs a global search, generating a new population through selection, crossover, and mutation operations, focusing on exploring potential high-quality regions in the solution space; the second optimization sub-model performs a local fine-grained search based on the current optimal solution, using a dual guidance mechanism of individual historical best and group best solutions to adjust process time windows and resource allocation schemes; the third optimization sub-model intervenes at the end of each iteration, applying controllable perturbations to the current optimal solution, helping the model escape local optima traps through a strategy of probabilistically accepting inferior solutions. During the three-stage optimization process, core parameters such as crossover rate, mutation rate, and inertia weight are dynamically adjusted, and an elite retention strategy is adopted to ensure that high-quality solutions are not lost. When the preset number of iterations or the solution set convergence threshold is reached, the optimization combinatorial model terminates and outputs the Pareto optimal solution set as the target solution set. The entire process leverages the complementary strengths of the sub-models to significantly improve convergence speed while ensuring solution diversity, effectively addressing the optimization challenges of large-scale complex scheduling problems.

[0196] Furthermore, it is necessary to optimize the combination of the three optimization sub-models or the parameters of the model when disturbance events occur. A dual-loop control architecture is adopted to achieve rapid response and continuous optimization for abnormal events (disturbance events).

[0197] Outer ring (event response ring):

[0198] 1. Intelligent identification of disturbance events based on trigger conditions:

[0199] Equipment failure: MTTR (Mean Time To Repair) > 1 hour;

[0200] Urgent orders: Delivery time reduction rate ≥ 30%;

[0201] Material anomaly: Kitting time delay > current work order buffer period;

[0202] 2. Tiered response strategy for disturbance events:

[0203] Local adjustment: Work order process reordering (response time < 2 minutes);

[0204] Tabu search is used to quickly generate alternative process sequences while preserving the original schedules of unaffected processes;

[0205] Global adjustments: Full optimization of the scrolling time window (response time < 8 minutes);

[0206] Freeze the status of processes that have already started;

[0207] Inner loop (continuous optimization loop):

[0208] 1. Time window rolling mechanism:

[0209] Incremental optimization is performed every 15 minutes (adjusting only the schedule for the next 2 hours);

[0210] A full optimization (restructuring the schedule for the next 8 hours) is performed every 4 hours;

[0211] 2. Digital twin simulation verification:

[0212] Rehearse and adjust the plan in a virtual environment;

[0213] Key indicators such as output capacity utilization rate fluctuation forecast and probability of delay risk;

[0214] The system automatically selects the option with the highest overall score on key performance indicators for implementation.

[0215] The production scheduling method of this invention acquires real-time production data, such as work order information, equipment status, and inventory, from Manufacturing Execution System (MES), Enterprise Resource Planning System (ERP), and Internet of Things (IoT) devices. Then, modeling and goal setting are performed by constructing a multi-objective optimization mathematical model and setting optimization objectives (such as minimizing production time, reducing costs, and improving resource utilization). Next, a hyperheuristic control strategy is employed, dynamically adjusting the optimization algorithm based on search progress and selecting the optimal optimization method at different stages. Following this, hybrid optimization is used, combining first, second, and third optimization sub-models through cross-iteration to improve search efficiency and solution quality. Then, dynamic adjustment and real-time optimization are implemented, using an adaptive adjustment mechanism to cope with sudden changes in the production environment, making the scheduling scheme more flexible. Finally, the optimized scheme is output, and the final target production schedule is automatically pushed to the Manufacturing Execution System (MES) for execution. The entire process combines multiple optimization strategies to ensure high-quality production scheduling schemes are obtained under different production scenarios.

[0216] like Figure 3 As shown, an embodiment of the present invention provides a production scheduling device 200, comprising:

[0217] The first acquisition module 201 is used to acquire real-time production data through the intelligent manufacturing system;

[0218] The first determining module 202 is used to determine the constraints of the target optimization model based on the real-time production data.

[0219] The second acquisition module 203 is used to acquire production targets in multiple dimensions;

[0220] The second determining module 204 is used to determine a composite objective function based on the multiple dimensions of the production target;

[0221] The third determining module 205 is used to determine the target optimization model based on the constraints and the composite objective function;

[0222] Processing module 206 is used to iteratively solve the target optimization model using at least one optimization model to obtain the target solution set;

[0223] The control module 207 is used to control the production scheduling module to output the target production schedule based on the target solution set.

[0224] Optionally, the first determining module 202 is specifically used for:

[0225] Based on the real-time production data, the following data are obtained: process time data for each piece of equipment, available machine time data for each piece of equipment, start-up time data for each process, material availability time data and safety buffer period data, sequence classification data for preset key processes and allowable adjustment range data, skill level data for personnel and skill requirement level data for each process, historical utilization rate data and fluctuation range limit data for each piece of equipment, historical energy consumption data per unit of output value and energy consumption reduction rate target data.

[0226] Based on the process time data and available machine time data of each piece of equipment, the first constraint condition is determined;

[0227] Based on the start time data, material availability time data, and safety buffer period data for each process, the second constraint condition is determined.

[0228] Based on the preset key process sequence classification data and the allowable adjustment range data, the third constraint condition is determined;

[0229] Based on the skill level data of the personnel and the skill requirement level data of each process, the fourth constraint condition is determined;

[0230] Based on the historical utilization data and fluctuation range limit data of each device, the fifth constraint condition is determined;

[0231] Based on the historical energy consumption data per unit of output and the target data for energy consumption reduction rate, the sixth constraint condition is determined.

[0232] Based on the first constraint, the second constraint, the third constraint, the fourth constraint, the fifth constraint, and the sixth constraint, the constraints of the target optimization model are determined.

[0233] Optionally, the second determining module 204 is specifically used for:

[0234] Based on the production objectives across multiple dimensions, determine the core objective function;

[0235] A composite objective function is determined based on the preset auxiliary objective function and the core objective function.

[0236] Optionally, a composite objective function is determined based on a preset auxiliary objective function and the core objective function, including:

[0237] Obtain a preset auxiliary objective function; the preset auxiliary objective function includes maximizing equipment utilization.

[0238] Based on the preset auxiliary objective function and the core objective function, a composite objective function is determined; the composite objective function is min(C-λU);

[0239] Where C = ∑(P + S + E); min(C - λU) is the composite objective function, C is the total cost in the core objective function, λ is the weighting coefficient, U is to maximize equipment utilization, P is the delivery deviation penalty cost, S is the changeover cost, and E is the energy consumption cost.

[0240] Optionally, processing module 206 is specifically used for:

[0241] Based on real-time monitoring data of the production equipment, determine the equipment disturbance events;

[0242] Identify order-level disturbance events based on real-time order data;

[0243] Based on real-time personnel and material supply data, identify resource-level disturbance events;

[0244] Based on the equipment disturbance event, the order-level disturbance event, the resource-level disturbance event, and the multi-dimensional preset performance indicators, determine at least one combination of optimization models to obtain an optimized combination model;

[0245] The objective optimization model is iteratively solved based on the aforementioned optimization combination model to obtain the objective solution set.

[0246] Optionally, processing module 206 is specifically used for:

[0247] Obtain a preset random initial solution set;

[0248] A global search is performed based on the first optimization sub-model in the optimization combination model and the preset random initial solution set to obtain the first optimal solution;

[0249] A local search is performed based on the second optimization sub-model in the optimized combination model and the first optimal solution to obtain the second optimal solution;

[0250] The target solution set is obtained by perturbating the third optimization sub-model and the second optimal solution in the optimization combination model.

[0251] Optional, control module 207, specifically used for:

[0252] Based on the target solution set, a production task list is determined; the production task list includes the quantity, priority, and delivery date of each production task.

[0253] Based on the production task list, resource allocation is determined; the resource allocation includes equipment resources, human resources, and material resources.

[0254] Based on the production task list and the resource allocation, determine the target production schedule;

[0255] The production scheduling module controls the output of the target production schedule.

[0256] The production scheduling device proposed in this invention acquires real-time production data through an intelligent manufacturing system and uses this data to determine the constraints of the target optimization model. It acquires production targets from multiple dimensions and uses this data to determine a composite objective function. Based on the constraints and the composite objective function, it determines the target optimization model and iteratively solves the target optimization model using at least one optimization model to obtain the target solution set. Finally, based on the target solution set, it controls the production scheduling module to output the target production schedule, thereby improving the quality and efficiency of production line scheduling and enhancing its adaptability to dynamic environments.

[0257] It should be noted that this device corresponds to the method described above, and all implementations in the method embodiments described above are applicable to the embodiments of this device and can achieve the same technical effect. Further details will not be provided in this embodiment.

[0258] This invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any of the above embodiments. All implementations in the above method embodiments are applicable to the embodiments of this device and can achieve the same technical effects. Further details are omitted in this embodiment.

[0259] This invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method as described in any of the above embodiments. All implementations in the above method embodiments are applicable to the embodiments of this device and can achieve the same technical effects. Further details are omitted in this embodiment.

[0260] It should be noted that in the apparatus and method of the present invention, the components or steps can obviously be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Furthermore, the steps for performing the above series of processes can naturally be performed in the order described and in chronological order, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel, overlapping, or independently of each other.

[0261] It should be noted that in the above embodiments, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments described above is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0262] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A production scheduling method, characterized in that, include: Obtain real-time production data through intelligent manufacturing systems; Based on the real-time production data, determine the constraints of the target optimization model; Obtain production targets from multiple dimensions; Based on the production objectives across multiple dimensions, a composite objective function is determined; Based on the constraints and the composite objective function, determine the objective optimization model; The target optimization model is iteratively solved using at least one optimization model to obtain the target solution set; Based on the target solution set, the production scheduling module is controlled to output the target production schedule; Specifically, the objective optimization model is iteratively solved using at least one optimization model to obtain the objective solution set, including: Based on real-time monitoring data of the production equipment, determine the equipment disturbance events; Identify order-level disturbance events based on real-time order data; Based on real-time personnel and material supply data, identify resource-level disturbance events; Based on the device disturbance event, the order-level disturbance event, the resource-level disturbance event, and multi-dimensional preset performance indicators, at least one combination of optimization models is determined to obtain an optimized combination model. The multi-dimensional preset performance indicators include convergence speed indicators, solution quality indicators, and resource consumption indicators. The resource consumption indicators are evaluated by monitoring the CPU time and memory usage. The optimized combination model is a first optimized sub-model, or a second optimized sub-model, or a third optimized sub-model, or a combination of the first and second optimized sub-models, or a combination of the first and third optimized sub-models, or a combination of the second and third optimized sub-models, or a combination of the first, second, and third optimized sub-models. The objective optimization model is iteratively solved based on the aforementioned optimization combination model to obtain the objective solution set; Wherein, when the optimization combination model includes a first optimization sub-model, a second optimization sub-model, and a third optimization sub-model, the target optimization model is iteratively solved according to the optimization combination model to obtain the target solution set, including: Obtain a preset random initial solution set; A global search is performed based on the first optimization sub-model and the preset random initial solution set to obtain the first optimal solution; A local search is performed based on the second optimization sub-model and the first optimal solution to obtain the second optimal solution; The target solution set is obtained by perturbation processing based on the third optimization sub-model and the second optimal solution.

2. The production scheduling method according to claim 1, characterized in that, Based on the real-time production data, the constraints of the target optimization model are determined, including: Based on the real-time production data, the following data are obtained: process processing time data for each piece of equipment, available machine time data for each piece of equipment, start-up time data for each process, material availability time data and safety buffer period data, sequence classification data for preset key processes and allowable adjustment range data, skill level data for personnel and skill requirement level data for each process, historical utilization rate data and fluctuation range limit data for each piece of equipment, historical energy consumption data per unit of output value and energy consumption reduction rate target data. Based on the process time data and available machine time data of each piece of equipment, the first constraint condition is determined; Based on the start time data, material availability time data, and safety buffer period data for each process, the second constraint condition is determined. Based on the sequence classification data of the preset key processes and the allowable adjustment range data, the third constraint condition is determined; Based on the skill level data of the personnel and the skill requirement level data of each process, the fourth constraint condition is determined; Based on the historical utilization data and fluctuation range limit data of each device, the fifth constraint condition is determined; Based on the historical energy consumption data per unit of output and the target data for energy consumption reduction rate, the sixth constraint condition is determined. Based on the first constraint, the second constraint, the third constraint, the fourth constraint, the fifth constraint, and the sixth constraint, the constraints of the target optimization model are determined.

3. The production scheduling method according to claim 1, characterized in that, Based on the aforementioned multiple production objectives, a composite objective function is determined, including: Based on the production objectives across multiple dimensions, determine the core objective function; A composite objective function is determined based on the preset auxiliary objective function and the core objective function.

4. The production scheduling method according to claim 1, characterized in that, Based on the target solution set, the production scheduling module is controlled to output the target production schedule, including: Based on the target solution set, a production task list is determined; the production task list includes the quantity, priority, and delivery date of each production task. Based on the production task list, resource allocation is determined; the resource allocation includes equipment resources, human resources, and material resources. Based on the production task list and the resource allocation, determine the target production schedule; The production scheduling module controls the output of the target production schedule.

5. A production scheduling device, characterized in that, include: The first acquisition module is used to acquire real-time production data through the intelligent manufacturing system; The first determining module is used to determine the constraints of the target optimization model based on the real-time production data. The second acquisition module is used to acquire production targets from multiple dimensions; The second determining module is used to determine the composite objective function based on the multiple dimensions of the production target; The third determining module is used to determine the target optimization model based on the constraints and the composite objective function; The processing module is used to iteratively solve the target optimization model using at least one optimization model to obtain the target solution set; The control module is used to control the production scheduling module to output the target production schedule based on the target solution set. The processing module is used for: Based on real-time monitoring data of the production equipment, determine the equipment disturbance events; Identify order-level disturbance events based on real-time order data; Based on real-time personnel and material supply data, identify resource-level disturbance events; Based on the equipment disturbance event, the order-level disturbance event, the resource-level disturbance event, and the multi-dimensional preset performance indicators, at least one combination of optimization models is determined to obtain an optimized combination model; the multi-dimensional preset performance indicators include convergence speed indicators, solution quality indicators, and resource consumption indicators, and the resource consumption indicators are evaluated by monitoring the central processing unit time and memory usage; The optimized combination model is a first optimized sub-model, or a second optimized sub-model, or a third optimized sub-model, or a combination of the first optimized sub-model and the second optimized sub-model, or a combination of the first optimized sub-model and the third optimized sub-model, or a combination of the second optimized sub-model and the third optimized sub-model, or a combination of the first optimized sub-model, the second optimized sub-model and the third optimized sub-model; The objective optimization model is iteratively solved based on the aforementioned optimization combination model to obtain the objective solution set; Wherein, when the optimization combination model includes a first optimization sub-model, a second optimization sub-model, and a third optimization sub-model, the target optimization model is iteratively solved according to the optimization combination model to obtain the target solution set, including: Obtain a preset random initial solution set; A global search is performed based on the first optimization sub-model and the preset random initial solution set to obtain the first optimal solution; A local search is performed based on the second optimization sub-model and the first optimal solution to obtain the second optimal solution; The target solution set is obtained by perturbation processing based on the third optimization sub-model and the second optimal solution.

6. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 4.

Images