A production decision-making method, apparatus, electronic device, and storage medium
By converting the nonlinear loss calculation formula into a linear form and setting filtering and constraint conditions in the MIP model, and then solving the problem with weighted average, the financial cost of upgrading a production line from a single shift to a double shift, which is currently impossible in existing technologies, is solved. This enables multiple optimal production decisions and automatic location of data defects, thereby improving the accuracy and efficiency of production decisions.
Patent Information
- Application Number
- CN202611337934.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-31
- Publication Date
- 2026-11-17
AI Technical Summary
Existing mixed-integer programming (MIP) models cannot effectively calculate the real financial costs such as new employee training losses, ramp-up losses, and rework losses when a production line is upgraded from a single shift to a double shift. This makes it impossible to directly incorporate these costs into the objective function of a linear MIP model, thus failing to obtain production decisions that meet the requirements.
The nonlinear loss calculation formula is converted into a linear loss calculation formula by fixing some variable values to the average value of optional parameters. Filtering and constraint conditions are set in the MIP model, and multiple optimal combination schemes are output by weighted solution through multiple sets of business strategy weights. When no feasible scheme is found, the filtering and/or constraint conditions are relaxed and the solution is re-solved to locate and output the defects in the production line configuration list.
It enables accurate modeling and optimization of production line changeover losses, outputs optimal solutions under various business strategies, and automatically locates data defects when there is no solution, thereby improving the accuracy of production decisions and the efficiency of problem investigation.
Smart Images

Figure CN123117664A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of intelligent production scheduling technology, and more specifically, to a production decision-making method, apparatus, electronic device, and storage medium. Background Technology
[0002] Factory production system planning refers to the comprehensive decision-making process within each planning cycle (e.g., month) regarding production line takt time (JPH, hourly output), shift patterns (single / double shift), number of working days, working hours, and employee recruitment and turnover, aiming to reduce overall costs such as energy consumption, labor, and overtime while meeting order demands. Because these decision variables are interconnected, some industry solutions currently employ mixed integer programming (MIP) models to model and solve this problem.
[0003] However, existing MIP production planning schemes generally have the following shortcomings: when a production line is upgraded from a single shift to a double shift, real financial costs such as new employee training losses, ramp-up losses, and rework losses are incurred. Among these, the calculation formulas for ramp-up losses and rework losses contain some bilinear terms, meaning that there are two decision variables multiplied together. This makes it impossible to directly incorporate them into the objective function of the linear MIP model, thus failing to obtain production decisions that meet the requirements. Summary of the Invention
[0004] In view of this, embodiments of this application provide a production decision-making method, apparatus, electronic device, and storage medium to optimize production decisions.
[0005] In a first aspect, embodiments of this application provide a production decision-making method, the method comprising: Obtain the production line configuration list; the production line configuration list includes the decision data and decision constraints required by the factory to make production decisions in each month. The nonlinear loss calculation formula of the mixed integer programming (MIP) model to be input is fixed with some variable values as the average value of multiple optional parameters set in the decision data, so as to convert the nonlinear loss calculation formula into a linear loss calculation formula containing only one variable and output the corresponding linear coefficient. The linear coefficients and the production line configuration list are input into the MIP model. The MIP model is then used to obtain the optimal combination scheme corresponding to each set of business strategy weights under the constraints of filtering and constraint conditions, according to the preset weights of multiple business strategies. The filtering and constraint conditions are determined based on the decision data and the decision constraints. If the optimal combination scheme exists, then based on the solved optimal combination schemes, determine the cost of executing each of the optional parameters under the optimal combination scheme, and output it; If the optimal combination scheme does not exist, the MIP model is controlled to be re-solved by reducing at least one of the filtering conditions and / or the constraints, and the defects of the production line configuration list are located and output based on the solution results.
[0006] In one feasible implementation, the loss calculation formula includes a formula for calculating losses from new employee training, a formula for calculating losses from production ramp-up, and a formula for calculating losses from rework. The loss calculation formula is triggered only when the shift changes from single shift to double shift in the current month; the production ramp-up loss calculation formula includes a bilinear term multiplied by the actual working hours and the manpower of the current month; the rework loss calculation formula includes a bilinear term multiplied by the JPH value and the actual working hours; the partial variable is the actual working hours; and the average value is the average value of multiple selectable actual working hours in the decision data.
[0007] In a feasible implementation plan, a weighted solution is performed according to multiple preset business strategy weights, including: The minimum sum of the products of each cost indicator and its corresponding business strategy weight is used as the optimization objective. An overcapacity penalty is imposed to construct a weighted objective function. The cost indicators include overtime cost, energy cost, monthly salary cost, recruitment and training cost, and shift changeover loss. The relative importance of each cost indicator is different under different business strategy weights. The combination scheme corresponding to the minimum weighted objective function value of each set of business strategy weights is the optimal combination scheme under that business strategy weight.
[0008] In a feasible implementation, each set of business strategy weights in the multiple sets of business strategy weights includes: overtime cost weight, energy consumption cost weight, monthly human resource salary cost weight, recruitment and training cost weight, and shift changeover loss weight, and the sum of each weight is 1; the preset multiple sets of business strategy weights include at least two of the following sets: The first strategy weighting is based on overtime cost, which has the highest weighting. The second strategy weighting has the highest weight for energy consumption cost. The third strategy weighting has the highest weight for monthly human resource salary costs. The fourth strategy weighting has the highest weight for recruitment and training costs; The fifth strategy weighting is based on the shift switching loss weighting, which has the highest weighting.
[0009] In one feasible implementation, the filtering conditions include at least one of the following: The first filtering condition indicates that the determined shift matches the shift type set in the production line configuration list for the current month. The second filtering condition indicates that the number of working days does not exceed the total number of days in the month; The third filtering condition indicates that the production capacity demand of each model does not exceed the maximum hourly output JPH of that model. The fourth filtering condition indicates that the total production capacity is not less than the sum of the demand and the surplus required by the production line configuration list; The fifth filtering condition characterizes the positive correlation between working days and working hours.
[0010] In one feasible implementation, the constraints include at least one of the following: The first constraint represents the flow relationship between the number of employees at the beginning of each month and the number of employees in the previous month, the number of employees who left, and the number of employees hired this month. The second constraint indicates that only one hourly output JPH pattern is selected each month. The third constraint indicates that the actual number of employees each month is not less than the sum of the required number of employees and the additional number of employees required by the selected JPH mode. The fourth constraint indicates that monthly production capacity must not be less than the sum of demand and surplus. The fifth constraint indicates that the selected JPH mode each month will be used as the basis for calculating the shift changeover loss, and the shift changeover loss will be used to participate in the weighted solution calculation. The sixth constraint indicates that overtime pay only applies to the selected JPH mode; The seventh constraint indicates that the monthly excess capacity is not less than the difference between the actual capacity and the minimum demand; The eighth constraint characterizes the cumulative effect of shift changeover losses corresponding to new employee training losses, production ramp-up losses, and rework losses across all months.
[0011] In a feasible implementation, the defects in the production line configuration list are located and output based on the solution results, including: The filter conditions and / or constraints that were reduced during the re-solution process are identified as bottleneck conditions, and the corresponding month information, condition category, and gap value are output.
[0012] Secondly, embodiments of this application also provide a production decision-making device, the device comprising: The acquisition module is used to acquire a production line configuration list; the production line configuration list includes decision data and decision constraints required by the factory to make production decisions in each month. The conversion module is used to fix some variable values in the nonlinear loss calculation formula of the mixed integer programming (MIP) model to be input to the average value of multiple optional parameters set in the decision data, so as to convert the nonlinear loss calculation formula into a linear loss calculation formula containing only one variable and output the corresponding linear coefficient. The input module is used to input the linear coefficients and the production line configuration list into the MIP model, and obtain the optimal combination scheme corresponding to each set of business strategy weights by performing weighted solutions according to multiple preset business strategy weights under the constraints of filtering conditions and constraints; the filtering conditions and the constraints are determined based on the decision data and the decision constraints. The first output module is used to determine, based on the solved optimal combination schemes, the cost of executing each of the optional parameters under the optimal combination schemes, and output the cost if the optimal combination scheme exists. The second output module is used to control the MIP model to re-solve if the optimal combination scheme does not exist, by reducing at least one of the filtering conditions and / or the constraints, and to locate and output the defects of the production line configuration list based on the solution results.
[0013] In one feasible implementation, the loss calculation formula includes a formula for calculating losses from new employee training, a formula for calculating losses from production ramp-up, and a formula for calculating losses from rework. The loss calculation formula is triggered only when the shift changes from single shift to double shift in the current month; the production ramp-up loss calculation formula includes a bilinear term multiplied by the actual working hours and the manpower of the current month; the rework loss calculation formula includes a bilinear term multiplied by the JPH value and the actual working hours; the partial variable is the actual working hours; and the average value is the average value of multiple selectable actual working hours in the decision data.
[0014] In a feasible implementation plan, a weighted solution is performed according to multiple preset business strategy weights, including: The minimum sum of the products of each cost indicator and its corresponding business strategy weight is used as the optimization objective. An overcapacity penalty is imposed to construct a weighted objective function. The cost indicators include overtime cost, energy cost, monthly salary cost, recruitment and training cost, and shift changeover loss. The relative importance of each cost indicator is different under different business strategy weights. The combination scheme corresponding to the minimum weighted objective function value of each set of business strategy weights is the optimal combination scheme under that business strategy weight.
[0015] In a feasible implementation, each set of business strategy weights in the multiple sets of business strategy weights includes: overtime cost weight, energy consumption cost weight, monthly human resource salary cost weight, recruitment and training cost weight, and shift changeover loss weight, and the sum of each weight is 1; the preset multiple sets of business strategy weights include at least two of the following sets: The first strategy weighting is based on overtime cost, which has the highest weighting. The second strategy weighting has the highest weight for energy consumption cost. The third strategy weighting has the highest weight for monthly human resource salary costs. The fourth strategy weighting has the highest weight for recruitment and training costs; The fifth strategy weighting is based on the shift switching loss weighting, which has the highest weighting.
[0016] In one feasible implementation, the filtering conditions include at least one of the following: The first filtering condition indicates that the determined shift matches the shift type set in the production line configuration list for the current month. The second filtering condition indicates that the number of working days does not exceed the total number of days in the month; The third filtering condition indicates that the production capacity demand of each model does not exceed the maximum hourly output JPH of that model. The fourth filtering condition indicates that the total production capacity is not less than the sum of the demand and the surplus required by the production line configuration list; The fifth filtering condition characterizes the positive correlation between working days and working hours.
[0017] In one feasible implementation, the constraints include at least one of the following: The first constraint represents the flow relationship between the number of employees at the beginning of each month and the number of employees in the previous month, the number of employees who left, and the number of employees hired this month. The second constraint indicates that only one hourly output JPH pattern is selected each month. The third constraint indicates that the actual number of employees each month is not less than the sum of the required number of employees and the additional number of employees required by the selected JPH mode. The fourth constraint indicates that monthly production capacity must not be less than the sum of demand and surplus. The fifth constraint indicates that the selected JPH mode each month will be used as the basis for calculating the shift changeover loss, and the shift changeover loss will be used to participate in the weighted solution calculation. The sixth constraint indicates that overtime pay only applies to the selected JPH mode; The seventh constraint indicates that the monthly excess capacity is not less than the difference between the actual capacity and the minimum demand; The eighth constraint characterizes the cumulative effect of shift changeover losses corresponding to new employee training losses, production ramp-up losses, and rework losses across all months.
[0018] In one feasible implementation, the second output module is used to locate and output the defects in the production line configuration list based on the solution results, including: The filter conditions and / or constraints that were reduced during the re-solution process are identified as bottleneck conditions, and the corresponding month information, condition category, and gap value are output.
[0019] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the production decision method as described in any of the first aspects.
[0020] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the production decision-making method as described in any of the first aspects.
[0021] This application provides a production decision-making method, apparatus, electronic device, and storage medium. By fixing some variables in the nonlinear loss calculation formula to the average value of optional parameters and eliminating variable multiplication terms, the switching loss can be incorporated into the MIP model in a linear form for optimization. By setting filtering and constraint conditions in the MIP model, invalid combinations are eliminated before solving, reducing the solution scale. By using the weighted summation of multiple cost indicators and corresponding business strategy weights as the optimization objective, the MIP model can output multiple optimal combination schemes according to different strategy preferences. Furthermore, the cost is recalculated using the actually selected optional parameters to ensure that the final output cost is consistent with the actual business scope. Furthermore, by relaxing the filtering conditions and / or constraints and resolving them when no feasible solution is found, specific defects in the production line configuration list can be located and output.
[0022] Compared with existing technologies that use fixed penalty values to approximate switching losses, only output a single solution, and only report errors when there is no solution but cannot locate the cause, the embodiments of this application can accurately model and optimize production line switching losses, output the optimal solution under multiple business strategies, and automatically locate data defects when there is no solution, thereby improving the accuracy of production decisions and the efficiency of problem investigation.
[0023] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description
[0024] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0025] Figure 1A flowchart of a production decision-making method provided in an embodiment of this application is shown.
[0026] Figure 2 A flowchart of another production decision-making method provided by an embodiment of this application is shown.
[0027] Figure 3 A schematic diagram of the structure of a production decision-making device provided in an embodiment of this application is shown.
[0028] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation
[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0030] Factory production system planning refers to the comprehensive decision-making process within each planning cycle (e.g., month) regarding production line takt time (JPH, hourly output), shift patterns (single / double shift), number of working days, working hours, and employee recruitment and turnover, aiming to reduce overall costs such as energy consumption, labor, and overtime while meeting order demands. Because these decision variables are interconnected, some industry solutions currently employ mixed integer programming (MIP) models to model and solve this problem.
[0031] However, existing MIP production planning schemes generally have the following shortcomings: when a production line is upgraded from a single shift to a double shift, real financial costs such as new employee training losses, ramp-up losses, and rework losses are incurred. Among these, the calculation formulas for ramp-up losses and rework losses contain some bilinear terms, meaning that there are two decision variables multiplied together. This makes it impossible to directly incorporate them into the objective function of the linear MIP model, thus failing to obtain production decisions that meet the requirements.
[0032] Based on this, embodiments of this application provide a production decision-making method, apparatus, electronic device, and storage medium, which are described below through embodiments.
[0033] To facilitate understanding of this embodiment, the production decision-making method disclosed in this application will first be described in detail. For example... Figure 1 As shown, it includes the following steps: Step 101: Obtain the production line configuration list; the production line configuration list includes the decision data and decision constraints required by the factory to make production decisions in each month.
[0034] In one embodiment, the production line configuration list can be obtained when the factory begins to formulate its monthly production plan. This list can originate from the factory's production management system, enterprise resource planning system, or be compiled and entered by relevant personnel based on actual production conditions. The data and constraints contained in this list reflect the factory's objective conditions and business rules for the next one or more planning cycles, serving as the foundational input for the entire production decision-making process.
[0035] The decision-making data includes various parameters that objectively exist in the factory's production process and need to be met or relied upon. For example, the production speed, required number of workers, and energy consumption levels corresponding to different JPH modes of the production line; the order demand, capacity surplus rate, statutory working days, total number of days, and number of existing models to be completed each month; whether each day is a workday, rest day, or holiday; the unit price of energy at different times and different JPH levels; the worker training cycle and the proportion of trainees; the new employee turnover rate, ramp-up days, ramp-up JPH ratio, defect rate changes, and rework unit price involved in production line switching; the factory's statutory working hours, overtime unit price, employee monthly salary, initial manpower at the beginning of the month, and maximum JPH for each model. These data can be historical values obtained from actual factory operations or fixed values given according to industry standards or equipment manuals.
[0036] Decision constraints are used to define the business rules that production plans must meet. For example, a factory may only use single shifts, double shifts, or both in a given month, depending on the shift type; the maximum number of people a factory can hire within a planning period; and statutory working hours stipulate the maximum daily or monthly working hours for workers. These constraints are usually internal factory management regulations or labor law requirements and may vary between different factories.
[0037] In this embodiment, decision data and constraints are not required to be used in every step, but rather serve as the data foundation for subsequent MIP model construction and solution. During the MIP model construction process, relevant data or constraints are extracted from the data as needed for computation. Some decision data is also retrieved separately during the linearization stage to calculate linear coefficients. After all data is acquired, subsequent steps can be carried out based on this foundation.
[0038] In one embodiment, the production line configuration list contains eight data tables, each recording different dimensions of basic production information, which can be used to determine decision data and decision constraints.
[0039] The JPH mode table records the various combinations of production parameters that can be selected during production line operation. Each row of data includes the factory number, the corresponding JPH value for that mode, the required number of personnel for that JPH value, the number of shifts, and the JPH value and energy consumption level corresponding to single or double shifts. The number of shifts indicates how many shifts the production line will operate on in a day, and can be 1 (single shift) or 2 (double shift). The energy consumption level measures the production line's consumption of water, electricity, gas, and other energy resources under the corresponding JPH value and shift conditions. The same production line can have different energy consumption levels at different JPH levels, and the first and second shifts can use different JPH values and energy consumption levels.
[0040] The monthly data table records the production tasks and basic conditions of each factory in each month. Each row of data includes the factory number, month, target output, capacity surplus rate, statutory working days, total number of days in the month, shift type, additional manpower, turnover rate, and the quantity of each type of vehicle (e.g., the quantity of H53, H56, H97, and other models). The target output represents the total output required for the month, and the capacity surplus rate represents the proportion of surplus capacity that needs to be reserved during production scheduling. The shift type limits the shift format that the production line can use in the month; it can be single shift, double shift, or a combination of both. The additional manpower indicates the number of extra personnel needed in the month, and the turnover rate represents the expected percentage of manpower turnover in the month. The quantity of each type of vehicle is used to calculate whether the capacity requirements for each vehicle model are met.
[0041] The daily data table records the date type information for each factory for each day of the month. Each row contains the factory number, month, date, date type, and day of the week. The date type identifies whether the day is a workday, rest day, or public holiday. This data is used for subsequent daily plan breakdown and refined energy cost calculation, as energy prices and working hours differ for different date types.
[0042] The energy price table records multi-dimensional energy unit prices under different JPH levels, months, and energy types. Each row of data includes the plant number, JPH value, month, energy type, and energy consumption unit price under various scenarios, including production energy consumption unit price, nighttime energy consumption unit price, dining energy consumption unit price, short holiday energy consumption unit price, and long holiday energy consumption unit price, as well as fixed costs. The long holiday energy consumption unit price is used for energy consumption calculations when the number of consecutive rest days reaches or exceeds a certain number. Different energy types include water, electricity, and gas. This table provides all the unit price data required for energy cost calculation.
[0043] The training parameter table records relevant parameters for human resource training under different training types. Each row of data includes the factory number, training type, percentage of trainees, and number of training days. This table is used to calculate recruitment and training costs and determine the time required for new employees to complete training.
[0044] The loss configuration table records all the parameters required for calculating the three losses when a production line switches from a single shift to a double shift. Each row of data includes the factory number, new employee turnover rate, training period, training unit price, ramp-up days, ramp-up JPH ratio, ramp-up labor unit price, ramp-up capacity consumption unit price, pre-switch defect rate, post-switch defect rate, and rework unit price. The ramp-up JPH ratio represents the proportion of normal JPH that the production line can achieve during the ramp-up period, and its value is less than 1. The parameters in this table are used in subsequent steps to calculate the new employee training loss coefficient, ramp-up loss coefficient, and rework loss coefficient.
[0045] The common parameter table records global limit parameters shared by multiple factories, including the maximum number of people that can be hired per month. This parameter represents the maximum number of people that all factories can hire in a month, and is used to constrain the flow of personnel.
[0046] The factory parameter table records the fixed parameter information for each factory. Each row of data includes the factory number, statutory working hours, overtime rate, employee monthly salary, initial number of employees at the beginning of the month, and the maximum JPH value for each vehicle model. Statutory working hours are used to determine whether the number of working days and hours meet legal requirements; overtime rate and employee monthly salary are used to calculate overtime costs and monthly labor costs; the initial number of employees at the beginning of the month serves as the initial value for labor turnover calculations; and the maximum JPH value for each vehicle model is used to determine the maximum production capacity of the factory when producing a certain type of vehicle. These eight tables together constitute a complete production line configuration list input.
[0047] Each record in the JPH mode table corresponds to a selectable production mode for a production line. The JPH value represents the production rate of the production line per unit time and is a core indicator for measuring production line efficiency. The same production line can have different JPH values under different production modes. Generally, the higher the JPH value, the more products are produced per unit time, but the required number of workers and energy consumption will also increase accordingly. The number of shifts can be single or double, indicating how many shifts the production line will operate in a day. When the number of shifts is double, the JPH value and corresponding energy consumption level of the first and second shifts need to be recorded separately. The two shifts can use the same or different JPH values. The energy consumption level measures the level of energy consumption such as water, electricity, and gas during the operation of the production line. Different JPH levels and shift types can correspond to different energy consumption levels, which will serve as the basis for subsequent energy cost calculations. The JPH mode table also includes a corresponding number of workers field, indicating the number of workers required for the production line to operate normally under the given JPH value and shift conditions. This number of workers will serve as the basis for judging whether the manpower requirements are met in subsequent constraints.
[0048] Step 102: Fix some variable values in the nonlinear loss calculation formula of the mixed integer programming (MIP) model to be input to the average value of multiple optional parameters set in the decision data, so as to convert the nonlinear loss calculation formula into a linear loss calculation formula containing only one variable, and output the corresponding linear coefficient.
[0049] For example, the loss calculation formula includes a new employee training loss calculation formula, a production ramp-up loss calculation formula, and a rework loss calculation formula. The loss calculation formula is triggered only when the shift changes from single-shift to double-shift in the current month; the production ramp-up loss calculation formula includes a bilinear term calculated by multiplying the actual working hours by the total manpower for the current month, and the rework loss calculation formula includes a bilinear term calculated by multiplying the JPH value by the actual working hours; the partial variable is the actual working hours, and the average value is the average of multiple selectable actual working hours in the decision data.
[0050] In other words, in this embodiment of the application, when the factory upgrades from a single shift to a double shift, three losses will occur: new employee training loss, ramp-up loss, and rework loss. The new employee training loss itself contains only a single variable and can be directly incorporated into the MIP model in a linear form. The ramp-up loss and rework loss, because they involve the multiplication of two variables (e.g., actual working hours multiplied by the number of workers this month, or JPH value multiplied by actual working hours), are non-linear terms and cannot be directly included in the objective function of the linear MIP model.
[0051] The calculation methods for the three losses are as follows.
[0052] The formula for calculating losses from new employee training is: New employee training loss = (This month's manpower - Last month's manpower) × New employee turnover rate × Training cycle × Training unit price.
[0053] Except for the human resources variable, all other variables in this formula are constants (the production line configuration list will provide the values for other data), therefore it is inherently a linear expression. Extracting the constant components as the new employee training coefficient, the above formula becomes: New employee training loss = New employee training coefficient × (This month's manpower - Last month's manpower).
[0054] The formula for calculating creepage loss is: Loss due to overrunning production = excess man-hours × man-hours this month × (labor cost per unit of overrunning production + labor cost per unit of overrunning production).
[0055] The excess man-hours are calculated as: (Number of days of ramp-up production × Actual working hours) / (Ramp-up production JPH ratio) - (Number of days of ramp-up production × Actual working hours). Since the actual working hours are multiplied by the total manpower for the month, a bilinear term is formed. In actual production, working hours can be selected from several optional values, such as 7 hours, 8 hours, 9 hours, or 10 hours. To eliminate the multiplicative term, the arithmetic mean of these optional working hour values is taken as a fixed constant. After fixing the actual working hours at 8.5 hours, the excess man-hours become a constant, and the ramp-up production loss degenerates into a linear expression containing only the manpower for the month. Extracting the constant part as the ramp-up production coefficient, the above calculation formula is transformed into: Production overrun loss = production overrun coefficient × this month's manpower.
[0056] The formula for calculating repair losses is: Rework loss = JPH value × JPH ratio for ramp-up production × number of ramp-up days × actual working hours × rework unit price × (defect rate after shift change - defect rate before shift change).
[0057] In this formula, the JPH value is multiplied by the actual working hours, forming a bilinear term. After fixing the actual working hours at 8.5 hours, the rework loss degenerates into a linear expression containing only the JPH value. Extracting the constant part as the rework coefficient, the calculation formula is expressed as: Repair loss = Repair coefficient × JPH value selected this month.
[0058] The rationale for using the average value of working hours instead of an extreme value is that, when solving the MIP model, actual working hours will not always fall on extreme values, and the average value can statistically be closer to the expected level of actual decisions. By fixing the working hours, all three losses are transformed into linear expressions. The new employee training loss is already in linear form, so the coefficient is directly extracted; after eliminating the bilinear terms, the production ramp-up loss and rework loss can also be extracted as constant coefficients that are multiplied only by the monthly manpower or JPH value.
[0059] The three linear coefficients output in this step are all scalar constants and can be directly fed into the objective function of the MIP model in subsequent steps. During the MIP model solution process, these coefficients participate in the weighted summation calculation of the objective function to guide the solver in selecting the option with the smaller switching loss. Since the coefficients are calculated under the premise of a fixed average working time, their values deviate from the actual loss. Therefore, the calculation results in this step are only used for directional guidance in the MIP solution process and are not used as the final financial cost output. The final actual cost will be recalculated accurately in subsequent post-processing steps using the actual selected working time values.
[0060] Step 103: Input the linear coefficients and the production line configuration list into the MIP model, and obtain the optimal combination scheme corresponding to each set of business strategy weights by performing weighted solutions according to preset multiple sets of business strategy weights under the constraints of filtering conditions and constraints; the filtering conditions and constraints are determined based on the decision data and the decision constraints.
[0061] The three linear coefficients calculated in step 102 are input into the MIP model along with the relevant data from the production line configuration list. Before the MIP model begins its formal solution process, candidate combinations are first screened based on filtering conditions to reduce the variable scale during subsequent solutions. In one embodiment, a candidate combination refers to all possible (JPH mode, number of working days, working hours) triples for each month. Each triple represents a potential production scheduling method for that month.
[0062] The filtering conditions are set as follows. For each month, the MIP model enumerates all possible combinations of JPH patterns, working days, and working hours, and then checks whether each combination meets the various filtering conditions. There can be one or more filtering conditions.
[0063] For example, the filtering conditions include at least one of the following: The first filtering condition indicates that the determined shift matches the shift type set in the production line configuration list for the current month; the second filtering condition indicates that the number of working days does not exceed the total number of days in the current month; the third filtering condition indicates that the production capacity demand of each model does not exceed the maximum hourly output JPH of that model; the fourth filtering condition indicates that the total production capacity is not less than the sum of the demand and the surplus required by the production line configuration list; the fifth filtering condition indicates that the number of working days is positively correlated with working hours.
[0064] The first filtering condition is used to check whether the shifts in the candidate combinations match the shift types set in the production line configuration list for the current month. If the shift type for the current month is single shift, all double shift modes will be eliminated; if the shift type for the current month is double shift, all single shift modes will be eliminated; if the shift type for the current month is either single or double, both single and double shift modes will be retained.
[0065] The second filtering condition checks whether the number of working days in the candidate combination does not exceed the total number of days in the current month. The total number of days in the current month is provided by the total number of days for this month field in the monthly data table. If the number of working days in a combination exceeds this value, the combination is rejected.
[0066] The third filtering condition is used to check whether the production capacity demand of each model exceeds the maximum allowable JPH value for that model. Specifically, for models such as H53, H56, and H97, the monthly demand for that model is divided by the product of the maximum JPH value for that model and the number of shifts, working hours, and working days in the candidate combination. If the result is greater than 1, it means that the combination cannot meet the production capacity demand for that model, and the combination is eliminated. This check is performed independently for each model.
[0067] The fourth filtering condition is used to check whether the total capacity is lower than the sum of demand and surplus. Specifically, the total capacity of the candidate combination is obtained by multiplying the JPH value of the candidate combination by the working hours and the number of working days. Then, the total capacity is compared with the result of multiplying the demand by (1 plus the surplus rate) and rounding up. If the total capacity is less than the result, it means that the combination cannot meet the demand including surplus capacity, and the combination is eliminated.
[0068] The fifth filtering condition is used to check whether the correspondence between working days and working hours is reasonable. Specifically, when the number of working days in a candidate combination is less than the statutory number of days, the working hours must not exceed the statutory working hours; when the number of working days is not less than the statutory number of days, the working hours must not be less than the statutory working hours. Combinations that do not meet this correspondence are eliminated.
[0069] All of the above filtering conditions can be used, or some can be selected according to actual needs. Any combination that fails to pass all the set filtering conditions is eliminated before the solution is obtained and will not proceed to the subsequent variable creation and solution process. In this way, the number of candidate combinations is significantly reduced, and the computational burden during the subsequent MIP model solution is also reduced.
[0070] After filtering, the MIP model constructs a complete optimization model based on the remaining candidate combinations. This model includes multiple constraints to ensure that the final output solution is business-compliant. The specific content of the constraints can be flexibly configured according to the actual management needs of the factory; different factories can choose different combinations of constraints.
[0071] For example, the constraints include at least one of the following: The first constraint represents the flow relationship between the number of employees at the beginning of each month, the number of employees from the previous month, the number of employees who left, and the number of employees hired this month. The second constraint represents that only one hourly output JPH mode is selected each month. The third constraint represents that the actual number of employees each month is not less than the sum of the required number of employees and the additional number of employees required by the selected JPH mode. The fourth constraint represents that the monthly production capacity is not less than the sum of the demand and the surplus. The fifth constraint represents that the selected JPH mode each month is used as the basis for calculating shift changeover losses, and the shift changeover losses are used in the weighted solution calculation. The sixth constraint represents that overtime pay is only effective for the selected JPH mode. The seventh constraint represents that the monthly excess production capacity is not less than the difference between the actual production capacity and the minimum demand. The eighth constraint represents the cumulative effect of shift changeover losses corresponding to new employee training losses, ramp-up losses, and rework losses across all months.
[0072] The first constraint defines the flow of manpower at the beginning of each month between the previous month's manpower, the number of employees who left, and the number of people hired this month. Specifically, the manpower at the beginning of this month equals the previous month's manpower minus the number of employees who left last month, plus the number of people hired this month. The number of employees who left is determined by multiplying the previous month's manpower by the turnover rate and rounding up. The manpower at the beginning of the first month equals the initial manpower at the beginning of the month plus the number of people hired in the first month.
[0073] The second constraint ensures that only one JPH mode is selected each month. Specifically, the sum of the binary selection variables corresponding to all candidate combinations for the month equals 1, ensuring that the solution outputs only one unique scheduling method for each month.
[0074] The third constraint is to ensure that the actual monthly manpower is not less than the sum of the required manpower and additional manpower for the selected JPH model. Specifically, the actual monthly manpower must be greater than or equal to the authorized number of employees for the selected model plus the additional manpower for that month; otherwise, the plan cannot meet the manpower requirements.
[0075] The fourth constraint is used to limit monthly production capacity to no less than the sum of demand and surplus. Specifically, the JPH value of the selected combination multiplied by working hours and working days must be greater than or equal to the demand multiplied by (1 plus surplus rate) and rounded up.
[0076] The fifth constraint assigns the selected JPH mode and number of shifts for each month to the corresponding auxiliary variables. These auxiliary variables are used to calculate the rework portion of the shift changeover loss. Specifically, the selected JPH value for the month is recorded in the corresponding JPH auxiliary variable, and the selected number of shifts for the month is recorded in the corresponding shift auxiliary variable. The objective function then calculates the rework loss using these auxiliary variables.
[0077] The sixth constraint limits overtime pay to the selected JPH pattern only. Specifically, by introducing a sufficiently large constant, when a candidate combination is not selected, the overtime pay for that combination is forced to zero; when a candidate combination is selected, the overtime pay for that combination equals the monthly workforce multiplied by the overtime rate. Overtime pay will not be generated for combinations that are not selected.
[0078] The seventh constraint is used to ensure that monthly excess capacity is not less than the difference between actual capacity and minimum demand. Specifically, if the difference between the actual capacity and the minimum demand for the month is greater than zero, this difference is recorded as excess capacity and used as the excess capacity penalty term in the objective function.
[0079] The eighth constraint characterizes the cumulative effect of shift changeover losses corresponding to new employee training losses, production ramp-up losses, and rework losses across all months. Specifically, each month, the difference between the current month's and previous month's manpower, the current month's manpower value, and the selected JPH value are multiplied by the three linear coefficients output in step 102 to obtain surrogate values for the three losses. These surrogate values are accumulated in the objective function across all months and used as approximations of shift changeover losses in a weighted solution, guiding the solver to favor the option with the smaller changeover loss.
[0080] Some or all of the above constraints can be selected based on the actual business rules of the factory. Unselected constraints will not be added during model construction. The MIP model will only solve based on the actually added constraints.
[0081] After the model is built, the MIP model will perform weighted solving according to multiple pre-set sets of business strategy weights. Each set of business strategy weights corresponds to a different set of cost indicator weights, such as overtime cost weight, energy cost weight, monthly salary cost weight, recruitment and training cost weight, and shift switching loss weight. In the same set of weights, the higher the weight of a certain cost, the higher the priority of that cost indicator during optimization, and the more the solver will tend to reduce that cost when searching for the optimal solution. By solving multiple sets of business strategy weights separately, the MIP model can output multiple different optimal combinations, each corresponding to different business strategy preferences.
[0082] Ultimately, the MIP model outputs the optimal combination of weights for each set of business strategies. Each solution includes information such as the selected JPH mode for each month, the number of working days, working hours, and the corresponding number of personnel and recruitment positions. These solutions serve as inputs for subsequent steps, used for further calculations and outputs.
[0083] Step 104: If the optimal combination scheme exists, determine the cost of executing each of the optional parameters under the optimal combination scheme according to the solved optimal combination schemes, and output the cost.
[0084] In one embodiment, after the MIP model successfully solves and outputs the optimal combination scheme corresponding to the weights of each set of business strategies, post-processing is required for each scheme to obtain cost data consistent with the actual business scope.
[0085] The post-processing includes shift changeover detection, accurate recalculation of changeover losses, daily plan breakdown, and precise calculation of energy consumption costs.
[0086] Shift changeover detection is used to determine whether a change in shift changeover losses has occurred between adjacent months. Specifically, the number of shifts selected for each month is extracted from each optimal combination, and then each pair of adjacent months is checked sequentially. The calculation condition for shift changeover losses is only triggered if the selected shifts in a given month are both double shifts and the selected shifts in the previous month were single shifts. Other cases, including shifts remaining unchanged and shifts decreasing from double to single shifts, do not trigger shift changeover loss calculations.
[0087] The precise recalculation of switching losses is used to recalculate the three losses for the month using the actual selected work hours (instead of the arithmetic mean) when the triggering conditions are met. In step 102, the calculation of ramp-up losses and rework losses uses a fixed average work hour of 8.5 hours, which is only used for directional guidance during the MIP solution process. In the post-processing stage, the actual selected work hours for the month are extracted from the optimal combination scheme. This work hour value can be any one of 7, 8, 9, or 10 hours, depending on the specific scheme selected during the MIP model solution. Then, according to the original calculation formulas for the three losses, the actual work hours are substituted to recalculate the new employee training loss, ramp-up loss, and rework loss. For months where shift switching is not triggered, all three losses are counted as zero. In this way, the final output of switching losses is consistent with the actual costs incurred in the business.
[0088] Daily plan breakdown is used to allocate the total monthly output to specific days according to the workday schedule. Specifically, based on the number of working days selected in the optimal combination plan for the month, combined with the date type information in the daily data table, it is determined which dates are scheduled as working days. If the number of working days is greater than the legally mandated number of days, it indicates that there is overtime work in that month, and the extra working days are prioritized to be scheduled on Saturdays, Sundays, or public holidays. If the number of working days is less than the legally mandated number of days, it indicates that there is time off in lieu of leave in that month, and Fridays, Mondays, Thursdays, etc., are prioritized for rest. If there are fewer than five working days, they are concentrated on the earliest working day of the month. After completing the workday allocation, the daily output for each day is calculated based on the total monthly output and the JPH value.
[0089] Energy cost actuarial calculation is used to accurately calculate the energy costs for each dimension of the month based on the results of daily plan breakdown. Specifically, according to the date type of each day and the selected JPH mode and working hours for the month, the corresponding energy unit price is matched from the energy price table, including the production energy unit price, nighttime energy unit price, dining energy unit price, short holiday energy unit price, and long holiday energy unit price. Among them, long holidays are determined when the number of consecutive rest days reaches or exceeds five days, and the long holiday energy unit price is used for calculation. The energy costs of each day are added up to obtain the accurate energy cost for the month.
[0090] After completing the above post-processing calculations, the cost details of all options are aggregated by option number to form a multi-option comparison result, which is then output. The output includes the selected JPH mode, working days, working hours, number of personnel, number of recruits, detailed cost items, and total comprehensive cost for each option, allowing decision-makers to select the final implementation option based on actual business preferences.
[0091] Step 105: If the optimal combination scheme does not exist, control the MIP model to resolve by reducing at least one of the filtering conditions and / or the constraints, and locate and output the defects of the production line configuration list based on the solution results.
[0092] For example, such as Figure 2 As shown, based on the solution results, the defects in the production line configuration list are located and output, including: Step 201: Identify the filtering conditions and / or constraints that were reduced during the re-solution process as bottleneck conditions, and output the month information, condition category, and gap value corresponding to the bottleneck conditions.
[0093] After the MIP model is re-solved, the filter conditions or constraints that were reduced during the re-solution process become the bottleneck conditions that caused the original model to have no solution. Specifically, the original MIP model in step 103 is solved with all filter conditions and all constraints in effect simultaneously. If the solution process returns no feasible solution, in step 105, the diagnostic model will reduce the filter conditions or constraints one by one and observe whether the model can obtain a feasible solution after removing a certain condition. If the model has a solution after a certain condition is reduced, then that condition is determined to be the bottleneck condition that caused the original model to have no solution.
[0094] For each filter condition or constraint identified as a bottleneck condition, the following information should be output: Monthly information indicates which specific month the bottleneck condition corresponds to. For example, when a capacity constraint in a certain month results in no solution, the specific month number or year-month identifier is output.
[0095] The condition category indicates whether the bottleneck condition is a filtering condition or a constraint condition, and what type of condition it is. For example, outputting category names such as "Personnel Mobility Constraint" or "Fourth Filter Condition" allows users to quickly identify the type of parameter that needs to be adjusted.
[0096] The gap value indicates the specific degree to which the bottleneck condition is violated. For example, when the manpower constraint leads to no solution, the output shows the specific number of people missing in that month; when the capacity constraint leads to no solution, the output shows the specific difference in capacity shortage in that month; when the recruitment limit constraint leads to no solution, the output shows the specific number of people exceeding the maximum recruitment limit in that month.
[0097] The aforementioned monthly information, condition categories, and gap values are combined into a business-readable diagnostic message and output. This message can be displayed in text format in the user interface, written to a log file, or pushed to the upstream system via an interface. Based on this diagnostic message, users can quickly locate specific data items in the production line configuration list that need adjustment, such as increasing the recruitment cap for a certain month, decreasing the demand for a certain month, or adjusting shift type settings, and then resubmit the solution after correcting the data.
[0098] This application provides a production decision-making method, apparatus, electronic device, and storage medium. By fixing some variables in the nonlinear loss calculation formula to the average value of optional parameters and eliminating variable multiplication terms, the switching loss can be incorporated into the MIP model in a linear form for optimization. By setting filtering and constraint conditions in the MIP model, invalid combinations are eliminated before solving, reducing the solution scale. By using the weighted summation of multiple cost indicators and corresponding business strategy weights as the optimization objective, the MIP model can output multiple optimal combination schemes according to different strategy preferences. Furthermore, the cost is recalculated using the actually selected optional parameters to ensure that the final output cost is consistent with the actual business scope. Furthermore, by relaxing the filtering conditions and / or constraints and resolving them when no feasible solution is found, specific defects in the production line configuration list can be located and output.
[0099] Compared with existing technologies that use fixed penalty values to approximate switching losses, only output a single solution, and only report errors when there is no solution but cannot locate the cause, the embodiments of this application can accurately model and optimize production line switching losses, output the optimal solution under multiple business strategies, and automatically locate data defects when there is no solution, thereby improving the accuracy of production decisions and the efficiency of problem investigation.
[0100] In a feasible implementation plan, a weighted solution is performed according to multiple preset business strategy weights, including: The minimum sum of the products of each cost indicator and its corresponding business strategy weight is used as the optimization objective. An overcapacity penalty is imposed to construct a weighted objective function. The cost indicators include overtime cost, energy cost, monthly salary cost, recruitment and training cost, and shift changeover loss. The relative importance of each cost indicator is different under different business strategy weights. The combination scheme corresponding to the minimum weighted objective function value of each set of business strategy weights is the optimal combination scheme under that business strategy weight.
[0101] The MIP model uses the minimum sum of the products of various cost indicators and their corresponding business strategy weights as its optimization objective, while also imposing overcapacity penalties to construct a weighted objective function. The cost indicators include overtime costs, energy costs, monthly employee salaries, recruitment and training costs, and shift changeover losses. The relative importance of each cost indicator varies under different business strategy weights. The optimal combination scheme under each business strategy weight is the one that minimizes the weighted objective function value.
[0102] The expression for the weighted objective function is:
[0103] The meanings of each symbol are as follows: This indicates the overtime cost weight corresponding to the overtime cost; This represents overtime cost, which is the total amount of overtime pay corresponding to the selected work hour combination in the current month.
[0104] This indicates the energy cost weight corresponding to the energy cost; This represents the energy consumption cost, which is the energy cost corresponding to the selected JPH mode and working hour combination in the current month. It is calculated by accumulating multiple dimensions such as production, night, meals, short holidays, and long holidays.
[0105] This indicates the weight of the monthly human resource cost corresponding to the monthly human resource cost. This represents the monthly labor cost, which is the actual number of employees in the current month multiplied by the monthly salary per employee.
[0106] This indicates the recruitment and training cost weight corresponding to the recruitment and training cost. This represents recruitment and training costs, which is the number of people recruited in the current month multiplied by the average training cost.
[0107] This indicates the shift change loss weight corresponding to the shift change loss. The shift switching loss represents the new employee training loss, production ramp-up loss, and rework loss incurred when a single shift is upgraded to a double shift in the current month. The linear coefficient calculated in step 102 is used as a proxy term in the objective function calculation.
[0108] This represents a penalty for overcapacity, calculated as a fixed coefficient within the objective function, without any business strategy weighting. Its value is the difference between the actual monthly capacity and the minimum capacity requirement, multiplied by the fixed penalty coefficient. A penalty is applied when actual capacity exceeds the minimum requirement; otherwise, the penalty is zero. This overcapacity penalty is not adjusted by business strategy weighting and is applied with the same intensity under all strategies. The purpose is to prevent MIP from excessively over-producing, leading to increased costs.
[0109] The higher the weight of a cost indicator, the more important that indicator is in the optimization objective, and the more the MIP model will tend to reduce that cost when solving the problem. For example, if the weight of overtime cost is higher... If set to a higher value, the MIP model will prioritize combinations with lower working hours or less overtime pay; if shift switching loss weights are applied... Setting the value to a higher value will cause the MIP model to prioritize avoiding the switching operation from a single shift to a double shift. By configuring multiple different weight combinations for the same set of input data and solving them separately, the MIP model can output multiple optimal combination schemes corresponding to different business preferences, allowing decision-makers to choose according to actual needs.
[0110] In a feasible implementation, each set of business strategy weights in the multiple sets of business strategy weights includes: overtime cost weight, energy consumption cost weight, monthly human resource salary cost weight, recruitment and training cost weight, and shift changeover loss weight, and the sum of each weight is 1; the preset multiple sets of business strategy weights include at least two of the following sets: In the first strategy weighting, overtime cost has the highest weight, indicating that this strategy prioritizes reducing overtime costs. In the second strategy weighting, energy cost has the highest weight, indicating that energy costs are prioritized. In the third strategy weighting, monthly salary cost has the highest weight, indicating that labor costs are prioritized. In the fourth strategy weighting, recruitment and training cost has the highest weight, indicating that recruitment and training costs are prioritized. In the fifth strategy weighting, shift changeover loss has the highest weight, indicating that shift changeover loss is avoided. Different strategy weights correspond to different business preferences. The MIP model solves each strategy separately and outputs multiple solutions for decision-makers to choose from. The sum of the weights is 1 to ensure the comparability of the objective function values between different strategies.
[0111] Based on the same technical concept, embodiments of this application also provide a production decision-making device, such as... Figure 3 As shown, the device includes: The acquisition module 301 is used to acquire the production line configuration list; the production line configuration list includes the decision data and decision constraints required by the factory to make production decisions in each month.
[0112] The conversion module 302 is used to fix some variable values in the nonlinear loss calculation formula of the mixed integer programming (MIP) model to be input to the average value of multiple optional parameters set in the decision data, so as to convert the nonlinear loss calculation formula into a linear loss calculation formula containing only one variable and output the corresponding linear coefficient.
[0113] The input module 303 is used to input the linear coefficients and the production line configuration list into the MIP model, and obtain the optimal combination scheme corresponding to each set of business strategy weights by performing weighted solutions according to preset multiple sets of business strategy weights under the constraints of filtering conditions and constraints; the filtering conditions and the constraints are determined based on the decision data and the decision constraints.
[0114] The first output module 304 is used to determine, based on the solved optimal combination schemes, the cost of executing each of the optional parameters under the optimal combination schemes, and output the cost if the optimal combination scheme exists.
[0115] The second output module 305 is used to control the MIP model to re-solve if the optimal combination scheme does not exist, by reducing at least one of the filtering conditions and / or the constraints, and to locate and output the defects of the production line configuration list based on the solution results.
[0116] In one feasible implementation, the loss calculation formula includes a new employee training loss calculation formula, a production ramp-up loss calculation formula, and a rework loss calculation formula.
[0117] The loss calculation formula is triggered only when the shift changes from single shift to double shift in the current month; the production ramp-up loss calculation formula includes a bilinear term multiplied by the actual working hours and the manpower of the current month; the rework loss calculation formula includes a bilinear term multiplied by the JPH value and the actual working hours; the partial variable is the actual working hours; and the average value is the average value of multiple selectable actual working hours in the decision data.
[0118] In a feasible implementation plan, a weighted solution is performed according to multiple preset business strategy weights, including: The minimum sum of the products of each cost indicator and its corresponding business strategy weight is used as the optimization objective. An overcapacity penalty is imposed to construct a weighted objective function. The cost indicators include overtime cost, energy cost, monthly salary cost, recruitment and training cost, and shift changeover loss. The relative importance of each cost indicator is different under different business strategy weights. The combination scheme corresponding to the minimum weighted objective function value of each set of business strategy weights is the optimal combination scheme under that business strategy weight.
[0119] In a feasible implementation, each set of business strategy weights in the multiple sets of business strategy weights includes: overtime cost weight, energy consumption cost weight, monthly human resource salary cost weight, recruitment and training cost weight, and shift changeover loss weight, and the sum of each weight is 1; the preset multiple sets of business strategy weights include at least two of the following sets: The first strategy weighting is based on overtime cost, which has the highest weighting.
[0120] The second strategy weighting has the highest weight for energy consumption cost.
[0121] The third strategy weighting has the highest weight for monthly human resource salary costs.
[0122] The fourth strategy weighting is based on recruitment and training costs, which have the highest weighting.
[0123] The fifth strategy weighting is based on the shift switching loss weighting, which has the highest weighting.
[0124] In one feasible implementation, the filtering conditions include at least one of the following: The first filtering condition indicates that the selected shift matches the shift type set in the production line configuration list for the current month.
[0125] The second filtering condition indicates that the number of working days does not exceed the total number of days in the month.
[0126] The third filtering condition indicates that the production capacity demand of each model does not exceed the maximum hourly output JPH of that model.
[0127] The fourth filtering condition indicates that the total production capacity is not less than the sum of the demand and the surplus required by the production line configuration list.
[0128] The fifth filtering condition characterizes the positive correlation between working days and working hours.
[0129] In one feasible implementation, the constraints include at least one of the following: The first constraint represents the flow relationship between the number of employees at the beginning of each month, the number of employees in the previous month, the number of employees who left, and the number of employees hired this month.
[0130] The second constraint indicates that only one hourly output JPH pattern is selected each month.
[0131] The third constraint indicates that the actual number of employees each month is not less than the sum of the required number of employees and the additional number of employees required by the selected JPH pattern.
[0132] The fourth constraint indicates that monthly production capacity is not less than the sum of demand and surplus.
[0133] The fifth constraint indicates that the selected JPH mode each month is used as the basis for calculating the shift changeover loss, which is used to participate in the weighted solution calculation.
[0134] The sixth constraint indicates that overtime pay only applies to the selected JPH mode.
[0135] The seventh constraint indicates that the monthly excess capacity is not less than the difference between the actual capacity and the minimum demand.
[0136] The eighth constraint characterizes the cumulative effect of shift changeover losses corresponding to new employee training losses, production ramp-up losses, and rework losses across all months.
[0137] In one feasible implementation, the second output module is used to locate and output the defects in the production line configuration list based on the solution results, including: The filter conditions and / or constraints that were reduced during the re-solution process are identified as bottleneck conditions, and the corresponding month information, condition category, and gap value are output.
[0138] Figure 4 A schematic diagram of an electronic device provided in this application embodiment includes: a processor 401, a storage medium 402, and a bus 403. The storage medium 402 stores machine-readable instructions executable by the processor 401. When the electronic device runs the production decision method as described in the embodiment, the processor 401 communicates with the storage medium 402 via the bus 403, and the processor 401 executes the machine-readable instructions to perform the steps as described in the embodiment.
[0139] In this embodiment, the storage medium 402 may also execute other machine-readable instructions to perform other methods as described in the embodiment. For details on the specific execution steps and principles, please refer to the description of the embodiment, which will not be repeated here.
[0140] This application also provides a computer-readable storage medium storing a computer program that is executed by a processor to perform the steps as described in the embodiments.
[0141] In this embodiment, the computer program, when run by the processor, can also execute other machine-readable instructions to perform other methods as described in the embodiments. For details on the specific execution steps and principles, please refer to the description of the embodiments, which will not be repeated here.
[0142] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interface; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.
[0143] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0144] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0145] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.
[0146] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A production decision-making method, characterized in that, The method includes: Obtain the production line configuration list; the production line configuration list includes the decision data and decision constraints required by the factory to make production decisions in each month. The nonlinear loss calculation formula of the mixed integer programming (MIP) model to be input is fixed with some variable values as the average value of multiple optional parameters set in the decision data, so as to convert the nonlinear loss calculation formula into a linear loss calculation formula containing only one variable and output the corresponding linear coefficient. The linear coefficients and the production line configuration list are input into the MIP model. The MIP model is then used to obtain the optimal combination scheme corresponding to each set of business strategy weights under the constraints of filtering and constraint conditions, according to the preset weights of multiple business strategies. The filtering and constraint conditions are determined based on the decision data and the decision constraints. If the optimal combination scheme exists, then based on the solved optimal combination schemes, determine the cost of executing each of the optional parameters under the optimal combination scheme, and output it; If the optimal combination scheme does not exist, the MIP model is controlled to be re-solved by reducing at least one of the filtering conditions and / or the constraints, and the defects of the production line configuration list are located and output based on the solution results.
2. The method according to claim 1, characterized in that, The loss calculation formulas include formulas for calculating losses from new employee training, losses from production overruns, and losses from rework. The loss calculation formula is triggered only when the shift changes from single shift to double shift in the current month; the production ramp-up loss calculation formula includes a bilinear term multiplied by the actual working hours and the manpower of the current month; the rework loss calculation formula includes a bilinear term multiplied by the JPH value and the actual working hours; the partial variable is the actual working hours; and the average value is the average value of multiple selectable actual working hours in the decision data.
3. The method according to claim 1, characterized in that, The solution is obtained by weighting multiple pre-set business strategy weights, including: The minimum sum of the products of each cost indicator and its corresponding business strategy weight is used as the optimization objective. An overcapacity penalty is imposed to construct a weighted objective function. The cost indicators include overtime cost, energy cost, monthly salary cost, recruitment and training cost, and shift changeover loss. The relative importance of each cost indicator is different under different business strategy weights. The combination scheme corresponding to the minimum weighted objective function value of each set of business strategy weights is the optimal combination scheme under that business strategy weight.
4. The method according to claim 3, characterized in that, Each set of business strategy weights in the multiple sets of business strategy weights includes: overtime cost weight, energy cost weight, monthly salary cost weight, recruitment and training cost weight, and shift changeover loss weight, and the sum of each weight is 1; the preset multiple sets of business strategy weights include at least two of the following: The first strategy weighting is based on overtime cost, which has the highest weighting. The second strategy weighting has the highest weight for energy consumption cost. The third strategy weighting has the highest weight for monthly human resource salary costs. The fourth strategy weighting has the highest weight for recruitment and training costs; The fifth strategy weighting is based on the shift switching loss weighting, which has the highest weighting.
5. The method according to claim 1, characterized in that, The filtering conditions include at least one of the following: The first filtering condition indicates that the determined shift matches the shift type set in the production line configuration list for the current month. The second filtering condition indicates that the number of working days does not exceed the total number of days in the month; The third filtering condition indicates that the production capacity demand of each model does not exceed the maximum hourly output JPH of that model. The fourth filtering condition indicates that the total production capacity is not less than the sum of the demand and the surplus required by the production line configuration list; The fifth filtering condition characterizes the positive correlation between working days and working hours.
6. The method according to claim 1, characterized in that, The constraints include at least one of the following: The first constraint represents the flow relationship between the number of employees at the beginning of each month and the number of employees in the previous month, the number of employees who left, and the number of employees hired this month. The second constraint indicates that only one hourly output JPH pattern is selected each month. The third constraint indicates that the actual number of employees each month is not less than the sum of the required number of employees and the additional number of employees required by the selected JPH mode. The fourth constraint indicates that monthly production capacity must not be less than the sum of demand and surplus. The fifth constraint indicates that the selected JPH mode each month will be used as the basis for calculating the shift changeover loss, and the shift changeover loss will be used to participate in the weighted solution calculation. The sixth constraint indicates that overtime pay only applies to the selected JPH mode; The seventh constraint indicates that the monthly excess capacity is not less than the difference between the actual capacity and the minimum demand; The eighth constraint characterizes the cumulative effect of shift changeover losses corresponding to new employee training losses, production ramp-up losses, and rework losses across all months.
7. The method according to claim 1, characterized in that, Based on the solution results, locate and output the defects in the production line configuration list, including: The filter conditions and / or constraints that were reduced during the re-solution process are identified as bottleneck conditions, and the corresponding month information, condition category, and gap value are output.
8. A production decision-making device, characterized in that, The device includes: The acquisition module is used to acquire a production line configuration list; the production line configuration list includes decision data and decision constraints required by the factory to make production decisions in each month. The conversion module is used to fix some variable values in the nonlinear loss calculation formula of the mixed integer programming (MIP) model to be input to the average value of multiple optional parameters set in the decision data, so as to convert the nonlinear loss calculation formula into a linear loss calculation formula containing only one variable and output the corresponding linear coefficient. The input module is used to input the linear coefficients and the production line configuration list into the MIP model, and obtain the optimal combination scheme corresponding to each set of business strategy weights by performing weighted solutions according to multiple preset business strategy weights under the constraints of filtering conditions and constraints; the filtering conditions and the constraints are determined based on the decision data and the decision constraints. The first output module is used to determine, based on the solved optimal combination schemes, the cost of executing each of the optional parameters under the optimal combination schemes, and output the cost if the optimal combination scheme exists. The second output module is used to control the MIP model to re-solve if the optimal combination scheme does not exist, by reducing at least one of the filtering conditions and / or the constraints, and to locate and output the defects of the production line configuration list based on the solution results.
9. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the production decision method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the production decision-making method as described in any one of claims 1 to 7.
