Closed-loop manufacturing production decision optimization method based on robust reconstruction combined with chance constraints

Abstract

The method for closed-loop manufacturing production decision optimization based on robust reconstruction combined with opportunity constraint can obtain the probability information of uncertain parameters by analyzing the probability distribution and perturbation range of production data, and decision makers can set the risk level of the model according to external policies and economic environment, so that the robust closed-loop manufacturing scheduling scheme is obtained; specifically including three steps of collecting the data of closed-loop manufacturing process, establishing the uncertain closed-loop manufacturing model and establishing the robust reconstruction opportunity constraint optimization model. The present application can hedge the influence of uncertainty on the optimal solution of the model, and the decision maker can control the conservativeness of the solution by adjusting the confidence probability to cope with the fluctuation of the external environment, which helps to provide a resource-saving scheme, reduces production waste, reduces the demand for distribution information of uncertain parameters, improves the solving speed and optimization quality, and can optimize the production fluctuation in the manufacturing and remanufacturing process.

Description

Technical Field

[0001] This invention relates to the field of production management technology, and in particular to a closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints. Background Technology

[0002] In recent years, with the advocacy of green, environmentally friendly, and resource-saving strategies in the industrial sector, the effective application of closed-loop processes such as recycling, reprocessing, and disposal to industrial manufacturing will provide significant economic and environmental benefits to industry. However, current technologies, limited by technological constraints, inevitably subject closed-loop production processes to uncertainties such as productivity fluctuations caused by long-term equipment operation, drastic external changes, inflation, and shifts in national policies. These uncertainties directly impact the profitability of production scheduling decisions. Therefore, researching closed-loop manufacturing scheduling optimization methods under uncertainty can, on the one hand, reduce the negative impact of internal uncertainties on production costs, and on the other hand, allow decision-makers to flexibly control the weighting of costs and production risks based on external policies and economic conditions. This can maximize the utilization of key measures by enterprises and enhance their competitiveness.

[0003] Current closed-loop manufacturing scheduling optimization management models often require large amounts of data to ensure the reliability of the calculation results, and do not adequately consider robustness. Insufficient robustness can adversely affect actual production. Therefore, it is particularly necessary to provide a method that can effectively optimize closed-loop manufacturing production decisions. Summary of the Invention

[0004] To overcome the shortcomings of existing closed-loop management production systems, as described in the background due to technological limitations, this invention provides a closed-loop manufacturing system for the manufacturing industry that considers manufacturing and remanufacturing under uncertainty. This system integrates factors such as manufacturing, recycling, reprocessing, and scrapping involved in production. Through the combined action of relevant steps and calculation methods, it can offset the impact of uncertainty on the model's optimal solution. Decision-makers can control the conservatism of the solution by adjusting the confidence probability to cope with fluctuations in the external environment. This enables enterprises to better cope with the uncertainties brought about by internal losses and external economic fluctuations, maximizing the quality of decision-making methods. It reduces the need for distribution information of uncertainty parameters in the solution, while improving solution speed and optimization quality. This provides favorable technical support for the concept of "sustainable development" and ecological environmental protection. This is a closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints.

[0005] The solution adopted by this invention to solve its technical problem is:

[0006] The closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints is characterized by obtaining probabilistic information of uncertainty parameters by analyzing the probability distribution and perturbation range of production data. Decision-makers can set the risk level of the model according to external policies and the economic environment, thereby obtaining a robust closed-loop manufacturing scheduling scheme. Specifically, it includes three steps: collecting data from the closed-loop manufacturing process, establishing an uncertain closed-loop manufacturing model, and establishing a robust reconfiguration opportunity constraint optimization model. The data collected from the closed-loop manufacturing process includes: supplier procurement data for components, data on the transportation of parts and modules from the warehouse to the manufacturing unit, and data on the transportation of semi-finished products. The manufacturing unit data, semi-finished products and modules, and finished products are inspected by the inspection unit in the warehouse; the uncertain closed-loop manufacturing model includes two sub-steps: calculating the objective function and calculating the constraints; the establishment of a robust reconstructive opportunity-constrained optimization model includes the following sub-steps: A: improving the robust optimization model with parameter uncertainty; B: describing the uncertainty parameters in the model; C: representing the constraints in the original model equivalently; D: using the norm uncertainty set to define the uncertainty parameters; E: reconstructing the traditional opportunity-constrained optimization model; F: calculating the probability of violation of the original constraints; G: solving for the confidence probability of each independent opportunity constraint.

[0007] Furthermore, the data on the procurement of components from suppliers also includes data on the production of components into different types of modules and the transportation of modules to the module warehouse; the data on the transportation of parts and modules from the warehouse to the manufacturing unit includes data on the assembly of parts and modules into different types of semi-finished products, and data on qualified semi-finished products being delivered to the semi-finished product warehouse.

[0008] Furthermore, the semi-finished products are transported to the manufacturing unit data, and the final products are assembled from the semi-finished products and the qualified final products are delivered to the final product warehouse data.

[0009] Furthermore, the finished products and modules are inspected by the detection unit in the warehouse, and the production data of defective products are also collected by the collection unit.

[0010] Furthermore, the formula used in calculating the objective function is as follows: In the formula, PC represents raw material procurement cost, MC represents manufacturing cost, RC represents remanufacturing manufacturing cost, and SC represents inventory cost. Indicates transportation costs, This represents the objective function.

[0011] Furthermore, the specific process of the calculation constraints is as follows: calculating that the material procurement quantity in any period cannot exceed the supplier's supply quantity range, calculating production capacity restrictions, calculating material conservation, calculating warehouse inventory within the period, calculating the data on the complete separation of non-conforming finished and semi-finished products in the remanufacturing unit, and calculating the uncertainty ratio between the transportation volume of non-conforming materials from the remanufacturing unit to the purification unit and the transportation volume from the remanufacturing unit to the warehouse; all decision variables in the above calculation data are non-negative variables.

[0012] Furthermore, the establishment of the robust reconfiguration opportunity-constrained optimization model mainly involves using a robustly optimized flexible perturbation interval to describe the uncertainty parameters in the uncertain closed-loop manufacturing model, and adding adjustment variables so that decision-makers can adjust the confidence level of the model according to the external environment, in order to balance the risk level and economic cost of the scheduling scheme.

[0013] Furthermore, in step A, the main focus is on adapting to the assembly model; in step D, the main focus is on better describing the fluctuation range of uncertainty parameters in the model; in step E, the main focus is on reconstructing the joint opportunity constraints into multiple independent opportunity constraints to solve the joint opportunity constraint optimization problem; in step G, for constraints in the model that require strong resistance to disturbances, robust optimization is used for solving, while for constraints with low resistance to disturbances, joint opportunity constraints are used for optimization.

[0014] The beneficial effects of this invention are as follows: This invention provides the manufacturing industry with a closed-loop manufacturing system that considers manufacturing and remanufacturing under uncertainty. It addresses the production decision optimization problem through a robust reconstruction-based joint chance constraint algorithm, which can mitigate the impact of uncertainty on the model's optimal solution. Furthermore, it allows decision-makers to control the conservatism of the solution by adjusting the confidence probability to cope with fluctuations in the external environment, thus contributing to resource-saving solutions and reducing production waste. Through this invention, manufacturing enterprises can better cope with the uncertainties brought about by internal losses and external economic fluctuations, maximizing the quality of decision-making methods. Compared with other existing chance constraint optimization methods, this invention uses a robust reconstruction method to solve joint chance constraints, significantly reducing the need for distribution information of uncertainty parameters while improving solution speed and optimization quality, and enabling the optimization of production fluctuations in the manufacturing and remanufacturing processes. Based on the above, this invention has good application prospects. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the closed-loop manufacturing process of existing electronic assembly as disclosed in this invention.

[0016] Figure 2 This is a schematic diagram illustrating the implementation process of the joint chance constraint algorithm based on robust reconstruction of this invention.

[0017] Figure 3 This is a schematic diagram of the probability distribution table of the uncertainty parameters of this invention.

[0018] Figure 4 This is a schematic diagram illustrating the impact of the uncertainty of the price and defect rate of this invention on the objective function value.

[0019] Figure 5 This is a schematic diagram of the cost sub-target value table of the present invention.

[0020] Figure 6 This is a schematic diagram illustrating the relationship between the target value and the constraint confidence probability of this invention.

[0021] Figure 7 This is a schematic diagram of the uncertainty parameter perturbation interval table under different confidence levels of the present invention.

[0022] Figure 8 This is a schematic diagram of the slope of the target value of the present invention.

[0023] Figure 9 This is a schematic diagram of the normalized slope of the objective function of this invention. Detailed Implementation

[0024] Figure 2 As shown, the closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints obtains the probability information of uncertainty parameters by analyzing the probability distribution and perturbation range of production data. Decision-makers can set the risk level of the model according to external policies and economic environment, thereby obtaining a robust closed-loop manufacturing scheduling scheme. Specifically, it includes three steps: collecting data of the closed-loop manufacturing process, establishing an uncertain closed-loop manufacturing model, and establishing a robust reconfiguration opportunity constraint optimization model.

[0025] The process of collecting data from the closed-loop manufacturing process in this invention is as follows: by collecting the production and scrap numbers and price fluctuations, data support is provided for the distribution of uncertainty parameters in the model. Figure 1This invention presents a closed-loop manufacturing process. Based on a real-world case study of a closed-loop assembly and recycling network in an electronics assembly plant, it remodels and optimizes the uncertainties involved in the manufacturing process. The specific closed-loop manufacturing system consists of 7 production units (departments), 4 warehouses, and 4 recycling units (departments). Raw materials used in production fall into two categories: components and parts (hereinafter abbreviated as CP and PT). Intermediate semi-finished products fall into two categories: modules and semi-products (hereinafter abbreviated as MD and SP). Finished products fall into one category: final-products (hereinafter abbreviated as FP). There are 9 models of CP, 3 models of PT, and 3 models each of MD, SP, and FP. Raw materials are sourced from three suppliers, each with varying procurement costs. The design considers a seven-cycle production requirement. The specific manufacturing process is as follows: Figure 1 As shown, the description is as follows: a. First, the manufacturer purchases components from suppliers; in the first stage, these raw materials (components) are transported to the manufacturer's manufacturing unit AC, where the components are manufactured into different types of modules, and then the modules are transported to the module warehouse. b. The parts and modules are transported from the warehouse to manufacturing unit DE; in the second stage, according to a specific ratio and a specific route, the parts and modules can be assembled into different types of semi-finished products, and then the qualified semi-finished products are sent to the semi-finished product warehouse. c. In the third stage, the semi-finished products are transported to manufacturing unit FG, where the final product is assembled from the semi-finished products; the qualified final product is delivered to the final product warehouse. d. Semi-finished products, modules, and finished products are inspected by the inspection unit in the warehouse. Defective products are sent to the collection unit for recycling and production. Since the product defect rate is an uncertain factor that affects the amount of materials entering the reverse process, the collection unit will store and classify all defective products in the reverse process. Then, these defective products need to be sent to the remanufacturing unit for dismantling. The subsequent purification unit receives the defective products that need to be scrapped from the remanufacturing unit and purifies these materials. The usable materials after processing by the remanufacturing unit will be sent to the corresponding warehouse for reuse. Defective products (CP, PT, MD) in the purification unit will be completely scrapped in the disposal unit.

[0026] Figure 2 As shown, an uncertainty closed-loop manufacturing model is established. Based on the data obtained from the above closed-loop manufacturing process, this invention patent constructs the following uncertainty closed-loop manufacturing model to hedge against the uncertainty of production qualification rate and price fluctuations during the production process. Simultaneously, it enables decision-makers to adjust the model's risk level in a timely manner according to the external environment to achieve better economic benefits. The specific steps are as follows: a. Calculate the objective function, as shown in the formula... As shown, the formula includes raw material procurement cost (PC), manufacturing cost (MC), remanufacturing cost (RC), inventory cost (SC), and transmission cost. The main purpose of this step is to find the minimum production cost of the closed-loop assembly network in the uncertain closed-loop manufacturing model. b. Calculate the constraints. The specific constraints of the closed-loop assembly network model are shown in equations (2)-(8), including the following sub-steps. 1): Calculate that the material purchase quantity in any period cannot exceed the supplier's supply quantity range, as shown in equation rs. min ≤qrm≤rs max As shown in (2); where rs min and rs max These represent the minimum and maximum values ​​of materials supplied by the supplier in each cycle, respectively. qrm is the purchase quantity per cycle. The purchase quantity of materials is limited by (2). 2): By formula MQ≤mu max (3), formula MY≤ru max (4) Calculate the production capacity limit. The output of a specific manufacturing unit cannot exceed its capacity limit at any time, as shown in equation (3), where MQ is the manufacturing quantity of the manufacturing unit in a unit period, and mu max The manufacturing unit's production capacity is capped within a unit cycle; the recycling unit's processing capacity must not exceed its processing limit at any time, as shown in equation (4), where MY is the recycling unit's processing capacity within a unit cycle, and ru max To limit the processing capacity of the recycling unit within a unit cycle, the production capacity of the manufacturing and recycling units is restricted through (3) and (4). 3): Calculate material conservation, that is, calculate that the amount of material entering each unit is equal to the amount of material output, specifically through the formula Perform the calculation, where, This represents the sum of the quantities of materials transported from other units j to j'. This represents the sum of materials transported from unit j to other units j, ensuring material inflow and outflow are conserved in each unit. 4): Calculate the warehouse inventory at period t, specifically using the formula IL. t =IL t-1 +QT j,j',t -QT j',j,t Calculate t∈T,j∈J (6); where IL t and IL t-1 QT represents the inventory of materials in the warehouse during periods t and t-1, respectively. j,j',t QT represents the quantity of material transported from other units j to unit j' within cycle t. j',j,tThis represents the quantity of material transported from unit j' to other units j within cycle t, where T represents the set of production cycles. 5): After the remanufacturing unit completely separates defective finished and semi-finished products, qualified parts will be returned to the corresponding warehouse, and completely unusable materials will be transported to the purification unit for purification. The data is specifically calculated using the formula... Calculate; where MY t QT represents the quantity of material i processed by the remanufacturing unit within cycle t. j,j',t This represents the quantity of material transported from remanufacturing unit j to other units within cycle t. 6) Represents the quantity of material i' obtained from disassembling each unit component i, and T represents the set of production cycles. Calculate the uncertainty ratio between the amount of non-conforming materials transported from the remanufacturing unit to the cleanroom unit and the amount transported from the remanufacturing unit to the warehouse, specifically using the formula... Perform calculations; Let QT be the yield rate of the material processed by the remanufacturing unit within cycle t, and let QT be an uncertainty parameter. j,j',t QT represents the quantity of materials transported from other units to unit j' within period t. j',j,t Let be the quantity of materials transported from unit j' to other units within period t. This formula restricts the proportional relationship of this unit. 7): All decision variables are non-negative variables, and the above constraints constitute a closed-loop manufacturing model.

[0027] Figure 2 As shown, a robust reconfiguration opportunity constraint optimization model is established. In the existing technology, a large amount of data is often required to ensure the reliability of the calculation results. At the same time, it does not consider robustness and is not robust enough, which will have an adverse effect on practical applications. In order to combat the uncertainty parameters in the closed-loop production process, this invention uses a robust optimization flexible perturbation interval to describe the uncertainty parameters in the model, and adds adjustment variables so that decision-makers can adjust the confidence level of the model according to the external environment to balance the risk level and economic cost of the scheduling scheme. The specific process is as follows. Step 1: Improve the robust optimization model with parameter uncertainty (such as the uncertainty pass rate in formula (8)). The specific formula used is as follows:

[0028]

[0029] In the formula, It is a left uncertainty coefficient matrix. It is a right-hand uncertainty coefficient matrix. It is the coefficient matrix in the objective function. R is the coefficient matrix of the equality constraints. m×n This is an m x n coefficient matrix. Step 2: Describe the uncertainty parameters in the model, using the following formula:

[0030] In the formula, a ij b i c i d ij The nominal portion (i.e., the determined value) of the corresponding uncertain parameter. It is the amplitude of the positive disturbance, ξ ij μ i σ ij This is a unit variable used to control the range of parameter fluctuations in a robust optimization model; the calculation of the control variable is shown in the following formula:

[0031]

[0032] In the formula,

[0033] χ i The confidence probability is set by the decision-maker. for quantile function Let j be the uncertainty parameter in the i-th constraint. This step is used to obtain the value of the control variable. Step 3: Represent the constraints in the original model in an equivalent way, as shown in the following formula:

[0034]

[0035] x is the decision variable, and the other parameters and uncertainty parameters are the same as those in the previous text.

[0036] Step 4: Define the uncertainty parameters using a norm-based uncertainty set to better describe the fluctuation range of the uncertainty parameters in the model; to simplify calculations, a box-shaped uncertainty set is used to represent the uncertainty parameters, as shown in the following formula.

[0037]

[0038] From the above set of uncertainties, the robust corresponding optimization model constraints are obtained:

[0039]

[0040] Ψ represents the perturbation interval of the uncertainty parameter; other parameters and decision variables are described above. Steps one, two, three, and four use data obtained from a traditional robust correspondence optimization model. Based on this, and considering the adjustability between risk and return, this invention introduces a robust approximation algorithm with opportunity constraints to address the uncertainty of material cost parameters and scrap rate in the model. Simultaneously, robust optimization is used to handle other uncertain parameters in the model to improve its feasibility and robustness, as detailed in the following steps.

[0041] Figure 2As shown, a robust reconstructed opportunity-constrained optimization model is established. Step five: The traditional opportunity-constrained optimization model in equation (15) is reconstructed into the model in equation (16):

[0042]

[0043]

[0044] This invention transforms the original robust optimization model for multi-source uncertain parameters into a hybrid model combining joint chance constraint and robust optimization. Specifically, as seen in equation (16), the original model is divided into two parts: joint chance constraint optimization and robust optimization, where, This is the lower bound of the prior probability of Δ, where Δ is the perturbation interval. The calculation method will be given in subsequent steps. Through this step, the present invention reconstructs the joint opportunity constraint into multiple independent opportunity constraints to solve the joint opportunity constraint optimization problem. Step Six: Assuming that the uncertain parameters in equation (16) are mutually independent and follow a symmetric bounded probability distribution, for any case, the probability of violating the original constraint is expressed as follows:

[0045]

[0046] θ is an intermediate variable. The corresponding explicit probability is:

[0047]

[0048] This step allows us to obtain the relationship between the perturbation interval and the constraint violation probability, which can be used for subsequent model solving.

[0049] Step 7: Assuming each uncertain parameter is independent, use the Bernoulli test model to solve for the confidence probability of each independent chance constraint using the lower bound of the population probability, as shown in the following formula:

[0050]

[0051] Therefore, the robust approximation model with chance constraints regarding the uncertainty of raw material cost parameters and the uncertainty of the yield rate is expressed as follows:

[0052]

[0053] In the formula, P price P wasterepresents the lower bound of the confidence probability for the price uncertainty parameter and the pass rate uncertainty parameter, respectively. These can be manually set based on market fluctuations; higher values ​​indicate a more conservative target value and increased reliability of the model's calculation results. z is an intermediate variable that transfers the uncertainty in the objective function to the constraints for solution. Other parameters are defined in the formulas above. For constraints requiring strong resilience to disturbances, robust optimization is used; for constraints with lower resilience requirements, joint chance constraints are used for optimization.

[0054] Through all the steps described above, this invention obtains probabilistic information about uncertainty parameters by analyzing the probability distribution and perturbation range of production data. Decision-makers can adjust the risk level of the model based on external policies and the economic environment, thereby obtaining a robust closed-loop manufacturing scheduling scheme. Specifically, firstly, a closed-loop manufacturing model is established, featuring a forward manufacturing process and a reverse recycling process, described by an objective function and constraints. Secondly, addressing the multi-source uncertainties in closed-loop manufacturing, the mixed-integer linear programming based on the joint chance constraint method is transformed into an independent chance constraint method using robust reconstruction, and the perturbation interval is embedded into the chance constraint programming to model the uncertainty parameters. In this way, decision-makers can control the conservatism of the solution by adjusting the joint confidence level and reduce the amount of parameter information required for optimization under uncertainty by designing robust reconstruction. From the perspective of environmental pollution and operating costs, the optimal confidence probability value is obtained. This invention can optimize production fluctuations in manufacturing and remanufacturing processes. Furthermore, decision-makers can adjust the risk level of the model based on external policies and the economic situation, which helps to provide resource-saving solutions and reduce production waste. Compared with other opportunity-constrained optimization methods, this invention uses a robust reconstruction method to solve joint opportunity constraints, which greatly reduces the need for distribution information of uncertainty parameters and improves the solution speed and optimization quality.

[0055] The following content uses the technology of this invention for simulation, and the simulation data is used to verify the opportunity-constrained optimization closed-loop manufacturing model based on robust reconstruction of this invention. Uncertainty parameters arising from market price fluctuations and equipment wear during the manufacturing process follow a normal distribution within a predetermined interval, such as... Figure 3 As shown. Since the manufacturing cost target value has different sensitivities to the uncertainty parameters of the two dimensions (i.e., price and manufacturing yield rate), the optimal confidence level selection scheme is sought by analyzing the impact of the confidence levels of the two uncertainty parameters on the target value. Simultaneously, the performance of the improved joint chance constraint robust reconstruction algorithm is evaluated through the objective function value and robustness. Considering the feasibility, production safety, and reliability of the model involved in the production process, the lower bound of the two-dimensional confidence probability fluctuation between [0.8, 0.997] is calculated to observe the impact of the parameters on the target value. The results are as follows. Figure 4As shown in Table 5, the results for each sub-target value are as follows.

[0056] Figure 4 As can be seen, production costs gradually increase with increasing confidence probability. The gray plane in the figure represents the objective value calculated through flexible robust optimization. Compared to ordinary robust optimization, this algorithm has a faster computation speed, allowing decision-makers to more intuitively observe the impact of uncertain parameters on economic benefits under multiple uncertainties. In most cases, the optimal objective value of flexible robust optimization is higher than that of joint chance-constrained optimization. Figure 4 It can be seen that, within the same confidence probability range (0.8–0.95), the greater the range of cost variation with the confidence probability of the non-conforming rate, the higher the sensitivity. Further observation of the impact of confidence probability on cost reveals the influence curves of the confidence levels of the two uncertain parameters on the objective function, as shown in the figure. Figure 6 As shown; Figure 6 There are four different curves, representing the impact of changes in the confidence level of the non-conformance rate uncertainty on the target value when the confidence level of the price uncertainty parameter is 0.8 and 0.997 (Pp = 0.8 and Pp = 0.997), and the impact of the confidence level of the price uncertainty parameter on the target value when the confidence level of the non-conformance rate uncertainty parameter is 0.8 and 0.997 (Pr = 0.8 and Pr = 0.997).

[0057] The robustness of the model designed in this invention is related to the chosen confidence probabilities. To observe the impact of confidence probabilities on model robustness, Pp and Pr (i.e., the confidence probabilities corresponding to the uncertainty parameters of price and pass rate) with different confidence probabilities were selected, thereby obtaining the perturbation intervals of the corresponding uncertainty parameters, such as... Figure 7 As shown. For example, the coverage of the perturbation interval increased by 20%, while the confidence probability increased from 0.925 to 0.975; therefore, the higher the confidence probability and the larger the perturbation interval, the lower the model's sensitivity to parameter changes, and the more conservative the model. To analyze the model's sensitivity, calculations were performed separately. Figure 6 The slope of each curve in the figure, the result is as follows Figure 8 As shown. Figure 8 It is evident that the slope values ​​are not easily observable. Therefore, the four curves are normalized to obtain the slope curves showing the change of the target value with the confidence levels of the two uncertainty parameters, as shown below. Figure 9As shown, points with a slope of 1 are selected on both curves; point A is defined on the price uncertainty curve, and point B is defined on the defect rate uncertainty curve. The confidence level of price uncertainty at point A reaches 92.374%, and the confidence level of defect rate uncertainty at point B reaches 93.364%. It can be seen that the slope value changes slowly until the confidence probability changes to the probability value corresponding to point A, at which point it begins to increase rapidly. Therefore, the sensitivity of the objective function value before point A is lower than the sensitivity of the objective function value after point A. Figure 9 Point B yields the same observation; therefore, the confidence probabilities corresponding to points A and B are chosen, resulting in an objective value of 58,939,721.5913925 (CNY). Compared to existing stochastic programming techniques, this scheme reduces costs by 3.17%, and allows for the artificial reduction of confidence probabilities, achieving even lower manufacturing costs. The decision scheme not only reduces production costs but also mitigates environmental impact through recycling, reducing pollution and improving resource utilization, thus achieving sustainable development. By robustly approximating reconstruction constraints and embedding opportunity-constrained optimization, decision-makers can adjust expected probabilities to balance costs and uncertainties, enabling the model to hedge against the impact of external policies and economic conditions. This invention employs a robust reconstruction-based opportunity-constrained optimization method in closed-loop manufacturing, providing a flexible and effective scheduling scheme for discrete manufacturing under uncertain conditions, bringing higher profits to enterprises and less pollution to the environment.

[0058] Those skilled in the art should understand that although this specification describes embodiments, the embodiments do not necessarily contain only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in the embodiments can also be appropriately combined to form other embodiments that can be understood by those skilled in the art. Therefore, the scope of protection of this application is defined by the claims.

Claims

1. A closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints, characterized in that, By analyzing the probability distribution and perturbation range of production data, probabilistic information of uncertainty parameters is obtained. Decision-makers can set the risk level of the model based on external policies and the economic environment, thereby obtaining a robust closed-loop manufacturing scheduling scheme. Specifically, this includes three steps: collecting data from the closed-loop manufacturing process, establishing an uncertain closed-loop manufacturing model, and establishing a robust reconfiguration opportunity-constrained optimization model. The data collection for the closed-loop manufacturing process includes: supplier procurement data, data on parts and modules being transported from the warehouse to the manufacturing unit, data on semi-finished products being transported to the manufacturing unit, and data on semi-finished products, modules, and finished products being inspected by the inspection unit in the warehouse. The uncertain closed-loop manufacturing model includes two sub-steps: calculating the objective function and calculating the constraints. The establishment of the robust reconfiguration opportunity-constrained optimization model includes the following sub-steps: A: Improving the robust optimization model with parameter uncertainty; B: Describing the uncertainty parameters in the model; C: Equivalently representing the constraints in the original model; D: Using a norm uncertainty set to define the uncertainty parameters; E: Reconstructing the traditional opportunity-constrained optimization model; F: Calculate the probability of violation of the original constraints, G: solve for the confidence probability of each independent opportunity constraint; establish a robust reconstruction opportunity constraint optimization model, which describes the uncertainty parameters in the uncertain closed-loop manufacturing model using a robust optimization flexible perturbation interval, and adds adjustment variables so that decision-makers can adjust the confidence level of the model according to the external environment to balance the risk level and economic cost of the scheduling scheme; in step A, adapt the assembly model; in step D, better describe the fluctuation range of the uncertainty parameters in the model; in step E, reconstruct the joint opportunity constraint into multiple independent opportunity constraints to solve the joint opportunity constraint optimization problem; in step G, for constraints in the model that require strong resistance to disturbances, robust optimization is used to solve them, and for constraints with low resistance to disturbances, joint opportunity constraints are used for optimization.

2. The closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints according to claim 1, characterized in that, The data includes data on the procurement of components from suppliers, data on the production of components into different types of modules, data on the transportation of modules to the module warehouse, and data on the transportation of parts and modules from the warehouse to the manufacturing unit, including data on the assembly of parts and modules into different types of semi-finished products, and data on qualified semi-finished products being delivered to the semi-finished product warehouse.

3. The closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints according to claim 1, characterized in that, Semi-finished products are transported to the manufacturing unit data, and the final product is assembled from the semi-finished products. Qualified final products are delivered to the final product warehouse data.

4. The closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints according to claim 1, characterized in that, Finished products and modules are inspected by the testing unit in the warehouse, and the data also includes data on non-conforming products sent to the collection unit for production data recovery.

5. The closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints according to claim 1, characterized in that, The formula used in calculating the objective function is as follows: In the formula, Indicates raw material procurement costs, Indicates manufacturing cost, Indicates the manufacturing cost of remanufacturing, Indicates inventory cost, Indicates transportation costs, This represents the objective function.

6. The closed-loop manufacturing production decision optimization method based on robust reconfiguration joint opportunity constraints according to claim 1, characterized in that, The specific process for calculating the constraints is as follows: calculating that the quantity of materials purchased in any period cannot exceed the supplier's supply quantity range; calculating production capacity limitations; calculating material conservation; calculating warehouse inventory during the period; calculating data on the complete separation of non-conforming finished and semi-finished products in the remanufacturing unit; and calculating the uncertainty ratio between the amount of non-conforming materials transported from the remanufacturing unit to the purification unit and the amount transported from the remanufacturing unit to the warehouse. All decision variables in the above calculations are non-negative variables.

Images