Multi-stage production process decision and cost optimization method based on sampling detection
Through multi-level production process decision-making and cost optimization algorithms based on sampling inspection, the problems of insufficient production efficiency and unsatisfactory cost control in industrial assembly production are solved, and efficient and low-cost production in complex environments is achieved.
Patent Information
- Application Number
- CN202510830068.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-20
- Publication Date
- 2025-10-10
AI Technical Summary
Modern industrial assembly production faces problems such as insufficient production efficiency, unsatisfactory cost control, and difficulty in controlling defective product rates. Traditional sampling and inspection methods lack flexibility and dynamism, and are difficult to adapt to complex and changing production environments.
A multi-stage production process decision-making and cost optimization algorithm based on sampling inspection is adopted, including a product inspection requirement module, a two-stage inspection model, a cost minimization dynamic model and a genetic algorithm module. The multi-stage production process is optimized by dynamically adjusting the sample size and decision-making strategy in combination with the genetic algorithm.
It achieves dynamic adjustment of sample size and decision-making strategy while ensuring product quality, improves the flexibility and efficiency of the production process, reduces production costs, and adapts to complex and changing production environments.
Smart Images
Figure CN120764745A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of industrial assembly production, and in particular to a multi-level production process decision and cost optimization algorithm based on sampling detection. Background Art
[0002] Modern industrial assembly production often faces diverse and complex decision-making challenges, such as insufficient production efficiency and suboptimal cost control. These issues directly impact a company's market competitiveness and production efficiency. With intensifying market competition and consumers' growing concern for product quality, companies are placing increasingly stringent demands on quality control during production. Traditional sampling inspection methods with a fixed defective rate are no longer able to meet companies' demands for high-quality, low-cost production. Since production processes often involve multiple stages and components, minimizing production costs while ensuring product quality has become a pressing issue for companies. While some algorithms exist domestically for production process optimization, they still struggle with sampling inspection, cost control, and defective rate control.
[0003] Current traditional sampling inspection methods often make decisions based on a fixed defective rate, lacking flexibility and dynamism, making them difficult to adapt to complex and changing production environments. Recent advances in genetic algorithms have helped us find cost-optimal solutions over multiple generations. This paper proposes a multi-stage production process decision-making and cost optimization method based on sampling inspection. This method effectively controls quality by employing scientific sampling methods at different production stages. Summary of the Invention (1) Technical problems solved
[0004] In response to multiple problems in the industrial assembly production process, such as insufficient production efficiency, unsatisfactory cost control, and difficulty in controlling the defective rate, the purpose of the present invention is to provide a multi-level production process decision and cost optimization algorithm based on sampling detection, which is used to solve the problems of insufficient production efficiency and unsatisfactory cost control in the industrial assembly production process. (2) Technical solution
[0005] In order to achieve the above object, the present invention is implemented through the following technical solutions:
[0006] The present invention provides a multi-stage production process decision-making and cost optimization method based on sampling inspection, including: a product inspection requirement module, clarifying the main decision-making links, a two-stage product inspection model, a preliminary screening and detailed inspection module, a cost minimization dynamic model, a genetic algorithm module, a multi-stage production optimization model, a defective rate optimization module, a dynamic optimization model based on sampling inspection, and a correction and optimization module.
[0007] A multi-stage production process decision-making and cost optimization method based on sampling inspection includes the following steps:
[0008] S1. Clarify product inspection requirements during the production stage
[0009] A finished electronic product is assembled from two parts. If either part fails, the finished product will definitely fail. Even if both parts pass, the finished product may still fail. For failed products, the company can choose to scrap them directly or pay for disassembly to recover intact parts.
[0010] S2. Two-stage test model
[0011] A two-stage testing model was constructed: the first stage is preliminary screening, in which we conduct small sample tests. If no clear conclusion can be drawn, we will enter the second stage, the detailed testing stage, and further sampling tests will be conducted to meet the confidence requirements.
[0012] Step (1): Statistical hypothesis testing, applying hypothesis testing to evaluate whether the defective rate of a certain spare part exceeds the nominal value. In order to conduct hypothesis testing, we can use a standardized process to convert the defective rate into a standard normal distribution value z-score for inference testing.
[0013] Step (2): We estimate the minimum number of tests using the sample size formula Based on the number of samples and the characteristics of the population, the minimum sample size can be calculated using the normal distribution approximation properties of the binomial distribution. By consulting the standard normal distribution table, we can obtain the corresponding critical value information at each confidence level.
[0014] Step (3): Design a cost-minimizing sampling test plan. Perform small batch sampling tests and analyze the samples. If no definite conclusion can be drawn, then proceed to the next sampling test stage until the predetermined reliability requirement is met. k~Binomial(n,p)
[0015] S3. Cost minimization dynamic model based on genetic algorithm To minimize total cost, decisions need to be made at different stages. We group these decisions into four categories: parts inspection and review decisions, finished product qualification inspection decisions, disassembly and reuse decisions, and return processing decisions.
[0016] The genetic algorithm is used to solve the problem, and the process is as follows:
[0017] Step (1): We assume that the defective rate threshold is 1.5%, and the ratio of the defective rate reduction of Part 1, Part 2 and the product is 0.65.
[0018] Step (2): During the loop, we comprehensively consider the inspection cost and defective rate. The decision analysis function intelligently decides whether to conduct inspection or disassembly, while simultaneously calculating and accumulating the total cost of each loop in real time. As the loop progresses, the defective rate decreases by a predetermined ratio until the set termination condition is reached, thus completing the optimization analysis of the entire decision-making process.
[0019] Step (3): We use the DEAP library to define the fitness function to calculate the total cost of the individual. We then create individuals and populations. The individual consists of three parameters (defective rate of component 1, defective rate of component 2, and defective rate of finished product). We then register the genetic algorithm operations, including selection, crossover, and mutation.
[0020] Step (4): Initialize the population and run the genetic algorithm to find the optimal solution through multiple generations of evolution. In each generation, the fitness of the individuals is evaluated and individuals with higher fitness are selected for crossover and mutation.
[0021] Step (5): After the algorithm ends, the individual with the best fitness is selected from the population as the final solution. This means the optimal solution with the minimum cost is obtained.
[0022] S4. Multi-stage production optimization model
[0023] Step (1): During each cycle, the cost formula for testing spare parts is:
[0024] Step (2): During each cycle, the cost formula for testing semi-finished products is:
[0025] Step (3): During each cycle, the cost formula for finished product testing is:
[0026] Step (4): The total cost function is the accumulation of the inspection, assembly, and disassembly costs of each cycle, representing the total cost of each operation in the entire production process. The formula is:
[0027] The system stops iterating if the defective rate drops to a predetermined threshold or the number of iterations reaches an upper limit, N. The system terminates promptly upon reaching the maximum number of iterations, N, preventing endless loops. Furthermore, once the system reaches the desired defective rate or the upper limit, it outputs the optimal decision solution.
[0028] S5. Dynamic Optimization Model Based on Sampling Detection
[0029] The defective rate is estimated by sampling and testing spare parts, semi-finished products and finished products, and the overall defective rate is estimated based on the proportion of defective products in the sample. The defective rate is evaluated through sample testing, so the process of reducing the defective rate is more complicated. Specifically, during each cycle, the defective rate The value of the reduction depends on the defective rate estimate obtained in the previous sampling. The reduction formula is:
[0030] Since the defective rate is dynamically revised after each sampling, the defective rate involved in the inspection decision must be updated after each inspection cycle. Therefore, the inspection decision algorithm is set as follows:
[0031] Model optimization performed well in reducing overall costs, particularly at high confidence levels. However, the model's strategies for inspecting spare parts and semi-finished products still require further optimization to improve overall inspection accuracy and efficiency. Beneficial effects
[0032] (1) The present invention proposes a multi-stage production process decision-making and cost optimization method based on sampling inspection. This method considers various uncertainties in the production process, such as the dynamic changes in the defective rate and the errors in sampling inspection. By dynamically adjusting the sample size and decision-making strategy, the algorithm can maintain efficient and low-cost production process optimization even when the error in the defective rate estimation is large.
[0033] (2) Compared with the traditional fixed defective rate sampling inspection method, this model has higher flexibility and dynamism, can adapt to complex and changing production environments, and provide support for enterprises' decision-making in complex production environments. At the same time, the model also introduces intelligent optimization algorithms such as genetic algorithms. Through multiple iterations and fitness evaluation, it can find the optimal solution in a complex parameter space, achieving a balance between inspection cost and quality control. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 This is a schematic diagram of the operational flow of a multi-stage production process decision-making and cost optimization method based on sampling detection according to the present invention;
[0035] Figure 2 This is a detailed algorithm step diagram of a multi-stage production process decision-making and cost optimization method based on sampling detection in the present invention. DETAILED DESCRIPTION
[0036] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. A multi-stage production process decision-making and cost optimization method based on sampling inspection includes the following steps:
[0037] S1. Clarify product inspection requirements during the production stage
[0038] A finished electronic product is assembled from two parts. If either part fails, the finished product will definitely fail. Even if both parts pass, the finished product may still fail. For failed products, the company can choose to scrap them directly or pay for disassembly to recover intact parts.
[0039] S2. Two-stage test model
[0040] A two-stage testing model was constructed: the first stage is preliminary screening, in which we conduct small sample tests. If no clear conclusion can be drawn, we will enter the second stage, the detailed testing stage, and further sampling tests will be conducted to meet the confidence requirements.
[0041] Step (1): Statistical hypothesis testing, applying hypothesis testing to evaluate whether the defective rate of a certain spare part exceeds the nominal value. In order to conduct hypothesis testing, we can use a standardized process to convert the defective rate into a standard normal distribution value z-score for inference testing.
[0042] Step (2): We estimate the minimum number of tests using the sample size formula Based on the number of samples and the characteristics of the population, the minimum sample size can be calculated using the normal distribution approximation properties of the binomial distribution. By consulting the standard normal distribution table, we can obtain the corresponding critical value information at each confidence level.
[0043] Step (3): Design a cost-minimizing sampling test plan. Perform small batch sampling tests and analyze the samples. If no definite conclusion can be drawn, then proceed to the next sampling test stage until the predetermined reliability requirement is met. k~Binomial(n,p)
[0044] S3. Cost minimization dynamic model based on genetic algorithm To minimize total cost, decisions need to be made at different stages. We group these decisions into four categories: parts inspection and review decisions, finished product qualification inspection decisions, disassembly and reuse decisions, and return processing decisions.
[0045] The genetic algorithm is used to solve the problem, and the process is as follows:
[0046] Step (1): We assume that the defective rate threshold is 1.5%, and the ratio of the defective rate reduction of Part 1, Part 2 and the product is 0.65.
[0047] Step (2): During the loop, we comprehensively consider the inspection cost and defective rate. The decision analysis function intelligently decides whether to conduct inspection or disassembly, while simultaneously calculating and accumulating the total cost of each loop in real time. As the loop progresses, the defective rate decreases by a predetermined ratio until the set termination condition is reached, thus completing the optimization analysis of the entire decision-making process.
[0048] Step (3): We use the DEAP library to define the fitness function to calculate the total cost of the individual. We then create individuals and populations. The individual consists of three parameters (defective rate of component 1, defective rate of component 2, and defective rate of finished product). We then register the genetic algorithm operations, including selection, crossover, and mutation.
[0049] Step (4): Initialize the population and run the genetic algorithm to find the optimal solution through multiple generations of evolution. In each generation, the fitness of the individuals is evaluated and individuals with higher fitness are selected for crossover and mutation.
[0050] Step (5): After the algorithm ends, the individual with the best fitness is selected from the population as the final solution. This means the optimal solution with the minimum cost is obtained.
[0051] S4. Multi-stage production optimization model
[0052] Step (1): During each cycle, the cost formula for testing spare parts is:
[0053] Step (2): During each cycle, the cost formula for testing semi-finished products is:
[0054] Step (3): During each cycle, the cost formula for finished product testing is:
[0055] Step (4): The total cost function is the accumulation of the inspection, assembly, and disassembly costs of each cycle, representing the total cost of each operation in the entire production process. The formula is:
[0056] The system stops iterating if the defective rate drops to a predetermined threshold or the number of iterations reaches an upper limit, N. The system terminates promptly upon reaching the maximum number of iterations, N, preventing endless loops. Furthermore, once the system reaches the desired defective rate or the upper limit, it outputs the optimal decision solution.
[0057] S5. Dynamic Optimization Model Based on Sampling Detection
[0058] The defective rate is estimated by sampling and testing spare parts, semi-finished products and finished products, and the overall defective rate is estimated based on the proportion of defective products in the sample. The defective rate is evaluated through sample testing, so the process of reducing the defective rate is more complicated. Specifically, during each cycle, the defective rate The value of the reduction depends on the defective rate estimate obtained in the previous sampling. The reduction formula is:
[0059] Since the defective rate is dynamically revised after each sampling, the defective rate involved in the inspection decision must be updated after each inspection cycle. Therefore, the inspection decision algorithm is set as follows:
[0060] Model optimization performed well in reducing overall costs, particularly at high confidence levels. However, the model's strategies for inspecting spare parts and semi-finished products still require further optimization to improve overall inspection accuracy and efficiency.
Claims
1. A multi-stage production process decision and cost optimization problem based on sampling inspection, mainly including the following steps: S1. Clarify product inspection requirements during the production stage The finished product of an electronic product is assembled from two parts. If any one of the parts is unqualified, the finished product will definitely be unqualified; even if both parts are qualified, the finished product may still be unqualified. For unqualified products, the company can choose to scrap them directly or pay for disassembly to recycle intact parts. S2. Two-stage test model A two-stage inspection model was constructed: the first stage is preliminary screening, where we conduct small sample inspections. If no clear conclusions can be drawn, we will enter the second stage, the detailed inspection stage. To meet the confidence requirements, we will further conduct sampling inspections. We use hypothesis testing to evaluate the defective rate of spare parts, use the sample size formula to estimate the minimum number of inspections, and design a cost-minimizing sampling inspection plan. S3. Cost minimization dynamic model based on genetic algorithm Use genetic algorithms to optimize total costs, involving decisions such as parts inspection, finished product qualification inspection, disassembly and reuse, and return processing. Set the defective rate threshold and reduction ratio, define the fitness function through the DEAP library, create individuals and populations, register genetic algorithm operations, and find the optimal solution. S4. Multi-stage production optimization model Define the cost formula for testing spare parts, semi-finished products, and finished products, as well as the total cost function. Stop iteration when the defective rate drops to a predetermined threshold or the number of cycles reaches an upper limit, and output the optimal decision plan. S5. Dynamic Optimization Model Based on Sampling Detection The defective rate is estimated through sampling inspection. The defective rate decreases depending on the previous sampling result. The inspection decision algorithm updates the defective rate after each inspection cycle. Model optimization performs well in reducing total costs, but the inspection strategy for spare parts and semi-finished products needs further optimization.
2. The multi-stage production process decision and cost optimization problem based on sampling inspection according to claim 1, characterized in that: The cost minimization dynamic model of genetic algorithm reduces the defective rate and lowers the total cost. A dynamic model of multi-stage production optimization is used to reduce the number of defective products entering the assembly link to reduce market losses and disassembly costs. A multi-stage production decision and cost optimization model based on sampling inspection is constructed to solve the inspection and processing decision-making problems in different production stages.