A Study on Multi-Stage Assembly Production Testing Decision Based on Binary-Coded Genetic Algorithm

Abstract

In the current manufacturing industry, intelligent optimization algorithms provide new ideas for quality control of multi-stage assembly production of complex products. However, the traditional inspection strategies and coding methods make it difficult to balance the inspection cost and product quality effectively. Aiming at this challenge, this paper proposes a multi-stage inspection decision model based on a binary coding genetic algorithm, which constructs a nonlinear profit maximization objective function by introducing 0-1 variables to describe the discrete decisions of spare parts inspection and finished product disassembly. The complex process of two processes and eight spare parts is downscaled to equivalent strategy combinations by the semi-finished product encapsulation method, and solved by combining violent search and genetic algorithm. The experimental results show that the model filters out the optimal solution among 65,536 strategies with a maximum profit of 52.94, which verifies the advantages of the algorithm in reducing the computational complexity (the number of strategies is reduced to 32×32×16) and improving the decision-making efficiency (the solution time is 1 second). The study provides an efficient solution to the cost-quality trade-off problem for multi-stage assembly production.