Adaptive Decision-Making in Multi-Stage Production: A Unified Framework for Cost Optimization via Dynamic Sampling and Evolutionary Algorithms

In modern manufacturing, particularly within the electronics industry, balancing quality control and cost optimization across multi-stage production systems presents a significant challenge due to the cascading effects of component quality uncertainty. This complexity necessitates a systematic approach to decision-making that can dynamically adapt to quality fluctuations while minimizing overall production expenses. This paper addresses this challenge by proposing a unified framework to determine the optimal strategy for inspection, disassembly, and returns under sampling uncertainty. We developed a dynamic optimization model, solved using a Genetic Algorithm, which integrates a two-stage sampling inspection mechanism to evaluate trade-offs between inspection, disassembly, and return decisions across multiple production cycles. The results consistently demonstrate that forgoing intermediate inspections in favor of final product checks and strategic disassembly of defective units yields the lowest total cost, a strategy that remains robust even when accounting for the statistical uncertainty inherent in sampling. Ultimately, this research provides a scientific framework that equips manufacturers with an adaptive decision-making tool to enhance system resilience, optimize resource utilization, and achieve a cost-effective balance between quality and efficiency.