Economic-Oriented Robust Optimization Design Considering Model Parameter Uncertainty

The quality of a product is not only reflected in its manufacturing process, but its ultimate value depends on the level of quality experienced by the customer. This paper develops a cost model from a whole life cycle perspective, encompassing tolerance cost, rework cost, and scrap cost in pre-sale, as well as post-sale warranty costs. It also considers that manufacturers with different marketing strategies (e.g., high volume low margin or low volume high margin) place varying levels of emphasis on different loss costs. Furthermore, limitations of experimental data and unknown random effects can cause significant estimation errors in parameters during modeling, potentially leading to unreliable quality design. To address this issue, the study combines the concepts of interval estimation and robust optimization to minimize the impact of parameter estimation errors caused by uncertainty factors on the optimization process. From the dual perspectives of economic design and interval estimation, the output performance of products/processes is optimized for both optimality and robustness. Finally, the proposed method's effectiveness is validated through simulation experiments and industrial examples. The research results indicate that interval estimation effectively addresses the interference caused by uncertainty factors. By balancing the relationships among various costs, it achieves the minimization of total product cost.