Robust Parameters Design of Categorical Responses under Modeling and Implementation Errors

International Journal of Decision Intelligence

مقاله 3، دوره 1، شماره 2، دی 2023، صفحه 25-44   

نوع مقاله: Original Article

نویسندگان

Milad Zamani¹؛ Arezoo Borji¹؛ Taha-Hossein Hejazi^{* 2}

¹Department of Industrial Engineering, College of Garmsar, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

²Department of Industrial Engineering, Amirkabir University of Technology (Tehran Polytechnic), Garmsar Campus, Iran

چکیده

Nowadays, improving quality is advocated as a strategy to increase market share, and failing to address this crucial issue results in exclusion from the competitive landscape. Most studies undertaken in recent years have investigated and optimized continuous response variables while ignoring categorical characteristics. This necessitates a change in statistical methods in this discipline to ones that take categorical responses into account. Statistical techniques have always provided researchers with estimates of parameters that are subject to uncertainty. Hence, considering uncertainty in modeling is essential for reducing errors and minimizing costs while increasing quality. In this study, we deal with the robust design of quality characteristics in categorical response problems to reach optimal levels of control variables, which can minimize the error caused by modeling and implementation and provide more accurate estimates. A portion of the uncertainty is considered while estimating the model parameters. However, the proposed approach assumes that the optimal settings of design variables during the implementation phase will also experience oscillations. Finally, in the optimization phase, multiple equations relating to response levels are modeled and solved using the goal programming approach. The results showed that our approaches could achieve solutions with robustness against the two main sources of errors.

کلیدواژه‌ها

Quality Engineering؛ Robust parameter design؛ dual response modeling؛ Logistic Regression؛ model imprecision؛ implementation error