An interactive approach to two-response product and process design optimization with statistical inferences

Thesis Type: Postgraduate

Institution Of The Thesis: Orta Doğu Teknik Üniversitesi, Faculty of Engineering, Department of Industrial Engineering, Turkey

Approval Date: 2015




In this study, an interactive approach has been developed for two-response product and process design optimization problems treating the single response problems as a special case. This approach considers preferences of the decision maker explicitly and the correlation between the responses. It uses a predefined set of objectives that are commonly encountered in the literature and industrial applications. However, instead of presenting all objective values at each iteration, a set of performance measures are used to represent the objectives in a way to communicate with the decision maker better. A significant part of this communication utilizes visual aids such as specification and prediction regions of a solution. Thus, the decision maker is able to decide better which objective can be sacrificed by how much in order to improve an unsatisfactory objective in the next iteration. The developed approach is demonstrated on two examples; one exemplifies binary response problem and the other is on single response problem. The advantages and disadvantages of the developed approach and issues for covering more than two responses are discussed.