Interactive multi response surface optimization under uncertainty for robust parameter design


Özateş Gürbüz M., Köksal G., Köksalan M. M.

29th European Conference on Operational Research, Valencia, Spain, 8 - 11 July 2018, pp.294

  • Publication Type: Conference Paper / Summary Text
  • City: Valencia
  • Country: Spain
  • Page Numbers: pp.294
  • Middle East Technical University Affiliated: Yes

Abstract

We develop an interactive approach for two-response robust parameter design problem considering preferences of a decision maker (DM) under model uncertainty associated with the parameters of response surface models assuming that the model structure is appropriate. It is desired to minimize the joint variation in the responses and the distances of response means from their target values by deciding on the parameter settings of the product. To provide relevant information on the consequences of solutions to the DM, we produce visual aids on performance measures such as joint response confidence and prediction regions of solutions under consideration. We show that prediction error of the response surface models affects the quality of the solutions and should be considered in evaluating them. We involve the problem analyst (PA) in the search for reaching the preferred solutions of the DM. At each iteration, the PA interprets and converts DM’s verbal preferences into mathematical expressions and continues searching the solution space systematically to identify and present solutions that are in line with the DM’s expressed preferences. The procedure continues until the DM finds a satisfactory solution.