Pareto Front Optimization for Multiple Process or Product Responses in the Presence of Model Parameter Uncertainty

• Activity Number: 373
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Quality and Productivity Section
• Author(s): Jessica Chapman*+ and Lu Lu and Christine Anderson-Cook
• Companies: St. Lawrence University and Los Alamos National Laboratory and Los Alamos National Laboratory
• Keywords: Optimizing Multiple Objectives ; Response Surface Methodology

Abstract

In many optimization situations, there are several responses associated with a product or process under consideration which need to be jointly considered. In this paper we present Pareto front multiple objective optimization as an option to complement other statistical and mathematical methods in the response surface methodology toolkit. Since the optimization is based on models with estimated parameters, any optimization procedure will be affected by uncertainty in those estimates. We investigate the effect that this uncertainty has on the Pareto front and the choice of overall solution. Using a simple case study with three responses and two explanatory variables, we present graphical tools that aid decision-makers in choosing an optimal solution using a Pareto front approach while accounting for response variability.