Decision Making and Constraint Tradeoff Visualization for Design Under Uncertainty
In design optimization problems under uncertainty, two conflicting issues are generally of interest to the designer: feasibility and optimality. In this research, we adopt the philosophy that design, especially under uncertainty, is a decision making process, where the associated tradeoffs can be conveniently understood using multiobjective optimization. The importance of constraint feasibility and the associated tradeoffs, especially in the presence of equality constraints, is examined in this paper. We propose a three-step decision making framework that facilitates effective decision making under uncertainty: (1) formulating a multiobjective problem that effectively models the tradeoffs under uncertainty, (2) generating design alternatives by solving the proposed multiobjective robust design formulation, and (3) choosing a final design using filtering and constraint uncertainty visualization schemes. The proposed framework can be used to systematically explore the design space from a constraint tradeoff perspective. A tolerance synthesis example is used to illustrate the proposed decision making process.