Multiple Objective Design Optimization by Estimating Local Utility Functions

This paper presents the application of an interactive multiple objective decision making (MODM) procedure to design synthesis. This procedure is composed of an interactive step trade-off method (ISTM) and the estimation of local utility functions. The ISTM method is used as a preference learning process to assist a designer (decision maker or DM) in generating a representative sub-set of efficient designs (solutions). In this learning process, the DM is expected to compare the generated solutions in a pairwise manner to express his preferences. A set of additive local utility functions are then estimated to help the DM search for his best compromise solution. A multiple objective design synthesis problem is examined using the procedure.