In many engineering application the need arises to solve multi-objective optimiza-tion problems that involve computationally expensive functions. Our approach for multi-objective optimization is Normal-Boundary Intersection (NBI) by Das and Dennis. We can

run NBI either on surrogate models of the simulations or directly on the simulations. Nei-ther of these alternatives is effective on complex functions; the surrogates are not accurate enough, and using the simulations directly is too costly. We have found a unique approach that solves this dilemma. Our approach combines the use of surrogates and simulations in

更多

a way similar to the general surrogate management framework by iteratively using both

the surrogate models and the simulations.

收起