제한조건을 포함하는 최적화 문제에 대한 순차적 크리깅 메타모델 기반 최적설계 연구 트랜드

Abstract

Design problems have recently become complicate and nonlinear and the number of constraints has been increased according to the development of new technologies. These changes can make feasible region small or disconnected, then the currenttechniques cannot guarantee to search the optimum satisfying all of the constraints. Therefore, it is necessary to analyze the research trends in constrained global optimization techniques to focus on the constraints and to find the globaloptimum successfully on the feasible regions. This research focus on the Kriging surrogate model based constrained global optimization techniques, because they utilize statistical models of the functions to determine the next sample pointswith the highest probabilities, statistically. It means that the optimization process can consider the current stochastic error of the prediction at an untried point. The sampling criteria and technique features of each technique are mainly analyzed. Finally, the current limitations of Kriging surrogate model based constrained global optimization techniques and ways to overcome them are discussed.