论文部分内容阅读
Many engineering optimization problems are multi-objective and have uncertainty.It is desirable to obtain solutions that are multi-objectively optimum and robust.In this work, the uncertain parameters with sufficient information are treated by random distributions, while some ones with limited information can only be given variation intervals.The robust multi-objective optimization model thus can be formulated with mixed variables.The mean and variance of random interval variables are calculated by the Monte Carlo simulation in the inner layer.Then, the Non-dominated Sorting Genetic Algorithm Ⅱ (NSGA-Ⅱ) are adopted as optimization operator in the outer layer.To improve the optimization efficiency, the Kriging models are constructed for the uncertain objective and constraint functions based on the Latin Hypercube Design(LHD).The multi-objective robust optimization method is combined with the approximation models to form an efficient robust multi-objective optimization method.The numerical examples are presented to demonstrate the effectiveness of the proposed method.