制約の感度を用いるハイブリッド型多目的PSOを利用した信頼性に基づく多目的最適設計

概要

論文の詳細を見る
This paper proposes a new algorithm for the reliability-based multiobjective optimization (RBMO) to obtain Pareto set under reliability constraints that consider uncertainties on design parameters such as material properties or load conditions. The RBMO algorithm consists of the hybrid-type multiobjective particle swarm optimization (MOPSO) with constraint satisfaction technique proposed by the authors and an efficient single loop reliability-based optimization approach. The hybrid-type MOPSO has a function to move the design candidate with constraint violation to the feasible boundary using constraint sensitivity information and the bi-section method. The feature is suitable for evaluating the reliability that requires sensitivity of the limit state function. In addition, the single-loop-single-vector method (SLSV) known as a single loop approach of the reliability-based design optimization is integrated to improve the computational efficiency. Through numerical examples, the effectiveness of the proposed algorithm is demonstrated.