Validity of Reliability-Based Multiobjective Optimization Incorporated MOPSO with SLSV
This study investigates the validity of the reliability-based multiobjective optimization (RBMO) algorithm proposed by the authors.The algorithm incorporates the hybrid multiobjective particle swarm optimization (MOPSO) with the constraint satisfaction technique into the single-loop-single-vector method (SLSV), known as the single loop reliability-based optimization technique.The hybrid MOPSO has constraint handling technique.When the design vector violates the constraints during searching iteration, the design vector is moved to the feasible domain using the constraint gradient information, and then, is moved to the feasible boundary using bi-section method.Because most Pareto optimum solutions exist on the feasible boundary, this algorithm has sufficient searching efficiency of the Pareto optimum solutions.The RBMO algorithm uses the concept of SLSV method that makes the reliability constraint into the approximate deterministic constraint by a proper approximation of the design point.The RBMO method is applied to an engineering design problem to demonstrate the validity of the RBMO method.
Reliability-Based Multiobjective Optimization Pareto Set Hybrid MOPSO SLSV Sensitivity Analysis
Shoichiro Kawaji Nozomu Kogiso Masayoshi Ohara Atsushi Ishigame Keiichi Sato
Dept.of Aerospace Engineering,Osaka Prefecture University,Sakai,Japan Dept.of Electrical Engineering and Information Science,Osaka Prefecture University,Sakai,Japan Research Laboratory,IHI Corporation,Yokohama,Japan
国际会议
黄山
英文
1-8
2012-06-18（万方平台首次上网日期，不代表论文的发表时间）