Multi-Objective Genetic Algorithm Concurrent Subspace Optimization (MOGACSSO) for Multidisciplinary Design

Abstract

Most real world design problems are complex and multidisciplinary, and there is almost always more than one objective (cost) function to be optimized simultaneously. While achieving an optimum design, it is often also desirable that the optimum design be robust to parameter variations. A design could be robust in terms of performance or feasibility or both. In this research, the capability of the Multi-Objective Genetic Algorithm Concurrent Subspace Optimization (MOGACSSO) Method is extended to generate Pareto points that are robust in terms of performance and feasibility. MOGACSSO can generate large number of non-dominated solutions for a multi-objective multidisciplinary design problem by formulating a multi-objective optimization problem in each disciplinary subspace. In the Robust-MOGACSSO method introduced in this paper, the multi-objective formulation within each subspace has been extended to include the mean and standard deviation of the performance of neighboring design points as additional objectives in order to drive the population of Pareto points towards robust designs. While evaluating the mean and standard deviation of performance around a design’s neighborhood, constraint violations are incorporated as a penalty to the performance measure, thereby helping to achieve feasibility robustness, as well.