基于多目标遗传算法的鱼雷多学科稳健优化设计

传统的多学科设计优化多为确定性优化,这种设计优化忽视了系统本身固有的不确定性因素,求得的最优解往往只具有数学意义,很难应用于工程实践。本文将多目标遗传优化算法与多学科设计优化方法相结合,同时考虑了稳健性对优化结果的影响,采用响应面近似方法获得系统耦合关系,通过效力系数把设计变量分配到相应的子空间中,在每一个子空间里都进行基于遗传算法的多目标优化,使各学科的模拟分析、优化设计得到并行设计。提高了优化效率。本文通过鱼雷总体多学科优化设计的算例证明了此方法的可行性。在工程实践中有一定的指导意义。

Multi-objective Genetic Algorithm for Multidisciplinary Robust Optimization Design of a Torpedo

Multi-objective genetic algorithm is applied to concurrent subspace optimization in multidisciplinary optimization design in this paper,and robustness is taken into account at the same time.Response surface is used to determine the coupling relationship.Design variables are distributed to each subspace using effective coefficient.Multi-objective optimization based genetic algorithm are carried out in each subspace parallelly.Optimization efficiency is greatly improved.Finally,optimal design of a torpedo is used as an example to validate the feasibility of the method.Results show that our method is feasible and effective.