改进的多学科协同优化方法

Improved multidisciplinary collaborative optimization method

针对标准协同优化方法系统级优化存在自身内部缺陷而导致计算困难的问题,提出一种改进的协同优化方法。利用凝聚约束处理方法,将系统级优化的多个等式约束凝聚成一个单约束,结合惩罚函数法将其转化为无约束优化问题,并针对系统级优化采用传统优化算法对初始点选取敏感等问题,将改进的粒子群优化算法应用于协同优化。利用数值计算和减速器设计两个典型算例,对所提方法的可行性和有效性进行了验证。

Aiming at the problem of calculation difficulty caused by collaborative optimization method＇s internal defect,an improved collaborative optimization method was proposed.By using the coherent constraint handling method,several constraint conditions of the system-level optimization equations were agglomerated into a single constraint condition.Combined with the penalty function method,this single constraint was transformed into no constraint problem.For those initial points selection which were sensitive in the conventional optimization methods,the improved swarm particle optimization method was applied in collaborative optimization.The proposed method was verified by two typical examples of numerical calculation and decelerator design,and the feasibility and effectiveness of this method was proved.