基于直接搜索方法的层合板多目标优化设计

直接多目标搜索方法(DMS)是一种不需要计算梯度信息并且能实现全局收敛的多目标优化方法。基于直接多目标搜索方法,以简支层合板铺设角度为设计变量,基频和声功率为目标函数进行层合板结构振动与声多目标优化。分别以4层、8层复合材料层合板为例,用DMS方法对其优化设计,并与传统的遗传算法(GA)对比。结果表明,对于4层复合材料层合板,DMS方法比GA方法优化速度快,且能得到全局最优解;对于8层复合材料层合板,用DMS方法比4层板优化所需时间多,但相比GA方法,DMS方法优化更快。

Multiobjective Optimization of Laminated Composited Plate Based on the Direct Multisearch Method

Direct MultiSearch (DMS) is a kind of multiobjective derivative-free optimization method and can realize the global convergence of the multi-objective optimization method. The fiber laying angles were chosen as design variables and the fundamental frequency and power were the objective function for laminated plates structural vibration and acoustic multiobjective optimization based on the DMS method. Four-layer and eight-layer laminated plates were used as examples,using the DMS solver and solutions compared with genetic algorithms. The results show that, for four-layer of composite laminated plates, DMS method converges faster than GA method and can get the globally optimal solution. For eight-layer of composite laminated plates, DMS method need more elapsed time than four-layer plates, but is faster than GA method while used in the eight-layer plates.