基于无网格及混合遗传算法的矩形加肋板肋条布置优化

Rib layout optimization for stiffened rectangular plates based on meshless method and hybrid genetic algorithm

将加肋板看作板与梁的结合体,基于无网格法并结合遗传算法,对矩形加肋板的肋条布置位置开展了优化,使加肋板在横向荷载的作用下中心点挠度最小。相较于传统有限元方法,利用本文的无网格法对加肋板进行肋条位置优化分析不需要对网格进行重构,离散在板与肋条上的节点始终不需要发生改变。本文在遗传算法的基础上加入了约束随机方向法形成混合遗传算法,使收敛速度加快,计算的重复率降低,遗传算法计算代数明显降低至两三代便可以得到较好的结果。

Recently, optimization problems have attracted more and more research interests. Stiffened plates are widely used in engineering, and optimization design and analysis on the plates possesses high engineering value. In current research, a stiffened plate is regarded as a combination of a plate and beams, and an optimization method based on meshless method and hybrid genetic algorithm is proposed for rib layout optimization of the plate to minimize its central deflection. Compared to the traditional finite element method, in the optimization process, the proposed method does not need remeshing of the plate, and the nodes on the plate and ribs do not need to be changed. As a modern bionics algorithm, the genetic algorithm has been applied to many optimization problems. In current research, the concept of the constraint random direction is added to the genetic algorithm, which results in a hybrid genetic algorithm to speed up the convergence rate and to reduce the repetition rate. The validity of the proposed method is demonstrated by several examples.