摘要
提出一种针对大型复合材料结构铺层厚度和铺层顺序同时进行优化的整数编码并行遗传算法,并将该优化方法应用于复合材料螺旋桨结构优化问题。首先,在区域划分的基础上,对每一单层设定一个长度控制因子来决定单层的铺设区域,以实现对大型复合材料结构的整体一次性优化。然后对遗传算法进行改进,使之成为一种能同时优化铺层厚度和铺层顺序的高效算法。并采用并行编程语言标准MPI(message passing interface)构建并行编程环境,利用主从式并行遗传算法框架,实现遗传算法在单机多进程上的并行计算。最后针对复合材料螺旋桨结构进行优化设计,以验证该方法的高效性,并分析并行遗传算法的加速效果。
An integer-coded parallel genetic algorithm is presented for solving large scale composite structures optimization problem.This optimization method,which is applied to the structure optimization of a composite propeller,makes staking sequence and staking thickness be optimized simultaneously.First,on the basis of region division,a length-control factor for every ply is added to decide the regions of composite plies.Second,the GA（genetic algorithm） is improved to make it a high-efficiency algorithm that can optimize staking sequence and staking thickness simultaneously.Then,MPI（message-passing interface） is adopted to build parallel program environment and master-slave parallel GA to fulfill the parallel calculation of GA through single computer multi-process method.Finally,the optimization method proposed is used to optimize a composite propeller to demonstrate the high efficiency.The acceleration effect of the parallel GA is also analyzed.
出处
《机械强度》
CAS
CSCD
北大核心
2012年第2期210-213,共4页
Journal of Mechanical Strength
关键词
复合材料螺旋桨
铺层厚度
铺层顺序
优化
并行遗传算法
Composite propeller
Staking thickness
Staking sequence
Optimization
Parallel genetic algorithm
