摘要
针对遗传算法在迭代过程中经常出现的未成熟收敛、振荡、随机性太大等缺点,引入一种新的遗传算子——单亲遗传算子,用于对标准遗传算法的改进。包含单亲遗传算子的改进遗传算法能直接计算具有应力约束和截面尺寸约束的离散变量结构优化问题,也能处理同时具有稳定约束和位移约束的多工况、多约束、多变量的离散变量结构优化设计问题,进而对框架结构的多种工况进行优化设计的结果进行了对比验证,结果表明:改进遗传算法比标准遗传算法有好得多的收敛特性,迭代次数明显减少,优化设计结果也远好于标准遗传算法。
In view of the standard genetic algorithm (SGA) having the shortcoming of convergence early, large number of iteration and over-randomization, an improved genetic algorithm (IGA) composed of a one-parent reproduce operator is proposed. This method can be used directly to work out the structural optimum design with discrete variables, which involve the stress constraints and cross-section size constraints, and the method can be also used to deal with the structural discrete optimization with multi-loading, multi-constraints and multi-variables,and to deal with different building structural optimization. The results of examples expatiate that IGA has a higher searching effectiveness, and its performance convergence is greatly enhanced. IGA, is an effective, efficient optimal method, superior to those of SGA.
出处
《辽宁工学院学报》
2004年第3期49-52,共4页
Journal of Liaoning Institute of Technology(Natural Science Edition)
关键词
建筑结构
结构优化
全局最优
离散变量
改进遗传算法
单亲遗传算子
global optimum
discrete variables
building structure
structural optimization design
genetic algorithm
improved genetic algorithm
