摘要
针对简单遗传算法中的线性适应度、恒定交叉与变异概率等不能动态地适应整个寻优过程,提出采用非线性适应度与自适应交叉、变异概率的改进遗传算法。以典型的遗传算法测试函数验证改进遗传算法的有效性与可行性,最后将改进遗传算法用于离散变量桁架结构优化设计,计算结果表明改进遗传算法是可行、有效的。
Aiming at simple genetic algorithm with linear fitness and invariable crossover and mutation probability dynamically unfit for the whole optimal process, a modified genetic algorithm with nonlinear fitness function and adaptive crossover and mutation probability is presented in this dissertation. The availability and feasibility of the modified genetic algorithms are proved by genetic algorithm＇ s testing function. Finally, modified genetic algorithm solve optimal problem of truss structure with discrete variables. The result obtained by modified genetic algorithms indicates that algorithm is effective and feasible on the structural optimization with discrete variables.
出处
《机械强度》
EI
CAS
CSCD
北大核心
2005年第6期766-769,共4页
Journal of Mechanical Strength
关键词
遗传算法
离散变量
结构优化
Genetic algorithm
Discrete variables
Structure optimization
