摘要
由于在进化算法中,一般的选择算子使适应度小的个体很早被淘汰,使它们的一些优良特征无法在种群中得到传播,但却增大了种群中优良个体繁衍的机会,因此可能会导致过早收敛或仅得到局部最优。对该算法加以改进,根据个体是否可行,将种群分为可行个体及非可行个体两部分,这样非可行个体也有机会参与到进化中。仿真结果表明,该方法对求解约束优化问题是很有效的。
In the evolutionary algorithms, some individuals with small fitness will soon be eliminated through selection, so the good character of them can not be passed to the next generation, Thus the chance of better individuals reproducing is increased, but early converging or local optimal is easy to occur, So a modified evolutionary algorithm is presented, Population is divided into two groups based on feasibility, so that each individual have chance to reproduce. The simulation result shows that the proposed algorithm performs well on constrained optimization problems.
出处
《控制工程》
CSCD
2007年第6期583-585,共3页
Control Engineering of China
基金
国家自然科学基金资助项目(70271066)
关键词
进化算法
可行性
可行个体
非可行个体
选择算子
evolutionary algorithm
feasibility
feasible individual
infeasible individual
select operator
