摘要
针对基本遗传算法存在容易"早熟",无法全局收敛的现象,设计了一种新交叉算子和变异算子,并在遗传算子构造中引入贪心控制策略.新算子的引入丰富了种群的多样性,提高了算法的全局搜索能力.实例仿真表明,改进遗传算法在迭代陷入局部最优时,能在较短的时间内跳出局部最优,继续寻找全局最优解.
Premature convergence usually appears in basic genetic algorithm. So, new crossover and mutation operators are designed. Greedy strategy is introduced in construction of genetic operator. Diversity of population becomes Rich because of introduction of new operators. New algorithm improves the ability of global search. The simulation indicates that the improved genetic algorithm can jump out of local optimum in a short time, and continue seeking the optimum.
出处
《计算机系统应用》
2012年第9期192-194,191,共4页
Computer Systems & Applications
基金
国家自然科学基金(50904032)
湛江师范学院青年基金(QW0712)
关键词
早熟
遗传算子
全局搜索
仿真
局部最优
premature convergence
genetic operator
global search
simulation
local optimum