半确定性的遗传算法

Half-determined Genetic Algorithm

在采用遗传算法进行函数优化的过程中存在的一个严重的缺陷是 :算法能很快地收敛到最优点附近 ,但要达到最优点需要很长的时间 .针对这一缺点通过如下 3个新的遗传算子构建了一个新的遗传算法 :半确定性的变异算子 ,Hamming- decreasing和设计变量的自适应定标 ,以加强遗传算法的局部搜索能力 .该算法充分利用基于概率的遗传算子的全局搜索能力和新算子较强的局部搜索能力 .算例表明 ,新的遗传算法具有很强的鲁棒性和较快的计算速度 .

There is a serious shortcoming in function optimization with GAs:althought they rapidly locate the region in which a global optimum exists,they locate the optimum with similar speed.A new GA with three new strategies-determined mutation,half determined Hamming decreasing and self adaptive scaling of design variables-is proposed for cutting this shortcoming.The new GA takes the advantages of the global searching of genetic operators based on probabilities and the advantages of the local searching of the new operators.The numeric examples show that the new GA is very robust and very fast.