遗传算法的改进策略及其应用

Improved Genetic Algorithm and Its Application

针对基本遗传算法在优化应用中遇到的诸如局部搜索能力差、计算量大、对较大搜索空间适应能力差和早熟收敛等棘手问题 ,本文将传统的单纯形搜索算法与遗传算法结合 ,提出了一种混合遗传算法 ,并在进化操作中将交叉 ,变异概率与个体的适应度结合提出了一种个体自适应概率 ,使算法在产生新个体的同时 ,最大限度地保护优秀个体不受破坏 ,从而提高了整个算法的运算效率 ;在上述基础上又引入了加速循环操作以加强对优秀个体的利用率 ,进一步提高算法的效率。通过典型测试函数的数值算例验证 ,表明本文方法不但可以有效地克服遗传算法的上述缺陷 ,而且计算速度、稳定性、精度都有明显提高。

Aimed at the limitation of genetic algorithm applied in practice such as poor local search ability, premature convergence, excessive computational cost and bad adaptability to large search space, a kind of hybrid genetic algorithm is developed by incorporating simplex algorithm into genetic algorithm, and individual self-adaptive probabilities of crossover and mutation related with individual fitness are proposed to protect high fitness individuals from being destroyed. As a result, the efficiency of genetic algorithm is improved. On the basis of the above-mentioned techniques, accelerated cycle operation is incorporated into the algorithm to strengthen the utilizing efficiency of high fitness individuals. Typical test function case studies show that the improved genetic algorithm is more efficient and reliable.