一种改进的遗传算法

A modified genetic algorithm

采用实数编码的遗传算法 ,在基于适应值比例和最优保留策略结合的选择、数值交叉及一致变异的基础上对算法作了两方面的改进 ,即对交叉操作采用剔除无效个体和保留父代进入选择 ,在算法出现未成熟收敛的趋势时插入混沌序列 ,增加群体多样性 ,以判断算法搜索特性。对函数优化仿真结果显示 ,新方法提高了搜索精度 ,较好地克服了未成熟收敛现象 ,取得了较满意的优化效果。

Based on the combination of fitness property selection and elitist model, numerical crossover and uniform mutation, two methods have been put forward to improve the efficiency of real-coded genetic algorithms. The first method is eliminating the inefficient individuals in crossover operation and reserving the parents for participating selection. The second method is inserting chaos serials into the population when a trend of immature convergence appears. Simulation results of function optimization shows that with the presented methods,the searching precision is enhanced,the phenomenon of immature convergence is effectively overcome,and a satisfying optimization result is obtained.