摘要
传统的方法解答非线性规划问题存在着有进入局部最优、效率低、甚至根本得不出结果的局限性,而遗传算法中用不变的p_c和p_m来控制进化,很容易导致早熟,降低算法搜索效率。根据适应度自动对交叉概率p_c和变异概率p_m进行调整,提出一种新的遗传算法。通过对6个测试函数的仿真实验,结果表明本算法是非常有效的。
The traditional methods to answer the question of non-linear programming have some limitations in its access to local optimum, low efficiency, and even no results obtained. And genetic algorithm with the same probability of crossover and mutation probability to control the evolution very easily lead to early maturity and reduce the efficiency of the algorithrn. The cross- over probability and mutation probability should be adjusted automatically According to fitness, thus a new genetic algorithm was proposed. Simulation results for six test functions, show that the new algorithm proposed in this paper is very effective.
出处
《河北工业科技》
CAS
2009年第6期461-463,共3页
Hebei Journal of Industrial Science and Technology
基金
陕西省教育厅专项科研资助项目(03jk065)
西安建筑科技大学基础研究基金资助项目(DD12006)
关键词
遗传算法
交叉概率
变异概率
适应度函数
genetic algorithm
crossover probability
mutation probability
fitness function
