摘要
遗传算法是一种借鉴生物界自然选择和进化机制的随机优化算法。它在求解一般全局优化问题时具有较好的鲁棒性,而且搜索不依赖梯度信息。但是,在用传统遗传算法解决较复杂的优化问题时,存在早熟及稳定性差的缺点。因而,针对这些缺点,出现了很多对传统遗传算法的改进。本文对遗传算法的3种改进方法进行了描述,并将它们应用到一个函数优化实例中。最后,通过比较3种改进方法与传统遗传算法优化所得结果,得出3种改进方法效果更好。
Genetic Algorithm is a kind of randomized search algorithm drawing on biological mechanisms of natural selection and evolutionary development.It has good robustness when it is used to solve the global optimisation problem,and the search does not depend on gradient information.But for more complex optimization problem,it has some disadvantages such as bad stability and premature convergence.There are many improved methods.In this paper, three improved methods of genetic algorithm are described,and are applied to an example of a function optimization.The results show us that they outperform traditional genetic algorithm.
出处
《电子测试》
2011年第3期38-40,共3页
Electronic Test
关键词
遗传算法
拟随机序列
变异概率
双种群遗传算法
genetic algorithm
quasi-random sequence
mutation rate
dual population genetic algorithm
