摘要
遗传算法的优良性能使其被广泛应用于现实许多工程领域中,但该算法由于随机搜索而带来的收敛速度慢、易产生局值、不稳定等问题,给其应用带来很大的困难。论文首先针对收敛速度慢,提出使用遗传迭代次数自适应控制选择算子,达到对收敛速度的自适应控制。其次,针对局值问题,提出一种新的改进自适应遗传策略,其交叉和变异算子能够根据前两代适应度变化进行自适应调整。最后,使用Matlab7.0对所选的函数进行优化仿真,通过比较仿真结果得出改进的自适应遗传算法在处理收敛速度和避免易产生局值方面具有较明显的优势。
Genetic algorithm has been widely applied to many actual engineering fields because of its excellent perform- ances. However, genetic algorithm has its drawbacks. For example, it has slow rate of convergence, problems of instability due to the random search and easy to produce local value. These drawbacks else bring a lot of difficulties to actual applica- tion. Firstly, for its first shortage, an optimal solution which uses the number of genetic interation to control selection opera- tor is proposed so that it could adaptively control rate of convergence. Secondly, to solve local convergence problem, a new modified self-adapting genetic policy is created. It can adaptively adjust crossover operators and mutation operators according to the change of two previous generation fitness. Finally, with employing Matlab7. 0 and simulating selected functions, it is concluded that this improved self-adapting genetic algorithm has evident advantages in convergent rate and avoiding local val- ue problem,
出处
《计算机与数字工程》
2014年第3期355-358,368,共5页
Computer & Digital Engineering
基金
甘肃省自然科学基金(编号:1010RJZA074)
甘肃省高等学校基本科研业务项目(编号:甘财教[2011]181)资助
关键词
遗传算法
自适应
性能仿真
genetic algorithm, self-adaption, performance simulation
