随机化均匀设计遗传算法

Genetic algorithm based on random uniform design

众所周知,遗传算法的运行机理及特点是具有定向制导的随机搜索技术,其定向制导的原则是:导向以高适应度模式为祖先的"家族"方向.以此结论为基础.利用随机化均匀设计的理论和方法,对遗传算法中的交叉操作进行了重新设计,给出了一个新的GA算法,称之为随机化均匀设计遗传算法.最后将随机化均匀设计遗传算法应用于求解函数优化问题,并与简单遗传算法和佳点集遗传算法进行比较.通过模拟比较,可以看出新的算法不但提高了算法的速度和精度,而且避免了其它方法常有的早期收敛现象,

It is known that the running mechanism and character of GA are random search technology with direction guidance whose principle leads to ＂family＂ direction with mode of high fitness for ancestor. Based on this conclusion and employing the theory and method of RUD, the crossover operation in GA is redesigned and a new algorithm called RGA is given. RGA is applied to finding the solution of function optimization, which is compared with simple GA and good point-set GA. Through simulated comparisons, it is observed that this new algorithm not only has superiority both in speed and accuracy but also avoids tbe early convergence phenomenon that the other methods often encounter.