摘要
The genetic algorithm is a prol,able search method based on Darwin's natural selection and survival of the fittest, which is presented by John Holland. In theory the genetic algorithm can form a more complete algorithm by using bilogical evolution and genetic theory,but there are many I3roblemsneeded to be solved in practice,and premature,which is a serious problem,is one of them. In this paperwe analyze its causes from the hand of diversity of population,and give two theorems about it,and present an effective method of preventing the premature of genetic algorithm,which adopts the concepl ofdiversity of population and adaptive crossover and mutation rate. At last we make a simulation of four function by using the method as stated above,and the result of experiments shows that the Inethod isvery effective.
从群体多样性角度对成熟前收敛的成因进行了分析，并提出了两个有关定理．根据分析结果，提出了基于群体多样性和自适应交叉、变异率的改进遗传算法．通过四个典型函数的实例验证，证明了本文提出的改进遗传算法是一种有效算法．
