摘要
退火遗传算法是模拟退火算法(SA)和遗传算法(GA)相结合的优化算法。其主要思想是利用模拟退火算法中的热槽法(概率分布函数)产生GA的群体初值,这样可以避免产生一些无用的群体,其结果既节省计算机空间,又节省计算时间,并且最终容易得到最优解。SA和GA交替运行,对GA的每一演化代,用SA对其进行局部搜索来改进GA的性能。利用退火遗传算法对地震波阻抗进行反演,目标函数可以达到10^(-5)数量级,这就使退火遗传算法的非线性方法能够应用于地震波阻抗反演中。
Annealing and genetic algorithm is an optimized
algorithm that integrate advantages of stimulated an-
nealing (SA) and genetic algorithm (GA). The heat-
bath method of SA is adopted to generate the initial
colonies of GA, which helps avoid useless colonies,
save computer memory and computing time, and ob-
tain optimal solution. The SA and GA run alternate-
ly. At each evolutionary generation, the SA is adopted
to improve the local performance of the GA. The an-
nealing and genetic algorithm can achieve an order of
precision as high as 10^(-5) in wave impedance inversion.
出处
《石油物探》
EI
CSCD
2004年第3期234-237,4,共4页
Geophysical Prospecting For Petroleum
关键词
退火遗传算法
地震波阻抗
反演
层速度
温度
反射系数
annealing and genetic algorithm
genetic algorithm
simulated annealing
wave impedance
inversion
