遗传模拟退火算法用于浅海声速反演的仿真研究

研究了浅海声速反演中经验正交函数的参数寻优算法.通过数值仿真分析了模拟退火算法和遗传算法的优缺点,并在此基础上,提出了将二者相联合而得到遗传模拟退火算法,数值算例验证了该算法兼备二者的优点,克服了它们的缺点.以射线理论为基础,应用遗传模拟退火算法,仿真计算了浅海2种声速分布下的声速反演问题,取得了较好的结果.该方法具有不依赖初始条件且解为收敛等特点.

Simulation study of sound speed profile inversion in shallow water using genetic simulated annealing algorithm

To improve sound speed profile inversion in shallow water,analysis was done on the efficiency and shortcomings of simulated annealing algorithms and genetic algorithms.Proposed is a new genetic simulated annealing algorithm,a combination of the traditional algorithms that was proven effective by numerical simulation.Using ray theory and applying the genetic simulated annealing algorithm,two kinds of sound speed profiles were simulated in shallow water,with favorable results obtained.Furthermore,it has features of insensitive initial condition and convergent solution.