在水声环境参数反演中,优化算法对反演结果有重要的影响,提出用遗传算法和Levenberg—Marquardt算法组成的混奢优化算法进行浅海环境参数匹配场反演,利用仿真数据对算法的有效性进行了评估,并与遗传算法、模拟退火算法进行了比较,最后采用实验数据验证了方法的有效性.
The optimization algorithms have significant effect on inverting the underwater acoustical parameters. Matched fieht inversion is conducted by using a hybrid method which combines the global genetic algorithms and the local Levenberg-Marquardt method. The combination of a global and a local method makes the inversion faster and results in a state closer to the global minimum. The inversion analysis based on simulated data and experimental data shows that the hybrid method works better compared with the individual global methods.