哈尔滨工程大学学报2025，Vol.46Issue(4)：643-651,9.DOI:10.11990/jheu.202309001
遗传算法优化神经网络在地声参数反演中的应用
Application of neural network optimized by genetic algorithm in geoacoustic parameter inversion
摘要
Abstract
To address the issue of low estimation accuracy of geoacoustic parameters by traditional matched-field in-version methods in shallow water environments,a BP neural network algorithm optimized by a genetic algorithm(GA-BP)is applied to geoacoustic parameter inversion.First,the sensitivity of the vertical spatial correlation coef-ficient of the noise field to changes in geoacoustic parameters is simulated and analyzed.Then,the performance of GA-BP in inverting geoacoustic parameters is evaluated.Finally,GA-BP is used to process measured ocean ambi-ent noise data to estimate the bottom density,sound velocity,and attenuation.Simulation and experimental results show that GA-BP has a faster net training speed and higher inversion accuracy compared to the BP neural network algorithm.GA-BP accurately inverts the geoacoustic parameters of the Pekeris waveguide,and the spatial correla-tion coefficient curve of the ocean ambient noise field obtained through inversion closely matches the experimental measurements,with a Pearson correlation coefficient of 0.98.The results of this study demonstrate the efficiency and reliability of the GA-BP algorithm in geoacoustic parameter inversion,providing a robust technical solution for passive geoacoustic parameter estimation based on ocean ambient noise.关键词
海洋环境噪声/空间相关特性/敏感度分析/遗传算法/BP神经网络/Pekeris波导/地声参数反演/海上实验Key words
ocean ambientnoise/spatial correlation characteristics/sensitivity analysis/genetic algorithm/BP neu-ral network/Pekeris waveguide/geoacoustic inversion/sea experiment分类
海洋科学引用本文复制引用
赵振星,李琪,黄益旺..遗传算法优化神经网络在地声参数反演中的应用[J].哈尔滨工程大学学报,2025,46(4):643-651,9.基金项目
国家自然科学基金项目(12074088). （12074088）
