利用雷达海杂波反演海洋蒸发波导的目标函数具有特殊复杂性,反演优化算法的性能严重影响蒸发波导参数的反演速度与精度。文中在蒸发波导两参数反演中引入了一种新的遗传-粒子群混合优化算法,该算法加入了Clerc 收缩因子,克服了传统单一算法在反演中的不足。通过仿真反演验证,结果表明,该算法较遗传算法和粒子群算法具有更快的收敛速度与更高的反演精度,波导高度和强度的反演平均差减小约47%以上。基于实测雷达海杂波数据的反演结果表明,该算法反演稳定性更高,更有利于海洋蒸发波导剖面反演的工程应用。

The objective function of optimization algorithm for the inversion of evaporation duct encountered in open oceans based on radar sea clutter data is distinguishingly complicated, so the performance of optimization algorithm affects the inversion speed and precision of evaporation duct deeply. In this paper, a new hybrid genetic algorithm-particle swarm optimization (GA-PSO) is introduced to the two-parameter inversion process of evaporation duct, which employs the Clerc convergence factor in the particle velocity update formula, overcoming the defects of the single optimization algorithm in duct inversion. Through the emulation inversion, the results show that the hybrid GA-PSO algorithm is faster than the GA and PSO in convergence and more accurate in inversion, the mean error of the inversed duct height and strength reduced by more than 47??. In addition, the inversion based on real sea clutter data indicates the GA-PSO algorithm has a higher inversion stability, which benefits the engineering-oriented application of oceanic evaporation duct inversion.

TN959

国家自然科学基金青年科学基金资助项目(61801446)