基于改进海洋捕食者优化算法和瑞雷波频散曲线的近地表地层参数反演

Inversion of near-surface parameters from Rayleigh wave dispersion curves based on an improved marine predators algorithm

面波勘探是获取近地表地质结构的重要手段,利用瑞雷波频散曲线可以反演得到地层横波速度和厚度或泊松比等参数.本文提出了一种新的优化策略来处理瑞雷波频散曲线的反演问题.改进海洋捕食者算法(ACMPA)是一种改进混沌初始化,引入自适应步长和精英竞赛机制优化的反演算法.采用混沌映射进行种群位置初始化,提高初始化种群位置的质量;加入自适应函数,增强全局搜索能力;引入精英等级制度,避免算法陷入局部极值点,从而从局部和全局寻优进行优化,提高反演的收敛速度.通过正演模拟和实测数据进行测试分析,证明了改进的海洋捕食者优化算法的有效性与稳定性.该方法可以利用瑞雷波频散曲线信息反演得到近地表地层的介质参数,其收敛精度和收敛范围明显优于其他优化算法,具有较强的应用前景.

Surface wave exploration is an essential method for obtaining the near-surface geological structure. The inversion of the Rayleigh wave dispersion curve can effectively obtain the stratum′s shear wave velocity and thickness or Poisson ratio. This paper proposes a new optimization strategy for the inversion from the Rayleigh wave dispersion curve. It introduces a Marine Predator Algorithm(ACMPA) with improved chaos initialization, adaptive step size, and elite competition mechanism optimization. The chaotic map is used for population position initialization to improve the quality of the initialized population position;the adaptive function is used to enhance the global search ability, and the elite hierarchy is introduced to prevent the algorithm from falling into the local extreme points, which can effectively balance global and local optimization and improve the convergence speed of the inversion process. Modeling and field data tests demonstrate the effectiveness and stability of the improved marine predator optimization algorithm. The proposed method can obtain the medium parameters of near-surface formation by using the Rayleigh wave dispersion curve. Its convergence accuracy and range are significantly better than other optimization algorithms, which have strong application prospects.