Research on the inversion of passive-source Rayleigh wave dispersion curves based on the MWHO algorithm

The passive-source surface-wave method infers subsurface structures by analyzing dispersion curves extracted from ambient noise, which offers advantages in operational simplicity and greater exploration depth. However, the inversion of dispersion curves remains challenging due to the complexity of underground media and structures, constituting a complex nonlinear optimization problem. Existing approaches—including observational methods, linear local optimization, and nonlinear global optimization—each exhibit limitations. This study proposes a modified wild horse optimizer (MWHO) for passive-source Rayleigh wave dispersion curve inversion to address the constraints of current methods in deep, complex geological exploration. Four intricate geological models were tested: a five-layer velocity-increasing model, a model with hard interlayers, a six-layer velocity-increasing model, and a six-layer model containing dual high-/low-velocity layers. Results demonstrate that MWHO outperforms particle swarm optimization (PSO), genetic algorithm (GA), and differential evolution (DE) in stratigraphic identification and noise resistance, accurately reconstructing geological structures. In field applications under ground-fissure conditions, MWHO successfully inverted passive-source Rayleigh wave dispersion curves, producing 2D shear-wave velocity profiles consistent with borehole data. Future research should focus on refining MWHO, exploring its applications to other nonlinear inverse problems, and integrating advanced optimization techniques to enhance computational efficiency and accuracy.