Weighting-Free Multi-Objective Global Optimization for Transdimensional Joint Inversion of Surface Wave Dispersion, Refraction, and Resistivity Data in Near-Surface Characterization

Abstract

We present a practical joint inversion framework for active-source Rayleigh-wave dispersion curves, refraction traveltimes, and electrical resistivity data to improve the accuracy and reliability of near-surface characterization. The framework uses a multi-objective global optimization strategy that avoids prescribing subjective weights in a single combined objective function. We develop a multi-objective Modified Barnacles Mating Optimizer by integrating Pareto dominance into the optimizer which enables simultaneous minimization of multiple data-misfit measures. Prior to inversion, modal and parameter-sensitivity analyses are conducted on a synthetic model. The experiments show that the problem is highly nonlinear, uncertainty-prone and characterized by strongly heterogeneous parameter sensitivities. The proposed approach is validated using synthetic data and benchmarked against conventional single-objective inversions using equal-weighting and randomly assigned weights under identical settings, and it is further demonstrated on some real datasets from T & uuml;rkiye. To support transdimensional inference, we also use a multiple-model space strategy that allows switching among model parameterizations without substantially increasing search complexity. Field results are interpreted using available geological and geophysical information, and post-inversion uncertainty analyses are conducted to assess solution credibility. The proposed framework provides a broadly applicable methodology for site characterization in geologically complex environments using transdimensional joint inversion.