Direct Search Simulated Annealing for Nonlinear Global Optimization of Rayleigh Waves

Abstract

Nonlinear global optimization of Rayleigh wave dispersion curves not only undergoes computational difficulties associated with being easily entrapped in local minima for most local-search methods but also suffers from the high computational cost for most global optimization methods due to its multimodality and its high nonlinearity. In order to effectively overcome the above described difficulties, we proposed a new Rayleigh wave dispersion curve inversion scheme based on Direct Search Simulated annealing (DSSA), an efficient and robust algorithm which hybridized direct search methods, as local search methods, and simulated annealing, as a meta-heuristic method. The performance of the proposed procedure is tested on a four-layer synthetic earth model and a real-world example. Results from both synthetic and real field data demonstrate that DSSA applied to nonlinear inversion of Rayleigh waves should be considered good not only in terms of computation time but also in terms of accuracy due to its global and fast convergence in the final stage of exploration.