Parameter tuning of differential evolution algorithm for microseismic location

Fast and accurate source location is crucial for microseismic monitoring. Stochastic optimization algorithm is derivative-free and just need random solutions as the initial model, and it is quite suitable for non-linear seismic location problem. In this work, we utilize differential evolution, which is a fast and robust global optimization method and belongs to evolutionary algorithms, to speed up microseismic location with waveform-based methods. Parameter tuning of differential evolution for two waveform-based location methods, namely diffraction stacking and cross correlation stacking, is studied and reference ranges of individual parameters are obtained. Field data examples indicate that parameter tuning is necessary to ensure the performance of differential evolution, and the convergence features of the imaging functions of different stacking operators for microseismic source location can be revealed.