A comparison study using particle swarm optimization inversion algorithm for gravity anomaly interpretation due to a 2D vertical fault structure.

A new approach to the inversion of gravity data utilizing the Particle Swarm Optimization (PSO) algorithm is used to model 2D vertical faults. The PSO algorithm is stochastic in nature; its development was motivated by the communal in-flight performance of birds looking for food. The birds are represented by particles (or models). Individual particles have a location and a velocity vector. The location vectors represent the parameter value. PSO is adjusted with random particles (models) and searches for targets by updating generations. Herein, the PSO algorithm is applied to three synthetic data sets (residual only with and without noise, residual plus regional, residual plus anomaly generated by a buried cylinder structure) and two field gravity data sets acquired across known faults in Egypt. Assessment of the synthetic data demonstrates that the PSO algorithm generates superior results if a first horizontal gradient (FHG) filter is applied first. The robustness of the PSO inversion algorithm is demonstrated for both synthetic and field gravity data. • New inversion approach is used to model 2D vertical faults based upon Particle Swarm Optimization (PSO) and first horizontal gradient (FHG). • PSO algorithm is applied to three synthetic data sets and two field gravity data sets. • The robustness of the new approach is demonstrated for both synthetic and field gravity data. [ABSTRACT FROM AUTHOR]