Estimating model parameters from SP anomaly of sheet-shaped sources using differential search and particle swarm optimization algorithms

Abstract

Abstract In this study, the efficiency of estimating the model parameters of sheet-shaped single and multiple sources of the self-potential (SP) anomaly using the differential search algorithm (DSA) is investigated. First, noise-free and noisy synthetic anomalies are calculated for a single sheet-shaped source, and its model parameters estimated by DSA. The DSA inversion is also done for a model consisting of three inclined sheets. To test the effectiveness of the method, the same processes are repeated with a more conventional algorithm, particle swarm optimization (PSO), and the solutions of both methods are compared. The results of synthetic anomaly analyses show that DSA can predict the parameters as accurately as PSO. Then, both algorithms are also applied to two field SP anomalies (Surda and Beldih) that have been evaluated by different algorithms in the literature. The source of the Surda anomaly is modelled as one sheet, whereas the source model of the Beldih anomaly is assumed to consist of three sheets. The five model parameters for each model are estimated using both algorithms and it is determined that they are in good agreement with the findings of the previous studies. The contribution of the regional background anomaly to the synthetic and field anomalies are also included and regional coefficients are estimated. Finally, we conclude that DSA can solve the source parameters without the need for the initial values required in conventional iterative inversion methods and is an efficient and promising algorithm for determining the parameters of SP sources.