Interpretation of residual gravity anomalies caused by simple geometrical shapes using the improved Grey Wolf Optimisation algorithm

In this research, an algorithm inspired by nature, i.e. the Improved Grey Wolf Optimiser (IGWO), is used in an iterative process to simultaneously estimate the optimal value of parameters related to simple geometric models (sphere, horizontal cylinder, and vertical cylinder) in a multi-objective problem. The variables of each model are the parameters of the amplitude coefficient (A), depth (z), shape factor (q), and position of the model centre (x). In this modelling, each of the wolves are a model having the dimensions of the numbers of the model parameters. This algorithm was verified in two stages. First, the accuracy of the algorithm was investigated for the gravity data generated from the artificial models of the sphere, horizontal cylinder, vertical cylinder, and a combination of different models in two states, i.e. with and without noise. The results show that the values of the parameters obtained with the IGWO are almost equal to the actual parameters. Also, by adding noise to the data, the results are satisfactory. In the next step, the real gravity data related to two abnormality profiles from Iran and America were used in this modelling. In this step, the comparison between the results of the IGWO algorithm and the results of previous studies indicated the proper performance of the proposed method. The inversion results obtained with the mentioned method show that the amplitude coefficient, depth, and shape factor of the salt dome located in America are respectively about -270.32 mGal×km2, 4.54 km, and 1.51, and of the salt dome located in Iran are -270.32 mGal×km2, 63.83 m, and 1.49. The advantage of IGWO inversion is that it has few parameters to adjust, it estimates the optimal value of the parameters quickly, and also converges with a nonlinear convergence coefficient without getting stuck in local minima.