Global optimization of controlled source audio-frequency magnetotelluric data with an improved artificial bee colony algorithm

Abstract

Abstract The inversion of controlled source audio-frequency magnetotelluric (CSAMT) data is a complex nonlinear problem. A linearization of this problem is easily trapped in local minima and the complexity of the artificial source makes CSAMT data interpretation more difficult than that of magnetotelluric (MT) data. This paper presents an improved artificial bee colony (ABC) algorithm for the 2.5D inversion of CSAMT data. New initialization and generation strategies are proposed to improve the optimization achieved by the original ABC algorithm. The global optimization of CSAMT by the improved ABC algorithm is realized based on 2.5D forward modeling theory and is used in the inversion of a complex model of water-bearing anomalous bodies in sandstone. Results show that the algorithm can accurately recover the resistivity and spatial distribution of strata and anomalous bodies. The survey data for a suspected collapse column in Shandong Province are also processed using the proposed method, and inversion via the algorithm accurately shows the water abundance of the suspected collapse column. Thus, the results of the theoretical modeling and practical data indicate that the improved ABC algorithm is effective for analyzing CSAMT data. Moreover, this algorithm improves the interpretational accuracy and resolution of CSAMT data.