A Radial Basis Function Neural Network Trained with Aquila Optimizer for Solving Inverse Modeling Problems

Inverse modeling refers to approximately design the reverse characteristics of a plant or system. It is one of the most important mathematical problem which finds several application in solving engineering problems like communication channel equalization, close loop controller design, direction of arrival estimation, spectral estimation, etc. In this manuscript, a nonlinear inverse modeling problem channel equalization is addressed. The modeling is carried out with a radial basis function neural network (RBFNN). The weights of RBFNN are trained by an Aquila optimizer algorithm. The Aquila optimizer is introduced in 2021 by Abualigah et al. is based on the search behavior of Aquila to catch the pray. Here the Aquila optimizer enhance the convergence and thus the inverse modeling becomes more accurate. Simulation studies are demonstrated on design of inverse models for two nonlinear digital communication channels under noisy scenario. The obtained results are analyzed with the same RBFNN model trained by nature inspired Gray Wolf Optimizer.