Temperature Forecasting Based on the Multi-agent Adaptive Fuzzy Neuronet

Abstract

This article presents a multi-agent adaptive fuzzy neuronet for average monthly ambient temperature forecasting. We fulfilled the agents of the MAFN based on neural networks. The automatic generation of the of the neuronet parameters of the optimal architecture is the most complex task. Due to train the optimum multi-agent adaptive fuzzy neuronet we modified the Ant Lion Optimizer and combined it with the Levenberg-Marquardt algorithm. We first applied the modified ALO to globally optimize the multi-agent adaptive fuzzy network’s structure in multi-dimensional space, and then we elaborated the Levenberg-Marquardt algorithm to speed up the convergence process. We generated an optimum multi-agent adaptive fuzzy neuronet architecture from the obtained global optimum which represented the MAFN optimum architecture’s parameters. The simulation results show that the proposed training algorithm outperforms the modified Ant Lion Optimizer and Levenberg-Marquardt algorithms in training the optimum multi-agent adaptive fuzzy neuronet for average monthly ambient temperature forecasting.