Using Hybrid Artificial Bee Colony Algorithm and Particle Swarm Optimization for Training Feed-forward Neural Networks

The Artificial Bee Colony Algorithm (ABC) is a heuristic optimization method based on the foraging behavior of honey bees. It has been confirmed that this algorithm has good ability to search for the global optimum, but it suffers from the fact that the global best solution is not directly used, but the ABC stores it at each iteration, unlike the particle swarm optimization (PSO) that can directly use the global best solution at each iteration. So the hybrid of artificial bee colony Algorithm (ABC) and PSO resolved the aforementioned problem. In this article, Hybrid ABC and PSO is used as new training method for Feedforward Neural Networks (FFNNs), in order to get rid of imperfections in traditional training algorithms and get the high efficiencies of these algorithms in reducing the computational complexity and the problems of Tripping in local minima, also reduction of slow convergence rate of current evolutionary learning algorithms. We test the accuracy of our proposal using FFNNs trained with ABC, PSO, and Hybrid ABC and PSO. The experimental results show that ABCPSO outperforms both ABC and PSO for training FFNNs in terms the aforementioned Imperfections.