Azwar Riza Habibi, Trisilowati, Ratno Bagus Edy Wibowo
This paper develops neural network (NN) method using conjugate gradient (CG) with combination of particle swarm optimization (PSO) and genetic algorithm (GA). The combination of PSO and GA is used for weight initialization to improve the computational process and minimize the errors. CG method can change every learning rate of neural network, so that the addition of CG can increase the rate of convergence. PSO, by its calculation velocity, can get quickly the solutions and GA act to expand the searching area of PSO solution. This new algorithm is called CN-PSOGA. There are some criterions used to compare the CN-PSOGA algorithms with others, i.e. accuracy degree, number of iterations, and computation time produced by each algorithm. Simulation results show that the proposed algorithm can increase the accuracy of solution approach. © 2015, Asian Research Publishing Network. All rights reserved.
Department of Mathematics, Brawijaya University, Indonesia
