A RPROP method with weight perturbation

In this paper we present an improved ´resilient backpropagation´ method using weight perturbation technique. This method is designed to provide a very simple and effective means of searching the minima of objective function directly without any computation of derivative values. Therefore, it can be applied in hardware implementations easily. We test this algorithm on benchmark problems. For all problems, the systems were shown to train efficiently by our method.