The Rprop algorithm proposed by Riedmiller and Braun is one of the best performing first-order learning methods for neural networks. We introduce modifications of the algorithm that improve its learning speed. The resulting speedup is experimentally shown for a set of neural network learning tasks as well as for artificial error surfaces.

The success of learning as well as the learning speed of an artificial neural network (ANN) strongly depends on the initial weights. If problem or domain specific knowledge exists, it can be transferred to the ANN by means of a special choice of the initial weights. In this paper, we focus on the choice of a set of initial weights, well suited to fast and robust learning of all particular problems out of a class of related problems. Our evolutionary approach particularly takes the learning algorithm into consideration in the design of the initial weights. The superior properties of the initial weights resulting from this algorithm are corroborated using a class defined by solving a differential equation with variable boundary conditions.

Acknowledgements This dissertation would never have come about without the support and direction of Dr. J. Robert Mahan. No one could hope for a more engaged advisor and better role model. The author owes a debt of gratitude to the Conseil Régional de Lorraine for its generous financial support of this research at the European Campus of the Georgia Institute of Technology, located in Metz, France. And although never directly involved in this work, my family – including a certain dangerous redhead – has perhaps made the largest contribution through years of constant and unwavering love and support. iii