OPTIMAL ROUTING IN MULTIPLE I O DATA NETWORK USING NEURAL-NETWORK WITH PERTURBED ENERGY FUNCTION/

An optimal routing problem in multiple I/O data network is one of the most important problems related on the performance of the network basi cally, and is formulated as a constrained nonlinear optimization probl em. When solving the problem considered using neural networks, we may obtain local minima, rather than global minimum, because the problem h as multimodals. In this paper, we introduce a perturbed energy functio n into the neural network based on a penalty method to solve the multi modal nonlinear optimization problem. (C) 1997 Elsevier Science Ltd.