On the use of a stochastic estimator learning algorithm to the ATM routing problem: a methodology

This paper presents a dynamic solution to the Routing problem in (ATM) networks suitable for real-time applications. The proposed methodology employs a Stochastic Estimator Learning Algorithm (SELA) specially designed for the Routing problem. The feedback of the SELA Routing algorithm is computed from the utilisation of the links that comprise a route which is a possible solution to a. specific Routing problem. The algorithm promotes the load balancing in light loading conditions, whereas as the load of the network increases it tends to behave as a shortest path algorithm trying to minimise the cell loss probability and the call dropping tate. Simulation results show the effectiveness of the proposed scheme in comparison with some of the most known Routing algorithms of the literature. Some interesting directions for future work are also discussed. (C) 1998 Elsevier Science B.V.