Heuristic Optimization for the Restoration Scheme of the ATM Network

This paper addresses the heuristic optimization algorithms for the restoration scheme of the ATM network based on VP (virtual path). According to the independence between bandwidth allocation and routing in ATM, it is possible to be implementation of real-time algorithm for the search the backup paths of the failed link or node. The performance of restoration depends on the search method for the alternative routes available to restore the failed target VP. Two heuristic optimization approaches are introduced and the application for the restoration scheme are examined to handle link and node failure by using a genetic algorithm (GA). We discuss the main components of the proposed GA, since the performance of a GA heavily relies on them. In particular, we develop a new genetic operator called heuristic crossover (HX) incorporating problem-specific information. Extensive experiments are carried out to evaluate the performance of the proposed GA