A Fault-tolerant Multicast Routing Algorithm in ATM Networks

This paper presents a genetic algorithm based method to solve the capacity and routing assignment problem arising in the design of selfhealing networks using the Virtual Path (VP) concept. Past research has revealed that Preplanned Backup Protection method and the Path Restoration scheme can provide a best compromise on the reserved spare capacity and the failure restoration time. Base on a set of customer traffic demands, we will determine the routings of the working and backup virtual paths to satisfy the demands, so that the traffic is 100% restorable under single point of failure and at the same time the amount of spare capacity can be minimized. In contrast to the past researches on this area, we will use Genetic Algorithm instead of linear programming. There are several advantages of using Genetic Algorithm which included faster and lower computation cost in reaching a reasonably good virtual path routing scheme, it can tackle multiple objectives function effectively, and less complex mathematical formulation. We will show that the results obtained using Genetic Algorithm is better than those results obtained by using heuristic approaches. Another contribution of this paper is our method not only can work on the Unicast traffic context just as the past researches, but also Multicast traffic.