A Simulated Evolution-Tabu Search hybrid metaheuristic for routing in computer networks

Routing in computer networks is a nonlinear combinatorial optimization problem with numerous constraints and is classified as an NP-complete problem. There are certain important QoS metrics which affect the performance of a network. One of these metrics is the average network delay, which should be minimized. In this paper, a routing strategy based on Simulated Evolution algorithm to find suboptimal routing solution for computer networks while optimizing the above metric is presented. To intensify the search, a hybrid variant of the proposed algorithm has also been implemented. This variant incorporates Tabu Search characteristics into the Simulated Evolution algorithm. Performance evaluation of the two approaches is done via simulation. Empirical results suggest that the hybrid variant performs better than the original Simulated Evolution algorithm.