Active GA Accelerated by Simulated Annealing to Solve SPP in Packet Networks.

This paper presents two approaches to deal with the shortest path problem (SPP) solution for routing network packets in an optimized way. The first one uses Simulated Annealing (SA), and the second one is a novel hybridization of the Genetic Algorithm with Dijkstra mutation accelerated by the SA (SGA). Also, two different case scenario configurations, each with 144 nodes, are employed to assess these two proposals, and the total time spent to fill out the routing tables, referring to the transmission of a packet from the initial to the destiny nodes, is measured. A statistical comparison is applied to identify differences among the algorithm's solutions. Experiments and simulations have shown that the SGA presented competitive results compared to standard SA and can solve the problem with fast convergence, which makes us conclude that it can operate efficiently in actual computer networks.