GAPS: The Genetic Algorithm-based Path Selection Scheme for MPLS Network

In the MPLS network QoS might be achieved by reducing the possibility of congestion while increasing the utilization of available links. Due to various interrelated factors, path selection for the MPIS network is known as an NP-hard problem. The widely employed path selection scheme has a shortcoming that the time overhead is high, especially when the number of nodes in the network is large. This paper proposes a new path selection scheme which uses the genetic algorithm along with the modified roulette wheel selection method. Performance evaluation through computer simulation shows that it finds the paths about 30% faster than the existing constrained shortest path first (CSPF) scheme while evenly distributing the traffic throughout the network.