Optimal path calculation for virtual networks using genetic algorithm
Man‐Soo Han · International Journal of Advanced Computer Research · 2019
Dijkstra's algorithm has been widely used for calculating a best routing path in networks for the PCE with the linear metric.Dijkstra's algorithm calculates the best path for the linear metric problem in a simple manner and within reasonable time.However, we find out that the Dijkstra algorithm cannot be applied in some cases that have performance costs in complex forms, especially when both link and performance costs are combined in a non-linear form.Many papers studied meta-heuristic algorithms for the VNE problem.In [3], a genetic algorithm (GA) based on a non-dominated sorting based multiobjective evolutionary algorithm is introduced to the VNE problem.But the link path between nodes is calculated by a shortest path first (SPF) method which is based on Dijkstra's algorithm.In [4], a particle swarm optimization method is used for the VNE problem.However, the SPF algorithm is used for calculation of the link path between nodes.In [5], an ant colony algorithm is introduced to the VNE problem.But, the link path between nodes is calculated by the SPF algorithm.We find out that the above methods cannot be applied to networks with non-linear costs, as all the methods above use the SPF algorithm for path computation.It is needed to