Routing and Spectrum Assignment Algorithm with Traffic Prediction and Periodic Rerouting in Elastic Optical Networks

Abstract

Routing and spectrum assignment (RSA) is a key issue in elastic optical networks (EONs). A new RSA algorithm entitled traffic prediction and periodic rerouting (TPPR) is proposed to optimize the utilization efficiency of network resource in this paper. In the proposed TPPR algorithm, we use the radial basis function neural network (RBFNN) to predict traffic variations on each fiber link, which is beneficial to select a better lightpath from all candidate lightpaths for the current arrival connection request. Simultaneously, we also periodically update link weights and thus reroute all candidate lightpaths for each source-destination nodes pair after the proposed TPPR deals with a batch of connection requests. Besides, distance-adaptive is also used in TPPR algorithm. Under both the Poisson traffic model and the self-similar traffic model, simulation results show that the proposed algorithm can effectively reduce the blocking probability and improve resource utilization.