Semi-dynamic Routing and Spectrum Assignment with variable bandwidth in Elastic Optical Networks. Bee-inspired Algorithms approach

Abstract

Elastic Optical Networks (EON) are a considerably new technology and have a promising future due to their fast speed, flexibility, and spectrum efficiency. The main point to consider in EON networks is the routing and spectrum assignment (RSA), an NP-complete problem. This work focuses on cases where there is semi-dynamic traffic with variable bandwidth. Initially, we broadly study techniques already proposed to face this problem, focusing on the best use of the spectrum and thus avoid spectrum reassignment, which will be the main objective of this work. This work proposes two promissory algorithms based on Artificial Bee Colony (ABC) and Bee Colony Optimization (BCO) with the encoding of bee based on routing and permutation. For both encodings, we use Fixed Alternative Routing and Mid-Fit and First-Fit spectrum assignment. Various and extensive simulations were conducted with different traffic loads and network topologies considering blocking probability and entropy measures. The efficiency of the algorithms varies according to the topologies. In general terms, routing-based ABC had better results in larger topologies, while permutation-based BCO having outstanding results in smaller topologies. Still, it does not have good results in the same way for larger topologies.