On the use of learning automata for energy saving in optical backbone networks

Optical backbone networks offer reliable long distance connectivity with low error rate and high bandwidth. Lightpath routing in these networks can be fulfilled while aiming at either performance or energy efficiency, according to current user needs. In this paper, an innovative hybrid scheme for creating node-to-node connections that is based on Learning Automata and can be oriented to performance or energy efficiency, is proposed. It is suited for network's Operation phase when a node bypass technology is absent. This phase requires low delay while routing, so complexity is reduced to a polynomial form by the use of an adaptive mechanism. Simulation results verify scheme's efficiency.