Dynamic Reconfiguration Policies for WDM Networks

We study the issues arising when considering the problem of reconnguring broadcast optical networks in response to changes in the traac patterns. Although the ability to dynamically optimize the network under changing traac conditions has been recognized as one of the key features of mul-tiwavelength optical networks, this is the rst in-depth study of the tradeoos involved in carrying out the reconnguration process. We rst identify the degree of load balancing and the number of retunings as two important, albeit connict-ing, objectives in the design of reconnguration policies. We then formulate the problem as a Markovian Decision Process and we develop a systematic and exible framework in which to study reconnguration policies. We apply results from Markov Decision Process theory to obtain optimal reconngu-ration policies for networks of large size. The advantages of optimal policies over a class of threshold-based policies are illustrated through numerical results.