Game Theoretical Issues in Optical Networks

In this paper we focus on the problem in optical networks in which selfish or non-cooperative users can configure their communications so as to minimize the cost paid for the service. Such a cost depends on the personal configuration and on the one of the other users. During a series of time steps, at each of which only one user can move to a better configuration, a Nash equilibrium is eventually reached, that is a situation in which no user can select an improved solution and thus is interested in further modifications. In such a setting, the network provider must determine suitable payment functions covering the network costs that induce Nash equilibria with the best possible global performances. We first present results in the classical scenario in which we are interested in optimizing the optical spectrum, that is in minimizing the total number of used wavelengths. We then outline possible settings in which the approach can be eventually applied to minimize the cost of optical routing due to specific hardware components such as ADMs or filters, that are typical examples of expensive elements whose price can be shared among different lightpaths under specific constraints.