µ-MaxWeight queueing network control with application to in-flight entertainment systems
Although having a multitude of practical applications, the control of large queueing networks is a notoriously difficult problem. It becomes especially challenging when, besides guaranteeing stability, a given cost metric shall be minimized. Recently, an interesting new policy design framework for the control problem, called h-MaxWeight, has been proposed which is a natural generalization of the famous MaxWeight policy. Stability of the policy is achieved through a perturbation technique, however, the stability crucially depends on the choice of parameters which has to be adapted in simulations. In this paper we use a different perturbation technique where the required properties are much easier to implement. This leads to a novel control framework, called µ-MaxWeight, which guarantees universal stability while still operating 'close' to the underlying cost criterion. To illustrate the approach we focus on the example of an in-flight entertainment system which has both wired and wireless components as well as queueing constraints, including buffer underflow constraints. We determine a suitable cost function for this scenario and perform simulations, which suggest that the new approach to policy synthesis can provide significantly higher gains irrespective of any further assumptions on the network model or parameter choice.