A mean field stochastic game approach to power management
Abstract
We consider large population of interacting mobile devices with individual state dynamics. Each mobile device has its own battery. The battery state is a modeled as a Markov decision process. The mobile devices are coupled via their payoff functions. Each mobile chooses a dynamic power management strategy in order to have a satisfactory long-term payoff. We propose a mean field stochastic game formulation when the number of mobile devices is very large. We establish a connection between finite stochastic game modeling and the evolutionary game modelling at the infinite population. Sufficient conditions for stationary mean field equilibria are given.