Ergodic stochastic optimization algorithms for wireless communication and networking

This paper introduces ergodic stochastic optimization (ESO) algorithms to solve resource allocation problems that involve a random state and where optimality criteria are expressed in terms of long term averages. A policy that observes the state and decides on a resource allocation is proposed and shown to almost surely satisfy problem constraints and optimality criteria. Salient features of the ESO algorithm are that it does not require access to the state's probability distribution, that it can handle non-convex constraints in the resource allocation variables, and that convergence to optimal operating points holds almost surely. ESO is applied to determine operating points of an orthogonal frequency division multiplexing broadcast channel that maximize a given rate utility.