Distributed resource allocation over stochastic networks: An application in smart grids

This paper focuses on an important problem involving distributed decision making, the optimal resource allocation in a network of agents. To address this problem, we explore a consensus-based algorithm that does not require any centralized computation. We use stochastic semistability theory to analyze the case when the communication network used by the agents to coordinate their decisions is not completely reliable, i.e., communication links are subject to a random failure process that is modeled by a continuous-time Markov chain. Some results on fault tolerance and optimality of the considered algorithm are provided. Finally, we present an application of the method to the distributed management of the customers of an aggregating entity in a smart grid environment.