Structural Results for Decentralized Stochastic Control with a Word-of-Mouth Communication

Abstract

In this paper, we analyze a network of agents that communicate through the ``word of mouth," in which, every agent communicates only with its neighbors. We introduce the prescription approach, present some of its properties and show that it leads to a new information state. We also state preliminary structural results for optimal control strategies in systems that evolve using word-of-mouth communication. The proposed approach can be generalized to analyze several decentralized systems.