Efficient Communication Infrastructures for Distributed Control and Decision Making in Networked Stochastic Systems
Date: July 05 - July 09, 2010
In networked systems, groups of agents achieve certain objectives via interaction at local levels in a distributed manner. the performance of such systems is determined by the communication infrastructure of the network as well as the system dynamics. In this paper we study the interdependence of the communication and collaboration graphs in a networked system, in the context of a coordination control and decision making problem. Each agent has to make a decision on whether to cooperate or not in a group effort. this is modelled by two-person coordination games between neighboring agents. the payoff of each agent is computed as the sum of the agent’s payoffs from each of these games. Each agent’s decision on whether to collaborate or not is based on its personal understanding of its own behavioral tendencies as well as its neighbors’. to account for this fact we endow the agents with a behavioral variable which indicates how risk averse each agent is. The effect of the agents on each other is governed by an influence matrix which is partially derived from the communication graph’s topology. This paper focuses on three major issues: The learning algorithm, the effect of communication network topology on the fast convergence of the scheme, the characterization of the effect of the number of like- minded agents and their well-connectedness as major decisive factors on which equilibrium is attained.