Routing as a Bayesian Coalition Game in Smart Grid Neighborhood Area Networks: Learning Automata-based approach

Abstract

Routing issues in the existing Smart Grid (SG) literature are focused on Home Area Networks (HANs), Neighborhood Area Networks (NANs), or Wide Area Networks (WANs). Among these, routing in NANs is the most challenging as it entails construction and maintenance of backhaul having various Mesh Routers (MRs). Wireless networks are generally used for communication between backhaul and centralized controller for power distribution. This triggers increased chances of congestion due to scarce resources of available bandwidth and number of channels. Keeping in view of the same, in this paper, we propose a new Efficient Routing Scheme (ERS) as a Bayesian Coalition Game (BCG). The solution strategy integrates the concepts of Learning Automata (LA) in NANs. LA are assumed to be the players in the game, which are deployed at the MRs in NANs. Coalition among the players of the game is scaffolded upon the concepts of Bayesian Networks. Each player in the game is allowed to move from one coalition to another depending upon the payoff function. Corresponding to each move of the player in the game, its action may be rewarded or penalized from the environment. Based upon reward/penalty from the environment, each player updates its action probability vector. The proposed scheme is evaluated with respect to various performance evaluation metrics such as load utilization factor, user satisfaction levels, delay and probability of transmission.