Modeling Cooperative Behavior in Smart Grid and Cognitive Radio Systems

Abstract

Cooperation is an instinctive and evolutionary trait in humans that results in mutual benefits for all persons involved irrespective of whether we live in a hunter-gatherer society or a digital economy. Similarly, independent rational players in complex multi-user intelligent systems such as smart grids (SG) and cognitive radios (CR) also benefit from cooperative behavior. Quantifying and sharing the benefits of cooperation amongst all players in a fair and stable manner is a non-trivial problem of great interest. The focus of this thesis is on modeling cooperative interactions for achieving demand side management (DSM) in SGs and dynamic spectrum access (DSA) in CRs using cooperative game theory. A critical challenge in electricity delivery is that energy supply does not follow consumer demand with typical peaks and valleys at different time periods. The integration of distributed renewable energy sources (RES) and information technology in SGs allows for overcoming this mismatch by means of DSM. DSM involves modification of consumer energy demand through price-based demand response (DR) models or by means of local energy markets. Highly efficient DR algorithms are proposed for cost minimization and load balancing for households with energy storage systems (ESS) under time of use pricing. Using concepts from consumer theory and intertemporal trading, cost minimization is formulated as a linear programming problem, while load balancing is formulated as a geometric programming problem. The proposed load balancing method performs very well with peak to average ratio (PAR) values close to 1. Local energy trading is modeled separately as an exchange economy for households with ESS and as a production economy for minigrids with hybrid RES. Due to continuous increase in spectrum demand, certain bands face severe scarcity and yet, a large portion of spectrum is often under-utilized across time and space. Apparent scarcity in spectrum arises from rigid and inefficient spectrum allocation policy rather than actual physical shortage of spectrum. DSA facilitates flexible spectrum usage by providing the capability for unlicensed secondary users (SUs) to sense the spectrum and opportunistically share unused licensed bands without causing harmful interference to licensed primary users (PUs). A cooperative game for jointly modeling spectrum sensing and sharing problem in CRs is proposed, whereby idle spectrum is allocated to SUs based on their sensing performance. The characteristic function that forms the basis for fair division of benefits of cooperation among SUs is derived. The proposed cooperative game for joint spectrum sensing and sharing results in the formation of a grand coalition and provides the best balance between fairness, cooperation and performance in terms of data rate achieved by SUs.