Distributed sustainable generation dispatch via evolutionary games

Today´s power grid is provisioned conservatively for rarely occurring demand peaks. These peaks are served by flexible generation systems that are typically costly and have significant carbon footprint. Distributed power sources such as wind turbines and solar panels are sustainable but unreliable as these have inherently variable generation capacities. An effective power dispatch management system is necessary to harness the significant generation potential of these intermittent systems. In this work, a novel scheme is proposed which leverages upon the recent cyber-enablement in the power grid to distributively dispatch a large number of strategically interacting small-scale variable generators. We incorporate evolutionary game theoretic techniques into the formulation of the dispatch strategy as it provides an opportunity to model the aggregate behaviour of tactical agents making inter-dependent decisions and aids with establishing deterministic steady state predictions of the system state. Numerical and theoretical results presented in this work show that the proposed strategy is highly scalable and enables real-time power dispatch of intermittent systems while maintaining low computational overhead.