Dynamic peak demand pricing under uncertainty in an agent-based retail energy market

For a transition to a sustainable energy future, smart grids must adapt to the mass introduction of renewable energy sources and their inherent unpredictability. The Power TAC competition is a simulation of distribution grid market dynamics with autonomous retail broker agents. It seeks to reflect real-world scenarios and thus guide policy and business decision-making. In Power TAC, these autonomous agents (”brokers”) trade energy through markets and offer tariff contracts to retail customers who consume and produce energy. By periodic alignment with present real-world and alternative future scenarios, Power TAC utilizes the autonomous agent competition structure to investigate sustainable solutions to electricity supply questions. We explore how alignment activities between the 2014 and 2015 competition years, in particular adding a high volume of retail solar production, made net demand less predictable for brokers. It also made demand more volatile in the 2015 competition, leading to more extreme peak demand events. A principal alignment activity between 2015 and 2016 is the introduction of peak-demand charges for brokers. We design a new peak-demand pricing mechanism that reflects the costs of grid capacity usage, balancing realworld practice against the constraints of the simulation environment. We explore the effects of these changes on broker decisions that account for imbalance and peak demand.