Operating peer-to-peer electricity markets under uncertainty via learning-based, distributed optimal control

Towards the global endeavor of clean energy transition, there is a rapid development of distributed energy resources installed in the premises of residential or commercial users, enabling them to act as flexible energy prosumers. Empowering prosumers is envisioned as a catalytic development for modern energy economies, with recent research, as well as innovation and policy actions, pointing to the promising direction of decentralized energy markets, where active energy prosumers exchange energy in a decentralized fashion. Despite the vast amount of recent research on prosumer-centric peer-to-peer (p2p) energy markets, only a small subset of studies accounts for managing the inherent uncertainty of prosumers’ flexible demands.
In this paper, we consider the problem of controlling the decisions of energy prosumers’ within a p2p exchange network. The multi-bilateral economic dispatch is formulated as an optimal control problem. The proposed solution is based on a direct lookahead policy, effectively addressing the issues of dimensionality and local constraint satisfaction. Experimental simulations demonstrate the method’s efficiency and the system’s behavior. The proposed formulation and method is shown to effectively address the operation of p2p markets under uncertainty, closely tracking the performance of the (full information) optimal-in-hindsight benchmark.