Simplifying Preference Elicitation in Local Energy Markets: Combinatorial Clock Exchange
Researchers fuse combinatorial clock exchange with ML to cut market complexity, converging in just 15 iterations.
A team of researchers, Shobhit Singhal and Lesia Mitridati, has introduced a novel market framework designed to unlock efficient local energy trading. As solar panels, batteries, and other distributed energy resources (DERs) proliferate, the power grid needs new platforms for prosumers (producer-consumers) to trade electricity and grid services. The core challenge is that prosumers have complex, interdependent preferences but limited time and computational resources to navigate intricate auction formats. This new mechanism, accepted for the Power Systems Computation Conference 2026, directly tackles that friction.
Their solution, the Combinatorial Clock Exchange, is a multi-product market that radically simplifies the user experience. Instead of requiring prosumers to forecast prices or submit complex bids, the iterative mechanism only asks them to report their preferred bundle of products at posted prices in each round. Machine learning techniques are integrated to accelerate the price discovery process, guiding the market toward equilibrium. The authors also employ a linear pricing rule to ensure transparency. In numerical simulations, this AI-aided approach demonstrates robust performance, converging to stable, market-clearing prices in approximately 15 clock iterations—a practical timeline for real-world operation.
- Mechanism eliminates need for price forecasting or complex bids, users just pick preferred packages.
- Fuses combinatorial clock exchange with ML for faster convergence, reaching clearing in ~15 iterations.
- Uses linear pricing for transparency, designed for prosumers with solar, batteries, and other DERs.
Why It Matters
This could democratize energy markets, enabling broader participation in grid flexibility and accelerating the renewable transition.