Coalition-based energy management of grid-connected smart homes in the day-ahead electricity market

The increasing penetration of residential renewable energy resources (RESs), battery energy storage systems (BESSs), and flexible appliances enables the cooperative operation of smart homes. However, existing peer-to-peer (P2P) and coalition-based energy management approaches often rely on bi-level formulations, computationally complex market-clearing mechanisms, or game-theoretic approaches, which limit scalability under uncertainty. This paper proposes a scalable coalition-based Energy Management System (EMS) in which inter-household energy exchange capacity is explicitly introduced as a tunable control parameter. This mechanism regulates peer-assisted energy sharing, enabling a trade-off between expected cost reduction for the system and performance metrics such as the peak-to-average ratio (PAR). The problem is formulated as a single-level, two-stage stochastic mixed-integer linear programming (MILP) framework that optimizes scheduling under uncertainty in renewable generation, electricity prices, and demand, achieving coordinated operation without the computational burden of traditional structures. Simulation results for a four-household coalition demonstrate a 10.5% reduction in total electricity procurement cost relative to an independent no-exchange baseline, increasing to 14.9% during weekdays, mainly driven by the complementary consumption profiles of working and retired residents. A 20-household survey-based case study shows diminishing marginal benefits beyond moderate exchange capacities, indicating that unconstrained exchange between residents does not necessarily improve aggregate load-profile metrics. Finally, scalability tests with up to 200 households demonstrate the computational tractability of the proposed framework for large-scale residential deployment.