Optimal decision method of demand side flexible resources as power system operation reserve based on master–slave game

• Real-time coordination between operators and users reduces reserve costs. • Decentralized interaction allows secure exchange of capacity and price data. • Adaptive optimization improves speed to meet real-time power grid needs. • City-scale tests show lower costs and stronger reliability of power supply. The escalating penetration of renewables and simultaneous source–load uncertainties are critically eroding reserve margins in modern power systems, leaving the latent flexibility of demand-side resources largely untapped. To redress this shortfall, we devise a distribution-network operational-reserve optimisation framework based on the master–slave game and heuristic algorithm is proposed. First, a multi-agent interactive optimization model based on a master–slave game is constructed to fully activate the reserve potential of heterogeneous flexible resources on the demand side. Secondly, real-time scarcity pricing steers users toward consumption rescheduling that dynamically balances economic utility against reserve provision costs. Finally, under the premise of fully considering the multi-agent interest game and data privacy protection, we propose a distributed privacy-preserving decision engine that hybridises an improved grey-wolf optimiser with quadratic programming, which effectively shortens the solution time and improves the convergence speed of the algorithm. The results demonstrate substantial reductions in operating expenditure and marked alleviation of transmission-side reserve stress to ensure the system’s safe operation and reserve capacity constraints, as well as have higher economy and robustness than traditional methods.