Abstract: Considering the obvious benefit conflict between a power grid and the renewable energy power plants, a grid-power supply short-term optimal scheduling model is developed using the Stackelberg game. A wind-photovoltaic-pumped storage union optimizes its power output strategy via grid electricity prices to maximize its own benefits, while considering the union’s output strategy the grid will adjust its electricity prices dynamically at the objective of minimizing its total cost, so as to achieve a benefit balance. To solve such problems, we develop an improved cuckoo algorithm based on adaptive load changing trend and a dynamic inertia memory strategy, and verify its effectiveness by simulating the Hubei power grid. The results show that the model can adapt to different typical days of power dispatching, and its optimized schemes reduce significantly the rates of wind power curtailment and photovoltaic curtailment by 15.0% - 37.2% and 17.4% - 27.9% respectively, thus promoting the use of renewable energy. In the optimized schemes, the power grid is more stable with the mean square error of its residual load reduced by 0.64% - 2.75%, and the average balance degree is increased by 2.37% - 5.36%, a significant improvement and a more stable benefit distribution.

Key words: Stackelberg game, wind-photovoltaic-pumped storage grid, improved cuckoo algorithm, penalty cost of residual load fluctuation, balance degree