基于元模型优化算法的主从博弈多虚拟电厂动态定价和能量管理

A Stackelberg Game Model for Dynamic Pricing and Energy Management of Multiple Virtual Power Plants Using Metamodel-based Optimization Method

随着各社会资本逐步涌入电力市场,未来不同虚拟电厂将属于不同的利益主体,形成多决策主体相互竞争的博弈格局。为兼顾配电网运营商和虚拟电厂的利益,建立运营商和多虚拟电厂的一主多从博弈模型,研究运营商动态定价行为和虚拟电厂能量管理模型,并提出一种基于Kriging元模型的主从博弈均衡算法。该方法用Kriging元模型拟合并代替虚拟电厂能量内部管理模型,在迭代寻优过程中,结合粒子群优化算法生成新的优异采样点,有针对性地修正元模型,从而快速准确地得到虚拟电厂的交易电价和出力计划。该算法无需虚拟电厂的所有参数,不仅达到保护虚拟电厂隐私的目的,而且能够避免对下层优化模型的大量调用,有效减小计算量,提高求解效率。

With different virtual power plants(VPPs)belong to different stakeholders in the future,a competitive game pattern is forming.In order to balance the interests of the distribution system operators(DSO)and VPPs,a one-leader multi-follower game model is constructed,studying the dynamic pricing behavior of DSO and the price response behavior of VPPs.Then a new algorithm for equilibrium solution based on Kriging metamodel is presented.The algorithm fits a Kriging metamodel of transaction price and power to replace the VPP energy management model.The particle swarm optimization algorithm(PSO)is used to generate new excellent sampling points to modify the metamodel in a targeted manner.By avoiding redundant searches,the algorithm identifies the transaction price and scheduling plan of each VPP quickly and accurately.With reduced number of follower-level optimization model evaluations,it also protects the privacy of the VPPs.