Decentralized self-discipline scheduling strategy for multi-microgrids based on virtual leader agents

A decentralized self-discipline scheduling strategy based on virtual leader agents is proposed for the coordination of microgrids and distributed generations, which has incorporated the optimization for tie-line power, distributed consensus λ-iteration algorithm and reference tracking under the framework of multi-agent system. Using synchronous ADMM algorithm, the sub-optimization problems coupled with boundary constraints are transformed into decoupled sub-optimization problems. With the help of virtual leaders, both the scheduling of distributed generations and that of interconnected microgrids, are integrated. That is, it can infinitely approach to global optimization obtained by traditional centralized algorithm. On the other hand, it can avoid the sharing of participants' private information and the need for a control center with a dedicated wide-area communication network. Finally, an actual study case with 5 microgrids verifies the effectiveness and robustness under various uncertainties of the proposed scheduling strategy.