Optimised Scheduling for Distribution Networks, Microgrids and Demand-Side Using Multi-Level Game Theory

Abstract

This paper introduces a multi-level leader-follower game scheduling method, which accounts for the coordinated interests of multiple stakeholders in the integrated operation of distribution networks (DNs), microgrids (MGs), and flexible loads (FLs) systems. Firstly, to address the challenges posed by significant non-linearity and computational burden in the DN game model, optimisation efficiency is enhanced from two perspectives. An improved Bayesian optimisation algorithm is implemented to reduce the number of optimisation iterations for electricity prices, while a graph theory-based forward-backward sweep method is employed to enhance power flow calculation efficiency. Secondly, distinct optimisation strategies are applied to MGs with varying operational profiles: the internet data centre MGs implement the differential evolution algorithm, while the time-shift load MG employs the covariance matrix adaptation evolution strategy, striking a balance between computational speed and solution accuracy for the MG cluster. Moreover, the demand response load layer is modelled and solved as a linear optimisation problem. Finally, based on the IEEE-33 bus test case analysis, the proposed model fully integrates photovoltaic energy, resulting in a 24.5% improvement in the system's overall economic performance and increasing the benefits for each stakeholder.