Review of learning-assisted power system optimization

Abstract

With dramatic breakthroughs in recent years, machine learning is showing\ngreat potential to upgrade the toolbox for power system optimization.\nUnderstanding the strength and limitation of machine learning approaches is\ncrucial to decide when and how to deploy them to boost the optimization\nperformance. This paper pays special attention to the coordination between\nmachine learning approaches and optimization models, and carefully evaluates\nhow such data-driven analysis may improve the rule-based optimization. The\ntypical references are selected and categorized into four groups: the boundary\nparameter improvement, the optimization option selection, the surrogate model,\nand the hybrid model. This taxonomy provides a novel perspective to elaborate\nthe latest research progress and development. We further compare the design\npatterns of different categories, and discuss several key challenges and\nopportunities as well. Deep integration between machine learning approaches and\noptimization models is expected to become the most promising technical trend.\n