Multi-Objective Power Network Partition: Finding the Pareto Frontier

Abstract

Network partition techniques have been widely applied in power systems. While traditional researches focus on improving the precision and efficiency for a single objective, multi-objective network partition is also of vital importance in reality. The main characteristic of multi-objective network partition is the trade-off between different targets. The Pareto Frontier marks the best trade-off that can be achieved. This paper proposed a framework for realizing Pareto improvements using the combination of several efficient partition techniques based on spectral clustering. The efficacy of the framework is demonstrated by numerical experiments.