Optimal planning of charging station for electric vehicle based on hybrid RODDPSO and K-means algorithm

This research proposes a hybrid approach combining K-means with randomly occurring distributedly delayed particle swarm optimisation (RODDPSO) to strategically locate electric vehicle charging stations (EVCS). The method is structured in three phases: initial regional clustering for demand analysis, refined selection of charging pile locations within these regions, and consolidation into efficient charging stations. The approach enhances the traditional K-means by optimising initial centroids with RODDPSO, mitigating the risk of suboptimal solutions due to local minima. The Yancheng ride-hailing dataset is employed to validate the model, showcasing a significant improvement in utilisation rates and operational efficiency compared to the standard K-means algorithm. The findings underscore the hybrid method's potential to optimise EVCS placement for enhanced service coverage and economic viability.