EVNextTrade: Learning-to-Rank-Based Recommendation of Next Charging Nodes for EV-EV Energy Trading

A research team from Queensland University of Technology and Australia's CSIRO has published a breakthrough paper titled 'EVNextTrade: Learning-to-Rank-Based Recommendation of Next Charging Nodes for EV-EV Energy Trading.' The system addresses a critical gap in electric vehicle infrastructure by using machine learning to identify which charging stations are most suitable for peer-to-peer energy trading between EVs during journeys. Unlike previous research focused on transaction management or isolated mobility prediction, EVNextTrade formulates the problem as a learning-to-rank challenge, where each EV decision event is associated with multiple candidate charging locations.

The framework processes a large-scale urban EV mobility dataset containing millions of journey records, enriched with multidimensional features including EV energy levels, trading roles (provider vs. consumer), distance to charging locations, charging speeds, and temporal station popularity. To handle uncertainty from the mobility of both energy providers and consumers, the researchers employed probabilistic relevance refinement to generate graded labels for ranking. They evaluated gradient-boosted learning-to-rank models including LightGBM, XGBoost, and CatBoost, with LightGBM consistently achieving the strongest performance across standard metrics like NDCG@k, Recall@k, and MRR.

The experimental results demonstrate exceptional performance, with LightGBM achieving NDCG@1 of 0.9795 and MRR of 0.9990, indicating near-perfect ranking accuracy for the top recommendations. These results highlight the effectiveness of uncertainty-aware learning-to-rank for charging node recommendation and support improved coordination and matching in decentralized EV-EV energy trading systems. The approach represents a significant advancement toward practical implementation of peer-to-peer energy networks that can alleviate pressure on constrained charging infrastructure while improving supply-side resilience.