Dynamic Electric Vehicle Charging Pricing for Load Balancing in Power Distribution Networks based on Collaborative DDPG Agents
Leloko J. Lepolesa, Kayode E. Adetunji, Khmaies Ouahada, Zhenqing Liu, Ling Cheng · 2025
Read Dynamic Electric Vehicle Charging Pricing for Load Balancing in Power Distribution Networks based on Collaborative DDPG Agents by Leloko J. Lepolesa, Kayode E. Adetunji, Khmaies Ouahada, Zhenqing Liu, Ling Cheng (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-06398.txt.
Introduction
Leloko Lepolesa and co-authors set electric vehicle charging prices that shift with grid conditions. Mass charging can destabilize distribution networks never designed for that load. Their strategy pursues three goals at once: shaving peaks, filling valleys, and balancing load between nearby distribution networks.
The team first forecasts network demand from weather history, prices, and vehicle distributions across areas and times of day. Deep reinforcement learning then sets the dynamic price, and three algorithms were compared for that role, with the best one deployed for pricing. Simulations showed better use of the grid at the distribution level, pointing toward healthier operation at larger scale.
Catalog record
| CREATOR | Leloko J. Lepolesa, Kayode E. Adetunji, Khmaies Ouahada, Zhenqing Liu, Ling Cheng |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-06398
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Dynamic Electric Vehicle Charging Pricing for Load Balancing in Power Distribution Networks based on Collaborative DDPG Agents Leloko J. Lepolesa, Kayode E. Adetunji, Khmaies Ouahada, Zhenqing Liu, Ling Cheng Abstract The transition from the Internal Combustion Engine Vehicles (ICEVs) to the Electric Vehicles (EVs) is globally recommended to combat the unfavourable environmental conditions caused by reliance on fossil fuels. However, it has been established that the charging of EVs can destabilize the grid when they penetrate the market in large numbers, especially in grids that were not initially built to handle the load from the charging of EVs. In this work, we present a dynamic EV charging pricing strategy that fulfills the following three objectives: distribution network-level load peak-shaving, valley-filling, and load balancing across distribution networks. Based on historical environmental variables such as temperature, humidity, wind speed, EV charging prices and distribution of vehicles in different areas in different times of the day, we first forecast the distribution network load demand, and then use deep reinforcement learning approach to set the optimal dynamic EV charging price. While most research seeks to achieve load peak-shaving and valley-filling to stabilize the grid, our work goes further into exploring the load-balancing between the distribution networks in the close vicinity to each other. We compare the performance of Deep Deterministic Policy Gradient (DDPG), Soft Actor-Critic (SAC) and Proximal Policy Optimization (PPO) algorithms for this purpose. The best algorithm is used for dymamic EV pricing. Simulation results show an improved utilization of the grid at the distribution network level, leading to the optimal usage of the grid on a larger scale. I Introduction The Electric Vehicles (EVs) are rapidly penetrating the market. The main reason for adopting the EVs in place of Internal‐Combustion Engine (ICE) powered vehicles is to reduce CO2 emissions from end‐use sectors, of which the transport sector is reported to account for one third [ 10 ] . In [ 10 ] , it is further reported that the sales of EVs grew by 55% in 2022, attaining a record high of more than 10 million. In as much as the decarbonisation of the transport sector brings environmental benefits, it is worth considering that many distribution networks were not originally designed to support EV load. As a result, integrating the charging of EVs into these networks introduces an additional load, potentially leading to grid instability [ 30 ] . Researchers in [ 29 ] and [ 23 ] mention the challenges of mass EV penetration into the grid, that include increased peak load demand, increased voltage violations, increased system imbalance due to single phase chargers, increased power losses due to increased rate of consumption, and overloading of transformers and distribution feeders, which in turn will violate their thermal limits, thereby reducing their lifespan. In [ 19 ] , EV charging data was gathered from the utilities, public-charging service providers, and the data logger records in California state. It was combined with travel behaviour data to find the impact of EVs on the distribution grids. It was predicted that 50% of feeders must be upgraded by 2035, and 67% by 2045. In another case study made by Abiassaf and Arkadan [ 1 ] , the authors investigated the impact of EV charging on the distribution grid in Colorado Springs. By utilizing the household data, vehicle technology, passenger travel, and fuel types, they used the Transportation Energy and Mobility Pathway Options (TEMPO) model to predict the impact of EV charging on the distribution network in the next 30 years. They report that by 2050, the impact will be severe. To mitigate this, coordinated EV charging strategies are essential for minimizing their adverse effects on the grid and enhancing its overall efficiency [ 2 ] . To ensure the optimal grid operation and the satisfaction of EV users, careful considerations have to be made on when and where the EVs are charged on the grid. Researchers in [ 25 , 14 , 15 ] conducted studies that show the EV load demand behaviour when the prices are varied in different areas. Better road networks and the usage of smartphones facilitate the change in the EV charging demand in the neighbouring areas when changes in EV charging price are made in one area [ 14 ] . Therefore it is apparent that when the state of charge reaches the threshold required for an EV to be charged, factors that influence the area of charging for the user include the distance to the Electric Vehicle Charging Station (EVCS) and the EV charging price at that station. The existing research on scheduling the charging of electric vehicles for grid stability focuses on load peak-shaving and valley-filling. Upon studying load profiles at the level of distribution networks, it is observed that different distribution networks have different patterns of power consumption. While conventional Time-of-Use (ToU) pricing is intended to improve utilization of the grid, implementing a mechanism to optimize grid usage at the level of the distribution networks can further enhance grid efficiency and stability. If the conventional ToU pricing structure is applied to the charging of EVs, a new electricity demand peak is created when the charging price is lowest in the residential areas, as studied in [ 13 ] . In the commercial areas, authors in [ 9 ] found that the EV load peaks between 12:00 and 18:00 when the conventional load is also at peak, which can be a major issue if the capacity of the distribution network is reached. The integration of EVs into the power grid has the potential to cause problems in the grid if not well managed, hence the scheduling of charging and discharging of EVs has attracted many researchers globally. Researchers in [ 28 ] propose a hub synchronization method that uses dynamic K-means clustering for optimized EV charging. The EV charging data is analysed continuously to recognize the patterns in user conduct, and then predicts the future demand based on the historical data and reallocates EVs to have optimal distribution at the charging locations. However, the proposed method does not show how the users are convinced to charge at the different locations. Authors in [ 3 ] use Multi-Objective Particle Swarm Optimization (MOPSO) to design an optimal placement of EV Charging Infrastructure (EVCI) in a manner that minimizes the power loss and voltage deviation in a distribution network. They further propose a distribution network-level dynamic pricing strategy that can be predicted using Autoregressive Integrated Moving Average (ARIMA), to improve the grid stability, efficiency and revenue. In [ 11 ] , authors formulate the CS attraction function to the EV based on Reilly’s law of retail gravity, and suggest charging prices for each CS based on scenario-based stochastic optimization. While the method improves the CSO revenue and relieves the congestion of idividual charging stations in comparison to the fixed pricing strategy, the distribution networks congestion is not addressed. Das and Kayal [ 7 ] forecasts a day-ahead PV energy and propose a surface estimation (SE) charging and discharging scheduling algorithm to minimize the grid load variation, which outperforms water filling (WF) algorithm interms of computation time. Authors in [ 22 ] presented short-term EV charging load forecasting using Transfer Learning (TL)