EV charging scheduling under limited charging constraints by an improve proximal policy optimization algorithm

The rapid growth in the number of electric vehicles (EVs) has revealed critical limitations in existing charging infrastructure: 40 % of public charging stations experience power overload during peak hours, while 35 % remain underutilized during off-peak periods. Current optimization approaches, inc...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Energy (Oxford) Ročník 333; s. 137422
Hlavní autoři: Lin, Mingqiang, Zhong, Ming, Meng, Jinhao, Wang, Wei, Wu, Ji
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Ltd 01.10.2025
Témata:
algorithms
Charging scheduling
cost effectiveness
electricity
energy
Genetic algorithm
Hybrid optimization
infrastructure
issues and policy
New energy vehicles
Proximal strategy optimization algorithm
variance
Proximal strategy optimization algorithm
New energy vehicles
Hybrid optimization
Genetic algorithm
Charging scheduling
ISSN:0360-5442
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The rapid growth in the number of electric vehicles (EVs) has revealed critical limitations in existing charging infrastructure: 40 % of public charging stations experience power overload during peak hours, while 35 % remain underutilized during off-peak periods. Current optimization approaches, including genetic algorithms and standard reinforcement learning techniques, struggle to effectively coordinate user demand and grid stability due to static constraint handling and delayed responses to demand fluctuations. To tackle these issues, this paper proposes an improved Proximal Policy Optimization (PPO) algorithm to optimize EV charging scheduling. The improved PPO model dynamically adjusts the charging schedule while considering both the capacity limitations of charging stations and the time-of-use electricity pricing. Using Monte Carlo simulations to model user charging behavior, the proposed method efficiently allocates charging stations and power resources, thus alleviating the strain on the grid during peak demand and lowering total charging expenses. Compared to traditional methods, includes genetic algorithms, mixed integer linear programming, and standard PPO, our approach achieves a 6.46 % reduction in charging costs, a 7.64 % decrease in peak load variance, and a 24.5 % improvement in convergence speed, demonstrating significant advantages in cost-effectiveness, system stability, and computational efficiency. •The Monte Carlo method is employed to model the dynamic charging patterns and user behavior stochasticity.•The proposed charging coordination strategy synthesizes dynamic electricity pricing with charging station operational constraints, ensuring both grid safety and user demand satisfaction.•This article introduces a new soft clipping mechanism based on the traditional PPO algorithm and explores it together with the fixed clipping mechanism during the training process.•The framework demonstrates superior cost efficiency and load balancing capabilities compared to existing methods, with marked improvement in policy convergence characteristics.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0360-5442
DOI:10.1016/j.energy.2025.137422