Stochastic modeling of electric vehicle infrastructure using queueing-theoretical approach

• Introduces queue-based stochastic modeling as a solution for optimizing electric vehicle (EV) charging management systems. • Presents three different stochastic models and their associated differential-difference equations for EV infrastructure analysis. • Provides insights into using single and multi-server Markov queueing frameworks to address real-time charging challenges. • Offers managerial strategies to enhance resource allocation, service times, and reduce customer waiting periods in EV charging stations. • Analyzes the impact of renewable energy integration on EV charging efficiency and customer satisfaction. Queueing modeling is essential for enhancing battery management technologies, including battery swapping destinations, charging station management, and renewable energy management systems. The present study offers an in-depth analysis of queue-based stochastic modeling for the electric vehicle (EV) industries, extending the previous investigations. This investigation proposes three distinctive stochastic models: a single-server finite capacity queueing model, appropriate for smaller charging stations however restricted by its single-server assumption; a multiple-server finite capacity queueing model, more advantageous for larger stations while still presuming an infinite population; and a finite capacity and finite population queueing model, appropriate for circumstances with a predetermined number of individuals, although slightly more complex to analyze. The corresponding differential-difference equations for each developed model are formulated and examined, taking into account both transient and steady-state behaviors. Closed mathematical expressions of various system performance measurements are demonstrated for each established model to provide a comprehensive statistical overview. Additionally, to highlight the robustness of the finite population queueing model, several illustrations in terms of different graphs and tables are provided and emphasized with the help of an in-depth investigation for comparison purposes. The significant outcomes of the investigation indicate that integrating finite population considerations substantially enhances the accuracy of waiting time predictions, especially during peak demand intervals, in comparison to models that assume infinite populations. This research emphasizes the significance of model selection according to particular application constraints and presents essential managerial insights, such as optimal resource allocation, dynamic pricing strategies, and load balancing techniques, to assist system analysts and decision-makers. Concluding remarks and potential future perspectives are also included, with an emphasis on integrating cutting-edge technologies like machine learning, deep learning, etc. to enhance adaptive management and model prediction.