Electric vehicles charging station allocation based on load profile forecasting and Dijkstra's algorithm for optimal path planning.

The widespread adoption of Electric Vehicles (EVs) presents new challenges for efficient and timely access to Charging Stations (CSs), particularly under constraints of limited availability and variable demand. The current investigation addresses the EV charging station allocation problem, aiming to guide EVs to optimal CSs based on real-time and forecasted system dynamics. An integrated framework that combines load profile forecasting, optimal path planning, and drone-assisted edge computing is proposed to support decision-making. Specifically, a Nonlinear Auto-Regressive with Exogenous inputs (NARX) model is used to predict future load profiles at CSs, enabling proactive management of charging demand. To determine the most accessible stations, Dijkstra's algorithm for shortest-path computation based on the EV's current location and the locations of the CSs around is applied. Furthermore, drones with lightweight edge computing algorithms enabled real-time data exchange between CSs and EVs, providing up-to-date information on slot availability and local crowd conditions. For the forecasting component, the NARX model has provided a correlation coefficient of 90% for the CS real data collection. Dijkstra's algorithm was employed to effectively optimize the routing of EVs to their nearest charging stations by determining optimal shortest paths. The simulation results demonstrate that the proposed approach significantly enhances EV allocation efficiency while reducing both waiting times and travel distances. Further research is needed to address regulatory and logistical challenges associated with drone deployment in real-time applications.