Ant Colony Optimization with Dynamic Pheromones for Electric Vehicle Routing and Charging Decisions

Abstract

The increasing adoption of electric vehicles (EVs) for last-mile delivery requires adapting existing routes designed for internal combustion engine (ICE) vehicles. This study introduces an enhanced Ant Colony System (ACS) that optimizes EV routing by dynamically incorporating state of charge (SOC), charging station proximity, and time constraints. Unlike traditional methods, our approach adjusts pheromone deposition in real time, prioritizing charging stops only when necessary, significantly improving adherence to delivery times. Using real-world delivery data from Paris, our results show that routes under 90 km tend to remain energetically feasible, although intermediate time-window violations may occur due to cumulative charging delays. For longer routes, the need for additional charging stops introduces a risk of delays, but the system effectively manages these constraints to minimize disruption. These results provide fleet operators with a practical decision-support tool to identify which pre-optimized routes can be efficiently adapted to EVs, thus offering a pathway for the integration of electric vehicles into existing logistics without significant operational disruption. Future work will focus on enhancing the system by incorporating real-time traffic updates and charging station availability to further optimize the routing process.