A two-layer optimization model for electric vehicle charging station distribution using a custom genetic algorithm: Application to Montenegro

To minimize the negative impact of internal combustion vehicles, it is necessary to develop the supporting infrastructure and facilitate the widespread adoption of electric vehicles. A key aspect of this transition is the strategic planning and optimization of the locations of electric vehicle charging stations. The distribution of charging stations is influenced by various factors, including drivers habits, population density, road infrastructure etc. This paper aims to determine the optimal geographical locations and capacities of electric vehicle charging stations by balancing these factors along with budget constraints. A two-stage optimization algorithm is developed using graph-modeled data. The first stage focuses on preliminary planning, considering consumer satisfaction, traffic, electrical grid infrastructure, and construction costs. The second stage refines the optimization using simulation-generated statistics, including waiting times, system utilization and charging stations availability. Novel custom mutation and crossover operations are introduced in the genetic algorithm to enhance the optimization process. Proposed modified genetic algorithm outperforms the standard genetic algorithm by providing solutions that improve the utilization rate while simultaneously enhancing user satisfaction. Notably, the genetic algorithm modification introduced here is not limited to electric vehicle charging station optimization; it can be applied to any problem requiring the optimization of spatial distribution.