A HYBRID GENETIC-GRADIENT DESCENT ALGORITHM FOR OPTIMAL ELECTRIC VEHICLE CHARGING STATION PLACEMENT

The strategic deployment of charging infrastructure is important to accelerate the adoption of electric vehicle (EV) and reduce transportation emissions. However, optimal charging station placement presents a complex optimization challenge, constrained by multiple factors such as construction costs and user accessibility. Traditional optimization methods often struggle to find globally optimal solutions within this multi-dimensional constraint space. To address these challenges, we propose a novel hybrid optimization framework that integrates genetic algorithms (GA) with gradient descent (GD) methods for charging station location planning. Our approach uses GA to generate promising initial solutions, followed by gradient-based optimization for solution refinement. The methodology incorporates three variants of gradient descent, including adaptive, conditional, and proximal gradient. We evaluate our framework through comprehensive simulations across various scenarios, using a carefully designed virtual environment that models realistic user demand patterns and geographical constraints. The simulation results demonstrate the effectiveness and robustness of our proposed hybrid optimization framework for optimal charging station placement.