Abstract
The deployment of charging stations is critical for promoting the adoption of electric vehicles. This planning problem involves complex coupling among conflicting factors, including geographical distribution, power grid constraints, construction costs, and user satisfaction, posing significant challenges to existing optimization algorithms. To address this problem, a matrix-based evolutionary algorithm with local search (MEA-LS) is proposed. The proposed algorithm consists of three core components. First, a dual encoding strategy is designed for station location and sizing, accompanied by a heuristic initialization strategy that generates high-quality initial solutions based on demand distribution. Second, a matrix-based crossover and mutation is proposed, which reformulates traditional evolutionary operations through matrix calculations to achieve mathematical formulation and parallel execution. Third, a local search is introduced to refine the capacity configuration of active stations, thereby improving solution precision. These components synergize to accelerate convergence while preserving population diversity. Experimental results on multiple benchmark instances demonstrate that MEA-LS significantly outperforms state-of-the-art algorithms in terms of both solution quality and computational efficiency.