Electric vehicle (EV) adoption critically depends on equitable charging infrastructure deployment, requiring balance between economic efficiency and equitable service delivery across communities, yet traditional optimization prioritizes efficiency over spatial equity. This study presents a novel hybrid NSGA-II (Non-dominated Sorting Genetic Algorithm II) approach for charging station placement integrating Traveling Salesperson Problem (TSP) station prioritization with Multiple-Choice Multi-Dimensional Knapsack Problem (MMKP) configuration selection to simultaneously maximize demand coverage and locational fairness measured by Jain’s Fairness Index. Leveraging high-granularity GPS-based GEOTRA Activity Data from Sapporo, Japan, our approach captures realistic charging demand patterns across origin communities traveling to Chuo ward, enabling accurate representation of actual charging opportunities. We employ a two-phase computational architecture: Phase 1 pre-calculates value-cost-fairness metrics for all candidate station-configuration pairs through comprehensive time-slotted demand simulation; Phase 2 optimizes via custom genetic operators preserving chromosome integrity while exploring the complete solution space. Applied across three budget scenarios (JPY 10, 50, and 100 billion), results reveal systematic coverage-fairness trade-offs along Pareto-optimal frontiers. Knee point analysis identifies balanced solutions that sacrifice less than 1% of potential coverage to secure 8-10% fairness improvements, demonstrating that equitable infrastructure deployment requires explicit multi-objective optimization from initial planning stages, providing actionable decision-support tools for municipal planners.

Keywords

Electric vehicle charging infrastructure, multi-objective optimization, hybrid NSGA-II, locational fairness, high-granularity human flow data.