Published Paper Details:

This paper presents a comprehensive intelligent public transportation framework integrating Electric Buses (EBs), Internet of Things (IoT) sensors, Artificial Intelligence (AI), Digital Twin technology, Vehicle-to-Grid (V2G) services, and Renewable Energy Sources (RES) to achieve sustainable urban mobility. The proposed framework introduces a novel Hybrid Multi-Agent Reinforcement Learning and Transformer-Based Optimization Algorithm (MARL-TBO) that simultaneously optimizes dynamic route planning, fleet scheduling, and charging management across a 100-bus urban transit network. A Digital Twin of the entire transit ecosystem provides real-time bidirectional synchronization for predictive decision-making. Battery Degradation-Aware Energy Management considers State of Charge (SOC), State of Health (SOH), thermal dynamics, and passenger load variations to extend battery longevity. AI-based passenger demand forecasting via spatiotemporal deep learning achieves over 95% prediction accuracy. Simulation results over 365 days of operation demonstrate 25-35% reduction in energy consumption, 30-40% reduction in passenger waiting time, 20-30% reduction in operational cost, 15-25% improvement in battery life, and 35-50% reduction in carbon emissions compared to conventional methods. Comparative evaluation against Rule-Based, Genetic Algorithm, PSO, DQN, and PPO baselines confirms MARL-TBO superiority across all performance metrics. The framework establishes a scalable blueprint for next-generation AI-powered electric transit systems.