People flow-driven energy optimization: A computational social systems approach to smart community management

Effective energy management in smart communities requires innovative approaches to balance energy supply and demand, particularly in the presence of dynamic people flow and shared communal spaces. This paper introduces a novel framework that integrates the Social Force Model (SFM), Reinforcement Learning (RL), and Distributionally Robust Optimization (DRO) to optimize energy distribution and promote sustainable consumption behaviors. The SFM predicts people flow patterns across communal spaces, such as EV charging stations, kitchens, and laundry rooms, enabling dynamic and responsive energy allocation. The RL model continuously adapts energy distribution policies based on real-time data, while the DRO framework ensures robustness against uncertainties in demand caused by fluctuating people movement. Furthermore, gamification techniques are incorporated to encourage energy-saving behaviors through incentives, fostering active participation from community members. The proposed framework is tested on a synthesized energy community consisting of 150 residential units and shared facilities with renewable energy generation. Results demonstrate a 28% reduction in peak loads, a 35% improvement in energy savings, and enhanced system robustness under worst-case scenarios compared to baseline methods. This paper highlights the importance of integrating human behavior modeling and advanced optimization techniques to achieve efficient and sustainable energy management in modern smart communities. Key findings underscore the framework’s ability to dynamically allocate resources, reduce energy waste, and promote cooperative energy-saving behaviors, paving the way for future advancements in community-level energy systems.