Human-centric IoT control: A framework for quantifying the impact of occupant behaviour on energy efficiency in shared offices

The building sector is a major contributor to global energy consumption and greenhouse gas emissions, necessitating innovative solutions for improved energy efficiency. Internet of Things (IoT)-based environment control systems offer promising strategies to optimise energy use; however, their effectiveness is highly dependent on occupant behaviour, which introduces significant variability and uncertainty in energy-saving outcomes. This study addresses a critical research gap by developing an agent-based modelling (ABM) framework that dynamically simulates interactions among occupants, heating, ventilation, and air conditioning (HVAC) systems, lighting, and IoT-based control strategies in shared office spaces. Our framework explicitly integrates behavioural factors—including occupant preferences and social interactions—into energy models to capture the complexity of human-building interactions more realistically. Results from our case study demonstrate that IoT-based control strategies can achieve a 31.71 % reduction in total energy use compared to traditional schedule-based controls, with a 13.27 % reduction observed for occupants exhibiting standard behaviour, while energy-conservative habits show minimal change. In addition, our approach is informed by recent advances in occupant-centric control and reflects a refined methodology that addresses earlier limitations related to static occupancy profiles. These findings underscore the necessity of incorporating dynamic behavioural data into building energy simulations. The insights provided here are valuable for policymakers, facility managers, and building engineers seeking to optimise energy management solutions while maintaining occupant comfort and operational efficiency. Future work will focus on integrating empirical occupant data to further calibrate the model for application across diverse building types. • Sensor-based control reduces energy use by 31.71 % in shared offices. • Occupant actions can change energy savings by 41.85 %. • An agent-based model simulates human–building interactions. • Social dynamics affect smart building system outcomes.