An agent-based simulation of occupancy schedule in office buildings

Occupant behavior is one of the most influential parameters affecting buildings' energy consumption. However, simplifying this parameter in building performance simulation tools can result in discrepancies between actual buildings and simulations. Because occupant behavior differs from individual to individual, the energy performance of a building is significantly related to its occupants. The present study proposes a novel agent-based occupancy simulation model to accurately simulate stochastic occupancy schedules in a real-time scenario. Instead of using sensory or time-use data, it considers occupancy parameters, including occupants' presence, location, and interaction with the building. The main contribution of this study is the proposal of a scalable simulation-based tool for predicting occupancy schedules in office buildings that does not require historical occupancy data. The developed framework is then implemented in an example that determines the stochastic occupancy schedule of an office building. The performance of lighting occupancy sensors is also modeled as an application of the simulated occupancy schedule. The results indicate that installing occupancy sensors in the office building reduced lighting energy consumption by 24-33%. It also appeared that lighting occupancy sensors may be more beneficial in buildings with fewer occupants. The proposed model can be used by researchers and practitioners to better understand the elements of the occupancy schedules in office buildings, as well as to analyze the impact of occupants on buildings' spaces and energy systems.