Exploring agent-based modeling approaches for human-centered energy consumption prediction

To counteract the high energy consumption in buildings, different interventions are used such as improving the envelope, using building systems that are more efficient, and reducing energy demand but not enough is done to address the end users, their activities, and energy use behavior. Productivity is important in office buildings and it can be linked to occupant wellbeing. Also vital is understanding the impact of occupant behavior on energy consumption as people take action to improve their comfort indoors. Human behavior is dynamic and challenging to accurately predict. Energy simulation tools are used in the design and sizing of building systems but they do not adequately capture the occupant-related factors (values, preferences, and behavior). Agent-based modeling (ABM) approaches have been used to represent human behavior in different contexts, they have also been used to model building occupants. This paper examines selected literature on ABM for occupant behavior modeling and it also discusses the practicality of ABM to study occupant-related factors in office buildings. Through this study, recommendations for potential applications of ABM for occupant modeling in buildings are presented. The need for actual data to better represent occupant-related factors was also observed.