Agent-based modelling using survey data to simulate occupancy patterns and occupant interactions for workplace design

Understanding the occupancy pattern and interaction potential afforded by design alternatives is particularly valuable in workplace design, where unplanned face-to-face interactions are valued as sources of innovation and collaboration. Many studies have been conducted to simulate occupants' spatial behaviour, but most abstract out interactions between occupants. This research presents an agent-based model (ABM) for simulating occupancy pattern and interactions within the domain of workplace design with the aim to test different office layouts for space planning using the amount of potential face-to-face interactions as the evaluation matric To enable the higher degree of customisation needed to simulate interactions, an online survey questionnaire was designed along with the ABM to calibrate the model with the survey data to simulate the occupancy behaviour of an existing organisation. The proposed methodology was tested in a case study office. The result of the simulation was compared to other occupancy data gathered from a limited area as an initial assessment to look for resemblance to real-life occupancy. The findings showed that the proposed methodology can capture the heterogeneous behavioural patterns of occupants to create an agent population that reflects an organisation's specific behaviour. They also showed that the methodology can be used to test different interior layouts and seating arrangements to quantitatively compare their influence on potential face-to-face interactions.