Occupant Behavior in Commercial Buildings: Synthetic Population, Co-Simulation with EnergyPlus and Agent Based Modeling

This technical manual serves as an outline of three separate research projects related to building energy performance modeling of occupant behavior, conducted at the Rutgers Center for Green Building (RCGB): a Synthetic Population project, a Co-Simulation of Synthetic Population with EnergyPlus project, and an Agent Based Modeling project. They all aim at improving interior building design through maximizing comfort, satisfaction, health and productivity of commercial-office building occupants.

The first section of the manual is related to the Synthetic Population project. It first starts by explaining the process of gathering existing data on occupant behavior from RCGB studies, identifying similar datasets from other researchers, aggregating them to a large database and ultimately utilizing them as the basis to create a synthetic population. It then covers a step-by-step instruction of constructing cohorts from the synthethic dataset.

The second section describes the steps involved in the Co-Simulation of Synthetic Population with EnergyPlus project. It includes writing co-simulation codes on obXML format by using a standardized occupant behavior modeling tool, obFMU and co-simulation with EnergyPlus using ExternalInterface.

The last section presents the processes involved in the Agent Based Modeling (ABM) project. It first describes the characteristics of building occupant agents and their interactions with the building and each other. It also covers a step-by-step instruction of setting up and running an ABM model on NetLogo, an ABM modeling tool written in Java.

The increased number of occupant behavior data contributes to construction of a reliable database that is later used to generate the synthetic population, which in turn ensures preserving confidentiality and allows transferability of findings. The process of setting up co-simulation with EnergyPlus using obFMU provides insights into the needs of standardized methods to better integrate multiple modeling tools. The cohort strategy, for example, enables modelers to utilize existing datasets on occupant behaviors for constructing an occupant behavior model. Finally, Agent Based Modeling (ABM) of occupant behavior contributes to construct what-if scenarios prior to retrofitting a building. The ABM approach does not only show adaptive actions made by individual occupants to building environment, but also their interactions among themselves to come up with a collective adaptive action in regards to their environment. The resulting outcomes help inform practice and contribute to the improvement of energy building performace models and simulations for the purpose of optimizing interior comfort and satisfaction.