Zarrinmehr, Saied (2016-12). A Computational Model for Simulation, Visualization and Evaluation of Mandatory and Optional Building Occupancy Scenarios. Doctoral Dissertation.
Evaluating design decisions is an important factor in a post-positivist design process. Understanding how people move in space is an important part of the evaluation processes. However, making accurate predictions of occupants' movements is a challenge mainly due to the differences between individual occupants, their unique preferences in relation to environmental qualities, the types of scenarios with which they become engaged, and multiple dimensions of the environmental factors that affect occupants' decisions. This study suggests a model to simulate and visualize mandatory occupancy scenarios, which are task-based, and optional occupancy scenarios, which are attraction-based. The impact of environmental qualities is largely overlooked in existing simulation models in both of these scenarios. Existing simulation models for mandatory scenarios are often based on finding shortest or fastest paths and for optional scenarios mainly rely on the field of visibility. The original contribution of the simulation models that this study suggests is simultaneous consideration of environmental qualities, path simplicity, and visibility in addition to desires such as travel time or distance minimization. The integration of these models unlocks new potentials that the individual components do not include. The individual techniques that will be used to develop the occupancy simulation models are validated in the exiting literature experimentally. However, this study does not include conducting field studies to validate the integrated model. If the observed walking trails of humans are provided, the suggested models in this study can be validated through a fine-tuning process that reproduces the observed trails. The simulation results can finally be used for evaluation purposes to help designers at the design phase and facility managers in after design phases to make informed decisions. This study provides a software solution that implements the suggested model to support its feasibility. This software uses a Building Information Model (BIM) to represent the built environment, an Agent-Based Model (ABM) to simulate the occupants, a list of research evidence to encode agent's reactions to the environment, a Discrete Event Simulation (DES) model to represent the tasks in mandatory scenarios, and the field of visibility (isovist) to simulate an occupant's viewshed. In this software, evaluation is a process of data query from the information collected by the agents during the simulations. The data query logic can be set according to the interests of designers or facility managers.