Agent-Based Model and Empirical Data Integration in Architectural Design Methodology for Corporate Offices

Abstract

This abstract outlines an ongoing research that uses an agent-based crowd simulation model to evaluate the social performance of a spatial design for corporate office environments within the field of architecture. Instead of focusing on the flow of movement which is the common approach in crowd simulation tools, the research described here aims to simulate occupancy behavior. The outlined methodology focuses on model initialization to calibrate the attributes and behavioral rules of the agents with empirical data for an agent based model with utility based decision making framework. The empirical data is gathered through a survey, CCTV camera footages and space observation in a case study, and the data gathering method is designed to capture organization specific occupancy patterns to inform the agent based model. Agents’ behavioral rules also consider the findings from research precedents in workplace communication patterns in the field of social science and behavioral and policy sciences. Using the agent-based model, probabilities for social interactions are compared in different design alternatives to evaluate spatial configurations. The initial set of experiments revealed nonlinear relationships between different aspects of design elements and frequency of social encounters.