Cognitive Agent-Based Life Process Modelling to Predict Social Performance in Workplace Design
DC I/O 2020 poster by PATRIK SCHUMACHER, TYSON HOSMER, ZIMING HE, SOUNGMIN YU, SOBITHA RAVICHANDRAN
Buildings are designed to enable high value human experience while being dynamically occupied for specific individual, interactive, and collaborative human functions, yet architects rarely collect data about how well their design performed from a social functionality perspective. As Frank Duffy said, “Because our heuristic seems to be ‘Never look back’, we are unable to predict the long-term consequences of what we design (Duffy 2008).” We present a cognitive agent-based simulation approach to predicting and evaluating the social performance of workplace design. Our simulation model is part of a larger multi-objective computational design framework for workplaces outlined below where we use it to evaluate social behaviour in workplaces in an iterative data driven process to achieve optimised multi-performance workplace design.The model is composed of three interrelated parts: Agent, Environment, and Data (I/O). Each Agent is an encapsulated object containing a dynamic internal state, visual perception, control mechanisms for path-finding / movement, and an autonomous decision making framework. A heterogeneous crowd of Agents take workplace related actions in relation to a workplace environment with other Agents, interactive destinations, zoning, planned, and unplanned events producing spatially situated social performance Data.
Agent-Based Modelling, Artificial Intelligence, Data-Driven Design, Simulation, [[Behaviour], Computational Design, Generative Design, Parametric Design, Machine Learning, Life Process Modelling.
thirst social wc motivation (100), internal state (60), public destination (60), home destination (50), interaction type (50), action utility factor (47), space occupied time (47), m max occupant (47), decision making (40), social performance (40), meeting table (40)
AI-driven development, Artificial Intelligence in Design, Automated Design Systems, Immersive Workspaces, Optimization, Semiology, Simulation.
Video Presentation: https://youtu.be/EHmg2FWgwdQ
Full text: Maciel, A. (Ed.), 2020. Design Computation Input/Output 2020, 1st ed. Design Computation, London, UK. ISBN: 978-1-83812-940-8, DOI:10.47330/DCIO.2020.QPRF9890