College of Design – School of Building Construction – PhD Proposal Defense –Qinghao Zeng

Date: December 8th , 2025

Time:  9:30am- 11am  Eastern Time

Location: Microsoft Teams

Meeting Information: 

Meeting ID: 262 739 265 280 74

Passcode: Te7pb9DW

Committee:

• Dr. Pardis Pishdad, School of Building Construction, College of Design, Georgia Tech (advisor and committee chair)

• Dr. Eunhwa Yang, School of Building Construction, College of Design, Georgia Tech

• Dr. Jing Wen, School of Building Construction, College of Design, Georgia Tech

• Dr. Duen Horng Chau (Polo), School of Computational Science and Engineering, College of Computing, Georgia Tech

• Dr. Xinghua Gao, Myers-Lawson School of Construction, College of Engineering, Virginia Tech

Title: Enhancing Human-Building Interaction in Educational Buildings: An Interdisciplinary Agent-based Framework for Optimizing Building Performance, Resilience, and Space Utilization

Abstract

Modern educational buildings, such as university campuses, often face increasingly urgent operational demands driven by stringent sustainability goals, dynamic occupancy levels, and the critical need for robust emergency preparedness. However, conventional building management systems are fundamentally limited to address these dynamic challenges, as static operational efficiency and structural integrity are usually prioritized while dynamic Human-Building Interactions (HBI), occupant perceptions of Indoor Environmental Quality (IEQ), and proactive performance prediction are neglected. As a result, this oversight could lead to suboptimal space utilization, increased vulnerabilities during emergencies, and diminished occupant well-being and productivity. Consequently, to overcome these multifaceted challenges, it requires not only a fundamental rethinking of conventional building management strategies, but also a paradigm shift towards an occupant-centric approach, where responsive and predictive models are used to proactively manage space, energy, and safety.

Therefore, this dissertation proposal aims to develop an interdisciplinary agent-based modeling (ABM) framework, specifically for educational buildings, designed to enhance human-building interaction and thereby optimize building performance, resilience, and space utilization. This framework introduces a multi-objective approach with three major functionalities. First, it conducts a longitudinal analysis of building utilization by modeling the impact of historical and future student enrollment growth on occupancy patterns, providing strategic insights for campus infrastructure planning. Second, it could enhance occupants’ indoor experience and space utilization of the building by acting as a smart recommender system, guiding users to optimal spaces based on personal IEQ preferences and real-time building data. Third, it could enhance building resilience towards emergent scenarios by providing dynamic indoor navigation, calculating the fastest and safest evacuation routes based on occupants’ location and real-time hazard alerts.

This research will be conducted through a structured, mixed-methods approach. The first part of the methodology would be a systematic literature review combined with selective interviews with involved professionals from the Office of Infrastructure and Sustainability at Georgia Tech to establish the theoretical imperative for enhancing space utilization and to synthesize the current practices and challenges of space management at Georgia Tech. Building upon this theoretical foundation, this research will proceed with empirical data collection to provide the essential input data for the Agent-Based model. Specifically, data collection will involve three primary streams: (1) historical Wi-Fi connection logs from OccuSpace sensors to model occupant behavioral patterns (e.g., occupancy rate through different times) that can be scaled to different enrollment periods, (2) IEQ monitoring data (e.g., temperature, humidity) from the Johnson Controls Metasys system to provide objective environmental conditions for the recommendation engine, and (3) Post-Occupancy Evaluation (POE) survey data to construct user profiles that capture subjective comfort preferences. These profiles will quantify a range of IEQ preferences (e.g., for natural light, crowd density, or acoustic privacy) and key behavioral traits such as typical group size, preferred library location, and accessibility needs (e.g., technical resources and infrastructure). The POE dataset will power the recommendation engine, and when combined with real-time occupancy sensor data, enable the generation of personalized space recommendation for users under normal condition and optimal evacuation routes for emergency scenarios. Subsequently, the integrated ABM framework will be validated through a detailed case study of two campus buildings, Price Gilbert library and Crosland Tower library. This validation will involve a multi-faceted simulation designed to quantitatively assess the performance of each objective: the accuracy of the predictive model in forecasting building utilization and performance, the effectiveness of the recommendation system in enhancing space utilization, and the reliability of the navigation engine in improving evacuation safety and efficiency.

This research could make significant theoretical and practical contributions. Theoretically, its primary novelty is a multi-objective optimization framework that holistically balances three often-contradictory goals: building performance, occupant satisfaction, and emergency resilience. This moves beyond conventional studies that typically address these issues in isolation. Methodologically, the novelty lies in creating a high-fidelity simulation framework parameterized with rich empirical data, introducing a method to analyze longitudinal occupancy changes by scaling agent populations, implementing a personalized recommender engine based on empirical POE/IEQ data, and applying ABM to model complex behavioral dynamics for dynamic evacuation.  Practically, it delivers a validated and scalable toolkit for the AECFM industry that integrates predictive, recommendation, and navigation functions into a single, cohesive system. The framework’s innovation is this holistic integration, marking a definitive shift from traditional, reactive building management to a proactive, occupant-centric paradigm that cohesively integrates human-well-being, environmental sustainability, and infrastructural flexibility.

Keywords:

Agent-Based Modeling, Human-Building Interaction, Building Resilience, Occupant-Centric Design, Educational Buildings