Year: 2025
Type: article
Abstract: In response to China’s national “dual carbon” goals, enhancing energy efficiency and reducing carbon emissions in the building sector has become a critical priority. Occupant behavior (OB) significantly influences building energy consumption, and accurately modeling OB is essential for evaluating and improving building energy performance. However, the stochastic, diverse, and complex characteristics of OB pose substantial challenges for understanding its nature and achieving accurate simulation. This paper presents a systematic review of existing research on OB modeling, outlining four key stages in the modeling process: data collection, feature and causal analysis, model development, and model evaluation. In the first stage, current data collection methods are categorized into in-situ field measurements, surveys, laboratory experiments, and big data approaches, with their respective advantages and limitations discussed. The second stage addresses feature and causal analysis, including temporal patterns, correlation analysis, and causal inference techniques. In the third stage, a comprehensive review of existing OB models is provided, spanning both occupancy and energy-related behaviors. These models range from traditional statistical regression approaches to machine learning-based techniques. Finally, model evaluation methods and the integration of OB models into building performance simulation (BPS) are examined. With evolving demands in the building sector and advances in technology, new challenges and opportunities have emerged for OB modeling. The paper identifies five key behavioral characteristics requiring deeper investigation: adaptability, diversity, flexibility, stochasticity, and immediacy. The critical role of OB characteristics in various engineering applications is also explored, including building thermal resilience design, urban-scale building energy system simulation, demand response, intelligent building control, and future policy-making. In the context of climate change, OB adaptability plays a crucial role in enhancing thermal resilience. In urban-scale energy modeling, accounting for behavioral diversity is essential. As renewable energy penetration increases, OB flexibility and immediacy offer untapped potential for demand response strategies. In smart building control, recognizing the stochastic and dynamic nature of OB is imperative. Furthermore, for effective policy development, behavioral adaptability and diversity must be incorporated. To address these emerging needs, future OB models must evolve beyond traditional frameworks. This review proposes several promising research directions: (1) modeling typical OB patterns, (2) developing high time-resolution models for OB prediction and detection, (3) constructing models to assess OB flexibility and sufficiency, and (4) introducing dynamic models capturing OB variations. Besides, advances in big data technologies, artificial intelligence (AI), and high-performance computing provide a solid foundation for these developments. Big data platforms now enable behavior analysis at both building and urban scales. AI technologies demonstrate great potential in understanding human-environment interactions and simulating complex behaviors. Additionally, improved computational capabilities support the training and deployment of complex models. This paper reviews the current landscape of OB modeling and highlights critical research directions to meet future challenges. By advancing OB modeling frameworks, this work aims to inspire future research and provide technical support to our national strategic goals.
DOI: 10.1360/tb-2025-0044
Language: English
Subfield: Building and Construction
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4409099052