Jia-Rui Lin

Jia-Rui Lin

Associate Professor

Generative Model-Based Building Evacuation Simulation for Safety Design

Journal of Building Engineering, 2025

Recommended citation: Han, J., Zheng, Z., Gu, Y., Lin, J.R.*, Lu, X.Z. (2025). Generative Model-Based Building Evacuation Simulation for Safety Design. Journal of Building Engineering, 116, 114644. doi: 10.1016/j.jobe.2025.114644 http://doi.org/10.1016/j.jobe.2025.114644 cited by count

Abstract

Evacuation simulation is essential for building safety design, ensuring properly planned evacuation routes. However, traditional evacuation simulation relies heavily on refined modeling with extensive parameters, making it challenging to adopt such methods in a rapid design iteration process in the early stages. Thus, this study proposes DiffEvac, a novel method to learn building evacuation patterns based on Generative Models (GMs), for efficient evacuation simulation and enhanced safety design. Initially, a dataset of 399 diverse functional layouts and corresponding evacuation heatmaps of buildings was established. Then, a decoupled feature representation is proposed to embed physical features like layouts and occupant density for GMs. Finally, a diffusion model based on image prompts is proposed to learn evacuation patterns from simulated evacuation heatmaps. Compared to existing research using Conditional GANs with RGB representation, DiffEvac achieves up to a 37.6 % improvement in SSIM, 142 % in PSNR, and delivers results 16 times faster—cutting simulation time to 2 min. Case studies further demonstrate that the proposed method not only significantly enhances the rapid design iteration and adjustment process with efficient evacuation simulation but also offers new insights and technical pathways for future safety optimization in intelligent building design. The research implication is that the approach lowers the modeling burden, enables large-scale what-if exploration, and facilitates coupling with multi-objective design tools.

graphical abstract

Download paper here

Download preprint here

The authors are grateful for the financial support received from the National Natural Science Foundation of China (Nos. 52238011, 52378306).

Financial Sources:

Leave a Comment

You May Also Enjoy

A Text Classification-based Approach for Evaluating and Enhancing the Machine Interpretability of Building Codes

Published in Engineering Applications of Artificial Intelligence, 2023

This study proposes a text classification approach based on a pretrained large language model, to identify computer-processible clauses and to evaluate the machine interpretability of building codes. Result show that the proposed algorithm outperforms SOTA methods with an F1-score of 93.6%, and could further improve downstream rule interpretation algorithm by 4%. It is also revealed that the overall interpretability of building codes is only 34.4%.

A Multi-Factor-Fusion Framework for Efficient Prediction of Pedestrian-level Wind Environment Based on Deep Learning

Published in IEEE Access, 2025

This study proposes a novel Multi-Factor-Fusion (MFF) framework that leverages deep learning techniques. This framework integrates Graph Convolutional Networks and Long Short-Term Memory networks to extract and fuse multiple factors and create an end-to-end neural network model capable of directly predicting wind fields, which, achieves a 10-minute prediction error of 3.52% and a 1-hour error of 10.23% in case studies.