Work

Machine learning-supported simulation and design optimization of acoustic comfort in healthcare waiting areas

Year: 2026

Type: article

Abstract: Acoustic comfort in hospital waiting areas is influenced by architectural configurations and occupant activity, shaping patient experience. This study integrates field measurements with design analysis across six hospitals, three public and three private, capturing variables such as enclosure type, glazing ratio, surface and partition materials, seating density, and acoustic treatment. Concurrently, nearly 600 patients reported perceived acoustic comfort via a five-item Likert questionnaire. Acoustic indicators including RT60, LAeq, L10, L90, and the speech transmission index were collected during standard operating hours. Correlation analysis identified L10 as the strongest positive correlate of comfort within the observed dataset ( r ≈ 0.70), while occupancy ( r ≈ −0.42) and RT60 ( r ≈ −0.36) were negatively associated; LAeq exhibited only a weak correlation ( r ≈ 0.19). The positive association between L10 and acoustic comfort is treated as a context-dependent finding, suggesting that intermittent activity-related sounds and possible masking effects may have contributed to comfort within the observed waiting-room conditions. Predictive models, including multiple linear regression, random forest, XGBoost, and support vector regression, achieved high apparent R 2 (0.80–0.86), though grouped hospital validation showed reduced generalizability. Feature-importance analysis indicated the dominant role of floor and partition materials, followed by L10 and occupancy, in shaping comfort. These results demonstrate that acoustic comfort is driven more by material selection and spatial modulation of crowd-related sound than by average sound levels alone, highlighting opportunities for simulation-guided design optimization in healthcare environments.

Source: International Journal for Simulation and Multidisciplinary Design Optimization

DOI: 10.1051/smdo/2026013

Authors: Negar Sioofy Khoojine, Ata Jahangir Moshayedi, Arash Sioofy Khoojine, Amin Kolahdooz

Institutions: Middle East Technical University, Dongguan University of Technology, Yibin University, De Montfort University

Language: English

Topic: Noise Effects and Management

Subfield: Acoustics and Ultrasonics

Field: Physics and Astronomy

Domain: Physical Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $791