Date of Award
6-14-2026
Thesis Type
PhD
Document Type
Thesis
Degree Name
Doctor of Philosophy (PhD)
Divisions
Faculty of Engineering
Department
Department of Civil Engineering
Institution
Universiti Malaya
Abstract
This research investigates trip generation patterns at fast-food drive-through establishments in selected urban areas in Klang Valley and Kuching, Sarawak. Trip generation refers to the estimation of vehicle trips entering and exiting a site, encompassing both dine-in and drive-through services. Although manuals such as the Trip Generation Handbook by the Institute of Transportation Engineers (ITE) and MyTripGen 2020 provide established reference rates, these frameworks typically rely on fixed predictors particularly gross floor area (GFA) within conventional land-use classifications. To complement these references, this study adopts a contextualised, data-driven approach tailored to fast-food drive-through facilities in Malaysia’s urban environment. Addressing this gap, the study contextualises trip generation modeling through a robust, data-driven approach tailored to Malaysian urban settings. The research highlights that trip generation is influenced by the interplay of spatial, temporal, and operational factors. Spatial factors include land-use type, development intensity, road connectivity, surrounding land-use mix, and site-specific features such as drive-through lane length, entry/exit configurations, and parking layout, all of which directly affect traffic flow. Temporal variations in trip rates are shaped by peak periods, day-of-week differences, and business hours. Data collected from 17 urban drive-through sites across both West and East Malaysia, particularly in Klang Valley and Sarawak, confirm that Gross Floor Area (GFA), Number of Employees (NE), Number of Parking Spaces (NPS), and Drive-Through Lane Length (DTL) are strong predictors of daily trip volumes (R² = 0.9712), with drive-through activity alone accounting for over 78% of total trips. In addition, Machine Learning classifiers were employed to capture non-linear relationships iv and uncover deeper interaction effects. A Random Forest model consisting of 50 decision trees, validated using Leave-One-Out Cross-Validation (LOOCV), achieved a classification accuracy of 61.11%. The analysis identified Number of Employees, Vehicle Waiting Time, and Queue Length as the most influential predictors, emphasizing the dominant impact of internal operational efficiency over generalized land-use metrics. A Decision Tree model achieved a certain accuracy level (55.56%), offering valuable interpretability through clearly defined decision rules, making it particularly useful for regulatory reviews and site-planning decisions. For example, outlets with fewer than 91 employees and limited parking (≤ 37 spaces) were frequently classified as high trip generators, suggesting that streamlined yet efficient operations can still result in significant vehicle demand. To supplement these models, drone-based heatmap analyses were conducted to visualize queue lengths and service times across peak and off-peak periods. Peak congestion was consistently observed during evening hours (6:00–7:00 PM), with queue lengths exceeding 220 feets and service times reaching 5.6 minutes. This underscores the importance of incorporating temporal load factors into operational planning and into the design of drive-through infrastructure. Overall, this study presents a comprehensive and contextualised predictive modeling framework for estimating trip generation at fast-food drive-through establishments. By combining regression models, queue-based service analysis, and interpretable machine learning techniques, it supports a shift from static, location-based trip rates toward dynamic, performance-oriented planning tools. The findings provide actionable insights for urban planners, traffic engineers, and policymakers aiming to optimize site design, mitigate traffic congestion, and promote sustainable development in rapidly urbanizing regions.
Additional Information
Thesis (PhD) - Faculty of Engineering, Universiti Malaya, 2026.
Recommended Citation
Let Hui, Tan, "Developing a contextualised trip generation model for drive-through fast-food restaurants in selected urban areas of Malaysia using predictive analysis" (2026). Student Works (2020-2029). 2024.
https://knova.um.edu.my/student_works_2020s/2024
Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
Included in
Spatial Science Commons, Transportation Engineering Commons, Urban, Community and Regional Planning Commons
Initial
khm