International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
|
Volume 187 - Issue 16 |
Published: June 2025 |
Authors: Araf Hasan Jhell |
![]() |
Araf Hasan Jhell . AI-Powered Bus Priority and Scheduling Optimization in Dhaka’s Overloaded Corridors: Leveraging GPS Data with a Gender-Sensitive Approach. International Journal of Computer Applications. 187, 16 (June 2025), 23-28. DOI=10.5120/ijca2025925204
@article{ 10.5120/ijca2025925204, author = { Araf Hasan Jhell }, title = { AI-Powered Bus Priority and Scheduling Optimization in Dhaka’s Overloaded Corridors: Leveraging GPS Data with a Gender-Sensitive Approach }, journal = { International Journal of Computer Applications }, year = { 2025 }, volume = { 187 }, number = { 16 }, pages = { 23-28 }, doi = { 10.5120/ijca2025925204 }, publisher = { Foundation of Computer Science (FCS), NY, USA } }
%0 Journal Article %D 2025 %A Araf Hasan Jhell %T AI-Powered Bus Priority and Scheduling Optimization in Dhaka’s Overloaded Corridors: Leveraging GPS Data with a Gender-Sensitive Approach%T %J International Journal of Computer Applications %V 187 %N 16 %P 23-28 %R 10.5120/ijca2025925204 %I Foundation of Computer Science (FCS), NY, USA
Dhaka, Bangladesh’s capital with over 12 million residents, suffers from severe transit congestion and chronic under-service on its bus network. Crowded buses, long waits, and safety issues disproportionately affect women commuters. This paper proposes an AI-driven framework to optimize bus scheduling and priority in Dhaka’s busiest corridors using simulated historical GPS data and survey-informed gender safety metrics. Five algorithms – reinforcement learning (RL), decision trees, support vector machines (SVM), clustering, and neural networks – are developed and compared. We first generate synthetic multi-day GPS traces for buses on major routes and incorporate a simulated female commuter survey capturing concerns (e.g., harassment risk, overcrowding, waiting time). Each AI model is designed to predict or prescribe schedules that minimize total waiting time and delays while weighting gender-safety factors. Models are trained and tested on 80/20 splits of the data, with performance measured by efficiency metrics (average wait time, on-time rate, average delay) and a Gender Safety Index (GSI) improvement score. Results indicate that the RL-based approach yields the greatest overall efficiency gains, reducing average passenger delay by ~30% over baseline scheduling, while increasing the GSI by ~25%. Decision tree and SVM models provide moderate improvements (15–20% delay reduction) with lower computational cost, while clustering and neural networks achieve 10–15% delay reduction. All AI methods outperform a naive timetable heuristic, and importantly produce schedules that reduce female commuters’ exposure to high-risk conditions (e.g. fewer off-peak long waits). This study demonstrates the promise of leveraging AI on transit GPS data for urban contexts like Dhaka, and highlights that incorporating gender-sensitive objectives can measurably improve equity and safety in bus service.