Hybrid Genetic Algorithm-Based Dynamic Route Optimization for Public Transport Efficiency

Public transport efficiency is often hindered by static schedules and unpredictable traffic conditions. This study introduces a Hybrid Genetic Algorithm (HGA)-based optimization model to dynamically adjust bus departure times, reducing transfer waiting times and improving route efficiency. The proposed HGA integrates global search (Genetic Algorithm) with local refinement, ensuring faster convergence and adaptability to transit delays. Experimental results demonstrate that HGA reduces total travel time by $\mathrm{1 2. 6 \%}$ and minimizes passenger waiting time by 18.4% compared to static schedules. The loss curve analysis reveals that HGA achieves optimal convergence within 10 iterations, significantly outperforming Dijkstra's algorithm (0.92s runtime) and A Search (1.16s runtime)* in computational efficiency. Additionally, HGA dynamically adapts routes, processing 11 nodes per iteration, whereas traditional models handle only 2 nodes due to their static nature. By optimizing multiple routes simultaneously and considering realtime constraints, HGA proves to be an effective alternative for large-scale urban transit networks, offering an adaptable solution where real-time data is unavailable. This study provides a scalable and computationally efficient framework for improving public transport scheduling without requiring real-time infrastructure.