Optimization of Public Bus Scheduling using Real-Time Online Information

Due to increasing populations and congested traffics, public transportation arrangements are crucial for smart cities. In this paper, a joint method is proposed to optimize public bus scheduling, using real-time online bus information. Specifically, we first introduce three parameters, i.e., number of buses, number of stops, and dwell time, to be the inputs of the optimization process. Then,
$K$
-means clustering and genetic algorithms are used to optimize collaboratively. Different from traditional genetic algorithms, the proposed method can effectively avoid local optimal results. Besides, we propose two metrics to evaluate the system performance. Real-time online data-based experiments show the effectiveness and robustness of our scheme. As a result, the number of buses and operation efficiency can decrease by 28% to 47% and enhance by 3% to 12%, respectively.