This research is motivated by the operations of a public transit company in Hong Kong. We investigate how real-time information can be utilized in combination with historical data to improve routing and scheduling decisions practically. A dynamic integrated vehicle and crew scheduling problem is studied where travel times are stochastic and time-dependent. The objective is to maximize the route frequencies and mileage to provide good passenger service and simultaneously minimize crew overtime and meal-break delays. To mitigate unexpected delays due to uncertainties in operations, various mathematical models are proposed for revising the schedules in real time under a rolling-horizon framework. Their efficiency and effectiveness are evaluated via simulation using real-world data. The simulation results also identify the potential benefits of revising the schedule dynamically in real time using optimization models. The results show that the proposed approaches can significantly reduce motormen overtime and meal-break delays while maintaining coverage and route frequency requirements.