Heuristic feeder-bus operation strategy considering weather information: a chance-constrained model

In this study, we present a heuristic strategy that makes use of weather information to improve feeder-bus operation. First, the Bayes' rule is adopted to establish the relationship between the weather information and the expected passenger arrival rates. We then develop an optimization model that decides the headway and the vehicle capacity to minimize the weighted sum of passengers' waiting time cost and the feeder bus operating cost. In light of the randomness in passenger demand, a novel chance constraint is introduced to control the overloading risk, thereby guaranteeing a high level of service (LOS). The chance-constrained model is solved efficiently via approximation and equivalence transformation. Based on the model, we propose a feeder-bus operation strategy that heuristically adjusts the headway and capacity according to the weather forecast. It is proved that the heuristic strategy outperforms the traditional fixed strategy, and that the advantage originates from the economies of scale in transit service. Numerical experiments are performed to demonstrate the merits of the proposed heuristic strategy. As observed, compared to the fixed strategy, the proposed heuristic strategy can simultaneously reduce the waiting time cost and the operating cost, thereby achieving a win-win situation. The numerical results also suggest that weather variability and forecast accuracy are two important factors impacting the performance of the heuristic strategy.