Abstract

This research addresses a flexible vehicle scheduling problem considering precedence constraints (FVS-P problem), which is prevalent in many scenic areas worldwide. Each group of tourists in the FVS-P problem comprises a set of visit requests that must be served by shuttle vehicles in a predefined order. A three-indexed integer linear programming model is introduced. Furthermore, an index-reduction strategy is proposed to strengthen the model. A math-heuristic route-segment generation algorithm embedded in a mathematical model is designed to solve the FVS-P problem, particularly in the case of large-sized instances. Experiments on different sizes of near-practical instances validate the mathematical models and the algorithm. The proposed math-heuristic can provide better solutions for large-sized instances in a much shorter time than the models. The trade-off between tourists’ feelings and the cost of the area is investigated to provide further insights for managers.