摘要：The BahnCard problem is an important problem in the realm of online decision making. In its original form, there is one kind of BahnCard associated with a certain price, which upon purchase reduces the ticket price of train journeys for a certain factor over a certain period of time. The problem consists of deciding on which dates BahnCards should be purchased such that the overall cost, that is, BahnCard prices plus (reduced) ticket prices, is minimized without having knowledge about the number and prices of future journeys. In this paper, we extend the problem such that multiple kinds of BahnCards are available for purchase. We provide an optimal offline algorithm, as well as online strategies with provable competitiveness factors. Furthermore, we describe and implement several heuristic online strategies and compare their competitiveness in realistic scenarios.
