Models for air-cargo allotment management and booking control.

摘要

Air carriers sell cargo capacity either as an allotment or on an ad-hoc basis. An allotment is a pre-negotiated amount of capacity on specific flights that is reserved for forwarders through medium-term contracts. In contrast, the carrier makes no prior commitment to accept ad-hoc shipments. This dissertation develops mathematical models to analyze the allotment-contract mechanism design problem and the ad-hoc booking control problem.; We study capacity contracts between carriers and forwarders when certain parameters such as the forwarder's demand, freight charges, and operating costs are its private information. Carriers use contracts to allot bulk capacity to forwarders who deliver consolidated loads throughout the course of a season. Carriers also sell capacity to direct-ship customers. We propose contracts with three parameters: a per-flight allotment, a lump-sum payment for the season, and a per-flight payment that depends on the volume tendered by the forwarder. We identify conditions under which relatively simple payment schemes can eliminate informational rents. We also study how the carrier's profit might be affected when the forwarder may not set its booking limit equal to the allotment. We show that under a linear payment scheme, the forwarder's strategic behavior strictly decreases the carrier's profit. However, the carrier's loss can be made arbitrarily small.; We consider the cargo-booking control problem on a single-leg flight with the goal of maximizing expected contribution. Each piece of cargo is endowed with a random volume and a random weight whose precise values are not known until just before the flight is set to depart. We formulate the problem as a Markov decision process (MDP). Because of its high-dimensional state space, an exact solution is not possible. Therefore, we develop six heuristics. The first four are based on different value-function approximations derived from one-dimensional MDPs. The remaining two heuristics are obtained from solving related mathematical-programming problems. We compare the carrier's profit under the various heuristics with that of the first-come-first-served policy as well as with an upper bound. Numerical experiments show that the value-function approximation derived from separate "volume" and "weight" problems offers the best approach and that the carrier's profit under this heuristic is close to the upper bound.