| Yield management or more generally revenue management, is the process of generating incremental revenues from existing inventory or capacity through better management of the sale of the good or service. That better management typically involves attention to pricing, segmentation and more active response to consumer demand. This dissertation focusses upon airline seat allocation, one of the levers by which airlines (and other industries) practice revenue management.; A common method of airline seat allocation, expected marginal seat revenue (EMSR), is investigated. Static and advanced static versions of EMSR are compared to dynamic programming for two fare planes. Through a series of simulations, it is determined that dynamic controls are superior to static controls and heuristic approaches are only slightly inferior to true dynamic control.; Given that airlines often reopen fare classes for sale if demand is less than expected, methods are developed for the determination of the probability of cheaper fare classes becoming available in the future. Simulations are performed to determine revenue losses resulting from buyers behaving strategically---i.e. from buyers whom are willing to delay purchase if the probability of a cheaper fare in the future is relatively high. These revenue losses from strategic behaviour can exceed 7%.; Lastly, a generalized model of customer arrivals is developed and applied optimally via dynamic programming. This generalized model is a nonparametric model free of numerous current assumptions. Through proper dynamic programming formulation, computational time for determination of seat allocation decisions is order Plane Capacity4 and grows linearly with the level of disaggregation (number of fare classes and number of decision periods), a significant improvement from earlier formulations that grow exponentially. |