The air transportation system in the United States is one of the most complex systems in the world. Projections of increasing air traffic demand in conjunction with limited capacity, that is volatile and affected by exogenous random events, represent a major problem in aviation system management. From a management perspective, it is essential to make efficient use of the available resources and to create mechanisms that will help alleviate the problems of the imbalance between demand and capacity. Air traffic delays are always present and the more air traffic increases the more the delays will increase with very unwanted economic impacts. It is of great interest to study them further in order to be able to more effectively mitigate them. A first step would be to try to predict them under various circumstances. A second step would be to develop various mechanisms that will help in reducing delays in different settings. The scope of this dissertation is to look closer at a threefold approach to the problem of congestion in aviation. The first effort is the prediction of delays and the development of a model that will make these predictions under a wide variety of distributional assumptions. The work presented here is specifically on a continuum approximation using diffusion methods that enables efficient solutions under a wide variety of distributional assumptions. The second part of the work effort presents the design of a parsimonious language of exchange, with accompanying allocation mechanisms that allow carriers and the FAA to work together quickly, in a Collaborative Decision Making environment, to allocate scarce capacity resources and mitigate delays. Finally, because airlines proactively use longer scheduled block times to deal with unexpected delays, the third portion of this dissertation presents the assessment of the monetary benefits due to improvements in predictability as manifested through carriers' scheduled block times.