Congestion and delay problems at airports have received much attention in recent years because of the unbalanced condition between demand and supply. Recent demand forecasts indicate that the problems are expected to increase in the next decade. Relieving congestion of the air transportation networks requires several strategies to enhance the runway capacity. Among these strategies is reducing the runway occupancy time a critical factor in affecting runway capacity. And one approach to reducing the runway occupancy time (ROT) is locating the high speed exits optimally. In addressing the reduction of the runway occupancy time, a full information on the distribution of aircraft landing distance is required. The landing performance at a specific airport may be found by observing the actual landings. However, this is costly and may not be transferable to other airports. An alternative approach is to use a simulation model. A simulation model was built at Center for Transportation Research at Virginia Tech based on point mass kinematics in the flying phase over runway and the ground roll phase on runway to predict the landing roll distance and time to a specified exit speed. Many influencing parameters were incorporated into the model, and then were calibrated using the field data obtained from real operations. The prediction of a nominal landing roll distance and time to decelerate to a specified exit speed is not sufficient for estimating ROT because the additional time to reach a designated exit should be taken into account. To compute the additional time, a braking adjustment scheme is selected from several alternative schemes. The combination of the selected braking adjustment scheme and the simulation model approximates very closely the observed ROT. An optimization model is formulated to determine the exit locations so as to minimize the weighted average ROT of the defined aircraft mix. A polynomial-time solution algorithm is developed for this model using Dynamic Programming technique. The major input parameters for the model are the distribution of the landing roll distance to the specified exit speed and the information on the aircraft mix. The model structured to address the problem of designing a new runway as well as the problem of improving an existing runway. A runway capacity model is used to convert the optimized ROT into capacity gains. Four scenarios are analyzed. Among the scenarios, one is based on the present Air Traffic Control procedures, and three are based on the future developments. The capacity analysis reveals that the ROT does not affect the runway capacity for landing operations. However, the ROT is found as a critical factor for the runway capacity for mixed operations. Hence, the ROT should be optimized for the current system and more crucially for the future developments. The capacity gains by optimizing the ROT under the current Air Traffic Control systems and standards are estimated 2 to 7 more operations per hour. These gains will increase to 20 more operations per hour in the future environment. / Ph. D.