In recent work a new and widely applicable approach to the theory of large deviations has been developed by Paul Dupuis of Brown University and this investigator. In this approach many aspects of the theory of large deviations are reduced to the theory of weak convergence of probability measures. This research treats new applications of this weak convergence approach to a number of important problems. The method also gives rise to several computational questions that are currently under investigation. Areas of application include a general class of queuing networks and several models in statistical mechanics. The theory of large deviations is an important area of probabilistic research having applications to a number of areas including statistics, statistical mechanics, and queuing theory. In recent work a new and widely applicable approach to the theory of large deviations has been developed by Paul Dupuis of Brown University and this investigator. This approach may give new insight into large deviation phenomena and allow researchers to determine computationally the accuracy of the theory. A main area of application of will be to queuing systems, which arise in the modeling of computer networks.