The integration of services that share resources on a single network has generated an in creasing interest in analyzing models of multiple service, multiple resource (or multi-rate) communication networks. The proposed research develops efficient computation algorithms for analyzing large multi-rate-loss networks. The principal investigator will develop decomposition algorithms that allow the exact analysis of a large class of models. By reducing the analysis to that of a set of subnetworks, the overall dimensionality of the problem and the computational costs are reduced significantly. This enables the exact efficient computation of performance measures (e.g. blocking probabilities) in larger models than hitherto possible. The PI also will develop an approximation method to analyze such models for any size and load. The success of this effort would result in the development of efficient methods for exact and approximate analysis of multi-rate networks of practical size and complexity.
