Stochastic discrete flow networks: Diffusion approximations and bottlenecks (Q1180566)

The congested discrete flow queueing networks, both open and closed, are described in terms of the bottlenecks of the networks. The main results stated as functional central limit theorems provide diffusion approximations for irreducible closed, open networks and networks with priorities. The joint limits for workload process, queue-size process, busy time process and sojourn time process are given. It is proved that the limits are reflected Brownian motions on nonnegative orthant (for open networks) and on the simplex (for closed ones).