Data networks
Scheduling networks of queues: heavy traffic analysis of a simple open network
Queueing Systems: Theory and Applications
Heavy Traffic Analysis of a Controlled Multiclass Queueing Network via Weak Convergence Methods
SIAM Journal on Control and Optimization
Heavy Traffic Convergence of a Controlled, Multiclass Queueing System
SIAM Journal on Control and Optimization
Introduction to Linear Optimization
Introduction to Linear Optimization
Sequencing and Routing in Multiclass Queueing Networks Part I: Feedback Regulation
SIAM Journal on Control and Optimization
Queueing Systems: Theory and Applications
State space collapse with application to heavy traffic limits for multiclass queueing networks
Queueing Systems: Theory and Applications
Heavy traffic resource pooling in parallel-server systems
Queueing Systems: Theory and Applications
Continuous-Review Tracking Policies for Dynamic Control of Stochastic Networks
Queueing Systems: Theory and Applications
Queueing Systems: Theory and Applications
Coding and control for communication networks
Queueing Systems: Theory and Applications
Hi-index | 0.00 |
As one approach to dynamic scheduling problems for open stochastic processing networks, J.M. Harrison has proposed the use of formal heavy traffic approximations known as Brownian networks. A key step in this approach is a reduction in dimension of a Brownian network, due to Harrison and Van Mieghem [21], in which the “queue length” process is replaced by a “workload” process. In this paper, we establish two properties of these workload processes. Firstly, we derive a formula for the dimension of such processes. For a given Brownian network, this dimension is unique. However, there are infinitely many possible choices for the workload process. Harrison [16] has proposed a “canonical” choice, which reduces the possibilities to a finite number. Our second result provides sufficient conditions for this canonical choice to be valid and for it to yield a non-negative workload process. The assumptions and proofs for our results involve only first-order model parameters.