151. A family of multi-path congestion control algorithms with global stability and delay robustness.
- Author
-
Feng, Wei-Jie, Wang, Lei, and Wang, Qing-Guo
- Subjects
- *
CONGESTION pricing , *ALGORITHMS , *TIME delay systems , *ROBUST control , *STABILITY theory , *UTILITY functions - Abstract
The goal of traffic management is to efficiently allocate network resources via adjustment of source transmission rates and routes selection. Mathematically, it aims to solve a traditional utility maximization problem in a fair and distributed manner. In this paper, we first develop a generalized multi-path utility maximization problem which features a weighted average of the classical Kelly’s formulation and the Voice’s model. Next, we design from this broader framework a family of multi-path dual congestion control algorithms whose equilibrium point can both achieve a desired bandwidth utilization and preserve a notion of fairness among competing users. Global stability can be guaranteed for the proposed schemes in the absence of delays by use of a totally novel Lyapunov function. Moreover, when heterogeneous propagation delays are taken into account, we establish decentralized and scalable sufficient conditions for robust global stability by constructing a reasonable Lyapunov–Krasovskii functional candidate. These conditions give estimates for the maximum admissible delays that the controller can tolerate without losing stability. Finally, we verify the results through simulation. [ABSTRACT FROM AUTHOR]
- Published
- 2014
- Full Text
- View/download PDF