Date of Award
2012
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Electrical Engineering
First Advisor
Bikdash, Marwan U.
Abstract
Network flow control methods have been studied for years to ensure the efficient usage of network resources. By adopting a representation reminiscent of the sliding surface representation, this work explores a highly adaptive network flow control strategy. The rate-based network flow control scheme is investigated through a novel model description in the matrix notation, and guaranteed to be stable under mild assumptions. The control law has a simple form but the proof of robust stability is quite involved via the Lyapunov theory. Simulations demonstrated that the developed adaptive algorithms are robust and effective in maintaining good link price and backlog tracking precision, with both constant and time-varying reference source rate inputs. The robustness and adaptation of the designed control law were also investigated in the presence of uncertain state dynamics and external disturbances. The salient feature of the proposed approach is its simple design procedures, easy real-time implementation and less on-line computations. The proposed algorithm was tested using simple network models as well as using more realistic models such as those developed using the small-world network topologies.
Recommended Citation
Mei, Xiaoyan, "Rate-Based Robust Adaptive Performance Tracking Control Of Network Flows" (2012). Dissertations. 35.
https://digital.library.ncat.edu/dissertations/35