Congestion control for the communication networks with random parameter jumps

This paper is concerned with the problem of congestion control for the class of communication networks with random parameters. The linear dynamic model of TCP network in congestion avoidance mode is considered which contain round trip delays in both state and input. The link capacity, round trip time delay and the number of TCP sessions is assumed to be random finite state Markov process. Active Queue Management (AQM) used to adjust the queue level of the congested link to a fixed value. A dynamic output feedback controller with mode dependent parameters is synthesized to stochastically stabilize the TCP/AQM dynamics. The control synthesis is implemented by solving a linear matrix inequality (LMI). The results are tested within a simulation example and the effectiveness of the proposed design method is verified.