Generally, congestion can be defined as follows: A result of a mismatch between the network resources (buffer space, transmission capacity, and so on) and the amount of traffic admitted for transmission. Congestion, if once it occurs, manifests itself in two ways: Time delay will increase substantially in the network, and network throughput may begin decreasing with an increase of the admitted traffic.
In this dissertation, congestion control problem in communication networks is considered as a viewpoint of system-theoretic control problem. First, the state equation is established to represent operation of the target network and congestion controller is designed to prevent congestion in the network. Then, we analyze the overall system at the steady state to derive the network performance. Moreover, the stability of the overall system is investigated analytically. In most of the previous results, queuing theory is used as a basic tool in analyzing the target network. However, in most of communication networks, it is almost impossible to analyze the networks. Thus, the system-theoretic approach based on the averaging theory is an alternative to queuing approach.
Here, we study on congestion control in various types of communication networks that can be divided into two categories: single-hop networks and multi-hop networks. As typical examples of the single-hop networks, token-passing networks and multi-channel frequency-hopping code division multiple access with slotted ALOHA (FH-CDMA-S-ALOHA) random access for packet radio networks are considered in this dissertation. Token-passing networks have been used as a media access control sub-layer of the data link layer (Layer 2 among ISO seven-layer) of MAP (Manufacturing automation protocol). CDMA communication techniques have been used as attractive alternatives to conventional time and frequency division multiple access (TDMA and FDMA).
Specifically, in token-passing networks, we formulate the performance charac...