ATM NETWORKS CONGESTION CONTROL OPTIMIZATION USING STOCHASTIC PROGRAMMING

ABSTRACT.
The Asynchronous Transfer Mode (ATM) is a major transport technology to support the
Broadband Integrated Services Digital Networks (B-ISDN). It supports variety of network
services such as World Wide Web, videoconference, e-commerce, network multimedia,
traditional file transfer protocol (ftp), email, newsgroup and etc. Many network applications
not only ask for high speed, broadband, but also require high quality of services (QoS). More
and more research works are focused on ATM networks. A number of researchers have
demonstrated the virtual path concept. The advantage of this concept is that it allows a large
group of virtual circuits to be handled and switched together, resulting in a lower computing
complexity, faster processing speed, and an efficient use of network resources. The ATM
network provides multiple services, a significant amount of traffic flows have stochastic
characteristics; that makes it difficult to solve congestion control problem.
Stochastic programming (SP) is introduced to solve the essential problems of ATM network
congestion control, Scenario Tracking (ST) approach of stochastic programming employed
for ATM network optimization. Three optimization models: Capacity Assignment (CA),
Capacity and Flow Assignment (CFA), and Flow Assignment (FA) are discussed; a proposed
virtual path based CFA model is then simulated.