Applying genetic algorithms to teletraffic control problems

• Main Author: Lim, Kim Hai.
• Other Authors: Tan, Chee Wah
• Format: Theses and Dissertations
• Language: English
• Published: 2008
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Control engineering
• Online Access: http://hdl.handle.net/10356/13324
• Institution: Nanyang Technological University

Broadband Integrated Service Digital Network (B-ISDN) opens new eras in network communication. It uses Asynchronous Transfer Mode (ATM) technology to transfer data, voice and video from the source to the destination. In order for the transmitted information to reach the destination with the minimum Quality of Service (QoS), it is important that the appropriate amount of network resources be allocated timely. Resource allocation strategies, such as Medium Term Reconfiguration strategy and Genetic Algorithm were studied by various authors using a 10 nodes ring network as the test bed. While GA performs better in achieving lower Worst Call Blocking Probability (WCBP), its CBP for individual data class is inferior to the Medium Term Reconfiguration strategy. By introducing weighting factors in the fitness function of the GA, its CBP can be further improved. In this study, it was discovered that the GA performs 16.5% better in terms of WCBP than the Medium Term Reconfiguration strategy and the CBP for various data classes is comparable with a difference of only 0.6% under 100% traffic fluctuation.