Fixed Optimal and Adaptive Neural Fuzzy Queue-Length-Threshold Control for Capacity Assignment in a Wireless ATM

• Main Authors: Chiang, Hsiang-Yu, 蔣翔宇
• Other Authors: Chung-Ju Chang
• Format: Others
• Language: zh-TW
• Published: 1996
• Online Access: http://ndltd.ncl.edu.tw/handle/26114982188480856305

碩士 === 國立交通大學 === 電信工程研究所 === 84 === Recently, a wireless ATM network is proposedwhich integrate different services to provide person-to-person communicationsin the same air interface and using ATM network as its backbone.For a wireless ATM network, in this paper we study the design of its buffermanagement and capacity assignment strategy.We first propose a fixed optimal queue-length-threshold (QLT) scheduling policyfor the buffer management and capacity assignment. And we show this QLT policy still outscore other policies suchas priority policy in a wireless ATM environment.Owing to the traffic fluctuation in a wireless ATM network system providingmultimedia services, the ATM network had better possess an intelligentQLT controller that can adaptively adjust parameters of QLT.Recently, fuzzy logic system have been widely applied to deal with controlrelated problems for ATM networks. Neural networks, like fuzzy logic controlsystems, have been successfully applied to cope with traffic control related problems. Different from the fixed QLT policy we mentioned above,we propose a neural fuzzy QLT (NFQLT) which can adjust the trafficparameters adaptively according to the status of voice and data traffic.The results show that applying the adaptive NFQLT can decrease thecombined total lost probability than the fixed QLT and other policies.