Record Details     « New Search       Brief Record Full Record MARC Record     Bibliographic Data Control Number UPD-00000179749 Date and Time of Latest Transaction 20090716222425.0 General Information 070712s1999 xx eng Title Statement A study on the use of neural network & fuzzy logic for ATM traffic control / by Il Ho Bae Cataloging Source DML Language Code eng Main Entry - Personal Name Bae, Il Ho Local Call Number LG 995 1999 C65 B34 Physical Description viii, 99 leaves : ill. +1 computer laser optical disc (4 3/4 in.) Summary, Etc. Nowadays, ATM is beginning to play as an important backbone in integrated network environments like B-ISDN. It supports various kinds of data such as ordinary data, image data and voice data. Forthcoming network environments will also have to support yet unknown services because our fast-changing information society demands such network environment and services. These new network environments require a different kind of traffic control mechanism in order to keep pace with the highly dynamic and complex nature of the data produced by the widely different services. The objective of controlling traffic in a network is changing from just aiming for better performance to striking a balance between performance and Quality of Service (QoS). Therefore, new approaches such as those using Neural Network or Fuzzy Logic are intensively being studied in this field. Studies are beginning to show that approaches employing Neural Network and Fuzzy Logic are promising compared to traditional models. In this paper, we tried to assess the promise of combining Neural Network and Fuzzy Logic for ATM traffic control. We combined flexible and knowledge-based control of Fuzzy logic and the learning and generalization capabilities of the Neural Network. To examine the performance of this combined model, we compared the results of three other models: the typical Two-Threshold ATM model, Fuzzy Logic Model, Neurofuzzy with One-Time Learning ATM Model and Neurofuzzy with Continuous Learning ATM model. Within the range of our experiments, the Neurofuzzy with Continuous Learning ATM Model showed around 1% improvement compared to pure Neural Network and Fuzzy Logic ATM models Subject Added Entry - Topical Term Fuzzy logic   Neural network Collection Category FI   UP Textual Physical Form Designator Thesis           Physical Location University of the Philippines Diliman: College of Engineering Library II LG 995 1999 C65 / B34           Digital Copy Not Available          Add to Book Cart  |   Download MARC