利用乏晰理論之寬頻整合服務數位網路訊務控制策略

Abstract

由於對多媒體服務的迫切需求，寬頻整合服務數位網路已漸受重視。寬頻 整合服務數位網路的訊務特性，與目前所使用的網路有相當大的差異。要 建立各式訊務特性與其服務品質需求之模型是非常困難的。因此類模型將 非常複雜，無法求解，故利用傳統排隊理論的分析法將不再適用於此類問 題。而我們確信模糊邏輯將可以成為解決寬頻整合服務數位網路之訊務控 制問題的一個適當工具。在此篇論文中，我們應用模糊邏輯來解決寬頻整 合服務數位網路之訊務控制問題。首先，我們提出應用一模糊邏輯器來解 決擁塞控制問題之方法。然後再提出整合一模糊允諾控制器與擁塞控制器 ，使其一同運作的策略。利用它來解決允諾控制及擁塞控制的問題。經由 系統模擬，我們發現模糊邏輯在此方面確實有不錯的表現。 Due to the demand for the multimedia services, the broadband integrated service digital network (B-ISDN) has increasingly received high attention. Since the characteristics of B-ISDN are quite different from the currently used network. It will be difficult to model the statistical behaviors of the diverse traffic and their service requirements. The traditional method, based on queueing model analysis, may not be suitable for the traffic control problems because the model will be very complicated and the solution could not be obtained. We believe that the fuzzy logic may become a suitable tool for dealing with the ATM traffic control problems. In this thesis, the fuzzy logic controller (FLC) is used to cope with the traffic control related problems. An fuzzy congestion controller is proposed to cope with the congestion control related problem. Furthermore, a strategy to integrate the operation of an admission controller and a congestion controller is also proposed. According to the simulation results, we can find that the FLCs have a great potential in these areas.