為了符合日前的多元傳輸以及應用,像是聲音傳輸、影音傳輸、高畫質網路電視、及時傳輸、非及時傳輸以及普通資料傳輸,乙太被動式光纖網路 (Ethernet Passive Optical Networks, EPONs)則是被設計來符合這些應用的需求。因此乙太被動式光纖網路需要在內部設計良好的品質服務分級 (Quality-of-Service, QoS)。本篇研究提出了一個以基因表示規畫法(Genetic Expression Programming GEP)實行適應性交錯式動態頻寬分配機制(Interleaved Dynamic Bandwidth Allocation, DBA)。首先,IDBA可以解決EPON上的閒置時間(Idle time)的問題。第二,不同的恤務特性也將在GEP預測中被探討。此外,剩餘頻寬以及保留頻寬將被公正的收集並再分配給下一個集合。模擬結果顯示在封包延遲以及浪費頻寬的一個系統表現。
Ethernet Passive Optical Networks (EPONs) are being designed to deliver multiple services and applications. To support these applications with their diverse requirements, EPONs require Quality-of-Service (QoS) mechanisms to build in. This study proposes an adaptive Interleaved Dynamic Bandwidth Allocation (IDBA) scheme incorporated with a Genetic Expression Programming (GEP) performance for traffic prediction mechanism in EPONs. Firstly, the IDBA can resolve the idle period problem in traditional Dynamic Bandwidth Allocation (DBA) mechanism to decrease bandwidth waste by interleaved transition. Secondly, the traffic characteristic in differentiated services is also considered in the proposed Interleaved Waited-Difference Bandwidth Allocation (IWDBA) and GEP prediction mechanism to provide more accurate prediction. Moreover, the remaining bandwidth will be collected and reallocate fairly in next group ONU. Simulation results show the preferable system performance for the proposed IDBA scheme in terms of packet delay, and the wasted bandwidth.