Memon, Kamran AliZhang, QiButt, Rizwan AslamMohammadani, Khalid HussainFaheem, Muhammadul Ain, NoorTian, FengXin, Xiangjun2021-01-182021-01-1820201943-06471943-0655https://hdl.handle.net/20.500.12573/450This work is supported by Natural National Science Foundation of China (NSFC) under Grants 61727817/61425022/61522501/61605013/61875248/61307086/61475024/61672290/61475094/61675030; in part by the National High Technology 863 Program of China under Grants 2015AA015501 and 2015AA015502, and in part by the Fund of State Key Laboratory of IPOC (BUPT).This study presents a dynamic inter wavelength migration scheme for the optical network units (ONUs) employing linear regression machine learning method to equalize the traffic volume on all the wavelengths in time and wavelength division multiplexed passive optical network (TWDM PON). The proposed traffic-adaptive wavelength and bandwidth assignment (TA-WBA) scheme not only decreases upstream traffic delays but also offers 2.3% and 30% less delay on the wavelengths balancing the excessive load and 7% less upstream bandwidth waste, when evaluated against other load-balancing scheme.enginfo:eu-repo/semantics/openAccessTWDM PONtraffic-adaptiveload balancingITU compliantDWBAPLOAMUS traffic delaybandwidth utilizationarticleVolume: 121