Linear matrix precoding, also known as vectoring, is a well-known technique to mitigate multiuser interference in the downlink digital subscriber line (DSL) transmission. While effective in canceling interference, vectoring does incur major communication overhead and computational overhead in terms of the transmission of idle symbols and precoder-data multiplications at each data frame, resulting in significant energy consumption when the number of lines is large. To facilitate energy efficient transmission, it has been recently proposed (in the G.fast standard) that each data frame is divided into a normal operating interval (NOI) and a discontinuous operating interval (DOI). In the NOI, all lines (or users) are involved in a common vectoring group, which requires a large matrix precoder, whereas in the DOI, the lines are subdivided into multiple small nonoverlapping vectoring subgroups and are transmitted in a time division multiple access manner, requiring small matrix multiplications and, thus, improving the energy efficiency. In this paper, we consider the key dynamic resource allocation problems in downlink DSL: given the real-time demands, determine the optimal transmission scheme: The optimal NOI and DOI size in each data frame as well as the optimal grouping strategy in the DOI, and optimally adjust the transmission scheme. We formulate these optimal dynamic resource allocation problems and propose efficient real-time algorithms to solve them to global optimality. Simulation results are shown to demonstrate the efficiency and the effectiveness of the proposed algorithms.
Bibliographical noteFunding Information:
Manuscript received October 14, 2016; revised March 11, 2017 and May 9, 2017; accepted May 12, 2017. Date of publication June 2, 2017; date of current version June 21, 2017. The associate editor coordinating the review of this manuscript and approving it for publication was Dr. Yue Rong. The work of Z.-Q. Yao is supported in part by NSFC under Grant 61372127 and in part by the Key Discipline of Hunan Province and Key Laboratory of Intelligent Computing & Information Processing, Xiangtan University, Ministry of Education, China. The work of Z.-Q. Luo is supported in part by NSFC under Grant 61571384 and in part by the Leading Talents of Guangdong Province program under Grant 00201510. Parts of this paper have also been presented in Globecom 2016. (Corresponding author: Zhi-Qiang Yao.) N. Zhang is with the School of Mathematical Sciences, Peking University, Beijing 100871, China (e-mail: email@example.com).
© 2017 IEEE.
- discontinuous operation
- dynamic programming
- energy efficiency
- resource allocation