Dynamic Spectrum Management For Digital Subscriber Lines

Digital subscriber line (DSL) technology is still the main way of modern broadband access.However, in the application of DSL technology, different lines indeed interfere with each other,also called crosstalk. And this crosstalk is a major issue in the DSL broadband access. Recently,IT power consumption has been identified as a significant contributor to the global warming, weput forward the concept of Green DSL. Dynamic spectrum management (DSM) is recognized asa key technology for tackling this crosstalk problem. By establishing the right dynamic spectrummanagement model, we can achieve the purpose of the “Green”. However, sub-gradientalgorithm usually used based on the dual decomposition has difficulty in adjusting step lengthwhen solves the dynamic spectrum management model. What’s more, the algorithm’sconvergence rate is slow and even not convergent. The improved algorithm given in this paper,that is, proximal center algorithm not only has the separable, but also do not need to adjust thestep length. In the case, this algorithm has faster convergence rate and better convergenceperformance.The main contents and the innovation points are as follows:1) Establish the dynamic spectrum management model with maximize data rate of theGreen DSL. The general solution of this model is sub-gradient algorithm usually used base onthe dual decomposition has difficulty in adjusting step length and the convergence rate is slow. Inthis paper, we gave a new algorithm which called proximal center algorithm. We use theproximal center algorithm to solve this model and also analyze the algorithm’s convergence.Matlab simulation results validate that the proximal center algorithm has faster convergence rate,compared with sub-gradient algorithm based the dual decomposition.2) Establish the dynamic spectrum management model with minimum the power of GreenDSL. The Green DSL reduced the power consumption as low as possible and does not affect thedata rate. Based on maximized the data rate through power control implement Green DSL, wedeveloped on a new model with the utility function of minimum the power directly. We also usethe proximal center algorithm to solve this problem. By Matlab simulation to validate thealgorithm has the advantage of faster convergence rate than sub-gradient algorithm.3) Matlab simulation to validate the dynamic spectrum management model with maximizedata rate of Green DSL achieved “Green” by control the. Meanwhile we also validate thedynamic spectrum management model with minimum the power of Green DSL achieved“Green”. What’s more, this paper introduces the basic DSL transmission model, analyzes thedynamic spectrum management mechanism of DSL and compares many DSL dynamic spectrum management algorithms.