An Evolutionary Algorithm for Improved Diversity in DSL Spectrum Balancing Solutions

Abstract

There are many spectrum balancing algorithms to combat the deleterious impact of crosstalk interference in digital subscriber lines (DSL) networks. These algorithms aim to find a unique operating point by optimizing the power spectral densities (PSDs) of the modems. Typically, the figure of merit of this optimization is the bit rate, power consumption or margin. This work poses and solves a different problem: instead of providing the solution for one specific operation point, it finds a set of operating points, each one corresponding to a distinct matrix with PSDs. This solution is useful for planning DSL deployment, for example, helping operators to conveniently evaluate their network capabilities and better plan their usage. The proposed method is based on a multiobjective formulation and implemented as an evolutionary genetic algorithm. Simulation results show that this algorithm achieves a better diversity among the operating points with lower computational cost when compared to an alternative approach.

Minimal Delay Violation-Based Cross-Layer Scheduler and Resource Allocation for DSL Networks

“…However, the question of whether or not these power allocations correspond to global optimums of the NUM problem from which they are obtained is now irrelevant with respect to the scheduler's performance. Some literature is available on selecting a representative set of DSL resource allocations in order to achieve good performance: [40] considers full-duplex DSL and constructs a setŜ containing two elements to obtain a performance gain over time-division duplexing, and [41]- [43] employ multiobjective evolutionary algorithms to obtain a larger setŜ containing resource allocations that are -in some sensediverse. We will however take a more heuristic approach towards compilingŜ by solving the following weighted sum rate maximization (WSRM) problem for a predetermined set of weight vectorsŴ:…”