Farklı Gölgelenen Kanallar Üzerinden Enerji-Verimli Veri İletimi için Geçmiş-temelli Su-Doldurma Algoritması

Abstract

In this paper, we tackle a resource allocation problem over multiple fading channels in wireless networks. This problem is investigated in two ways. First, we consider the problem over the whole multiple fading channels altogether with no power constraint. We look for an optimal solution for this problem by considering an offline waterfillling algorithm. Then, we also propose history-based online waterfilling algorithms for this problem. This online algorithm uses the history partially in order to determine a waterfilling level based on that part of history. Then, the online policy applies this history-based determined waterfilling level to transmit data over the time horizon of the problem. The relative performance of the online and offline policies is evaluated for various types of fading channels (Rayleigh, Rician, Nakagami, Weibull) over various time horizons. The numerical results demonstrate these online waterfilling algorithms shows close performance to offline waterfilling algorithms especially for longer time horizons and by using larger portions of history.