42
filling algorithm (explained in section 3.1).
In [34], Yin and Liu simulate their algorithm and show that it outperforms OFDMTDMA in
terms of achievable rate and outage probability. Their results clearly show the advantages of
OFDMA over OFDMTDMA. By considering the different channel gains seen by different users
at any given time, OFDMA outperforms standard OFDMTDMA. Despite their great results,
their algorithm suffers from having a high computational complexity of
O
(
N
4
). In [36], Rhee and
Cioffi attempt to overcome this drawback by developing an algorithm with a lower
computational complexity. Instead of maximizing the total data rate, [36] uses the RA approach
to maximize the minimum of each user’s data rate.
Rhee and Cioffi allow the variable
c
m,n
(defined by equation E15) to assume any real value in
between 0 and 1, instead of binary 0 or 1. This simple change turns the resource allocation
problem into a typical convex optimization problem. As was explained before, convex
This is the end of the preview. Sign up
to
access the rest of the document.
 Spring '09
 Prof
 Optimization, Computational complexity theory, data rate

Click to edit the document details