Improving Average Case Performance
n“Optimal Codes” for highest minimum throughput guarantees often have many more codewords than number of nodes in the network
nIdea: carefully select
subset of codewords that are
actually used (previous analysis assumes all codewords are equally likely to be used)
nThe proposed
code-selection procedure chooses
code-words that have mutual maximum-average
Hamming distance.
The code-selection procedure chooses
code-words that have mutual maximum-average Hamming distance. We tested the
performance of this algorithm by computing the sample-average throughput of a
given node using the selected code-words of the singly-extended RS codes that
maximize Eq. (1) subject to (2). The code-selection algorithm is formally described as
follows,
The code-selection procedure chooses
code-words that have mutual maximum-average Hamming distance. We tested the
performance of this algorithm by computing the sample-average throughput of a
given node using the selected code-words of the singly-extended RS codes that
maximize Eq. (1) subject to (2). The code-selection algorithm is formally described as
follows,