We now construct
a family of cycle-free codes that attain the bound of Theorem 3.6
with equality, for all values of n and k. The construction
is quite simple: as in Section 3.3,
we start with the single-parity-check code
k+1 of dimension k,
and repeat the symbols of
k+1 until a code
of length n
is obtained. It is obvious that the dimension of
is k, and
by Lemma 3.4 this code is cycle-free. The minimum distance of
will depend on the sequence of symbol repetitions.
The idea is to
repeat each symbol in
k+1 equally often and as much as possible.
For example, for n = 13 and k = 3, we obtain the following
parity-check matrix in reduced canonical form (0 bits have been
intentionally replaced with blank spaces to accentuate the structure
of the matrix):
which defines a (13, 3, 6) cycle-free code.
In general, the number of symbols to be repeated is k + 1,
while the number of positions available is n - (k + 1).
Write:
n - (k + 1) = a(k + 1) + b
where a, b are integers, and
0bk. This
decomposition of the number of available positions
means that in our construction exactly k - b + 1 symbols
of
k+1 will be repeated
times, while the remaining b symbols of
k+1
will be repeated a + 1 times. If bk - 1, then
at least two symbols of
k+1 are repeated
exactly a times. Since
k+1 contains
a codeword of weight 2 in every two positions,
the minimum distance of the resulting code
is
If b = k, then only one symbol in
k+1
is repeated a times, while all the other
symbols are repeated a + 1 times. In this
case, the minimum distance of
is
d = 2 + a + (a+1) = 2 + 1 |
(9) |
Notice that b = k if and only if
n + 1 0 mod (k + 1).
Hence it follows from (3.17) and (3.18) that the
code
constructed in this manner
attains the bound of Theorem 3.6 with equality.
Figure 3.6 schematically shows two alternative cycle-free Tanner
graphs for codes resulting from this construction (compare the Tanner graph
in Figure 3.6a with Figure 3.2b).
Figure 3.6:
Two alternative Tanner graphs for optimal cycle-free codes.
|
We point out that although cycle-free codes obtained by
repeating symbols in
k+1 have the highest possible
minimum distance, they are not the only codes with this property. For example, consider the following
parity-check matrix in reduced canonical form:
It is easy to see that this matrix defines a (13, 3, 6) cycle-free
code
, whose distance attains the bound of Theorem 3.6 with equality.
This code was obtained by repeating symbols in a (5, 3, 2) code.
It can be readily verified that
is not equivalent to
the (13, 3, 6) cycle-free code
, defined by the parity-check
matrix in (3.16) and obtained by repeating symbols
in
4. For instance,
contains the all-one codeword,
while
does not.
http://people.bu.edu/trachten