Bug in distance matrix

[verginer]

the HigherOrderNetwork.getDistanceMatrix() has a strange bug.
For small networks it works fine passing all test but for larger node sets i.e. 10 -> matrix 10x10
the result becomes unstable.

For example in one iteration the distance matrix looks like:

[[  0.   2.   1.   2.   2.   1.   1.   2.   2.]
 [  2.   0.   2.   1.  inf   3.   2.  inf   1.]
 [  1.   1.   0.   2.   2.   2.   1.   2.   2.]
 [  1.   2.   1.   1.   3.   2.   2.   3.   3.]
 [  2.   3.   2.   3.   0.   3.   2.   3.   1.]
 [ inf  inf  inf  inf  inf   0.  inf  inf  inf]
 [  1.   2.   1.   1.   1.   2.   1.   1.   1.]
 [  2.   2.   1.   3.  inf   3.   2.   0.   3.]
 [  1.   2.   1.   2.   2.   2.   1.   2.   0.]]

and in the next

[[  0.   2.   1.   2.   2.   1.   1.   2.   2.]
 [  2.   0.   2.   1.   3.   3.   2.   3.   1.]
 [  1.   1.   0.   2.   2.   2.   1.   2.   2.]
 [  1.   2.   1.   1.  inf   2.   2.   3.   3.]
 [  2.  inf   2.  inf   0.   3.   2.   3.   1.]
 [ inf  inf  inf  inf  inf   0.  inf  inf  inf]
 [  1.   2.   1.   1.   1.   2.   1.   1.   1.]
 [  2.   2.   1.  inf  inf   3.   2.   0.   3.]
 [  1.   2.   1.   2.   2.   2.   1.   2.   0.]]

2c2
<  [  2.   0.   2.   1.   3.   3.   2.   3.   1.]
---
>  [  2.   0.   2.   1.  inf   3.   2.  inf   1.]
4,5c4,5
<  [  1.   2.   1.   1.  inf   2.   2.   3.   3.]
<  [  2.  inf   2.  inf   0.   3.   2.   3.   1.]
---
>  [  1.   2.   1.   1.   3.   2.   2.   3.   3.]
>  [  2.   3.   2.   3.   0.   3.   2.   3.   1.]
8c8
<  [  2.   2.   1.  inf  inf   3.   2.   0.   3.]
---
>  [  2.   2.   1.   3.  inf   3.   2.   0.   3.]

Note the position of the inf.

The test to reproduce the bug has been added in 37811ad.

Observation: only at position where there would be a 3 does an inf appear

[verginer]

The order of the for loops was not in the order required by "Floyd-Warshall"