ExCC is a method exact for solving CC problem.
Currently, ExCC is slow (it takes 5-6 hours for a graph with 600 nodes).
Consequently, it is preferable to apply it on graphs up to 100-200 nodes.
TODO: Cplex is used with Open-Jdk. So the jar file called "cplex-partition-openjdk-1_8.jar"
is not compatible with oracle Jdk.

…negs for density=1

1) We include a new variable 'network.no.list' which is
a sequantial integer vector (i.e seq(1,10))
2) when density=1, prop negs is equal to prop extarnal links according to our model hypothesis.
Instead of ignoring prop negs in folder structure, we explicitely put prop negs folder
even though prop negs is changed by only graph size (not even with prop mispl).
This is a decision to make the system generic

…folder structure

Since we generate many graphs with the same parameter set, we updated plot.grap.stats().
Now, it generates the same plots as before, but for all networks (according to network.no.list)
On the other hand, we changed 'out' folder structure: out/networks and out/plots.
Lastly, we found a bug in plot.graph.stats(): legend was buggy when x-axis represents 'prop.negs', we corrected it

…iple of k