FYI. I put the dimacs file here: <a href="http://coral.ie.lehigh.edu/~magh/tmp/net.dimacs">http://coral.ie.lehigh.edu/~magh/tmp/net.dimacs</a><br><br>Note, I am using the -double option, since my costs are not integral.<br>
<br><br><br><div class="gmail_quote"><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Hi. I am trying to form a basis to hand a standard LP simplex solver (dual simplex) after running LEMON's network simplex. From my understanding, network simplex forms the basis as a spanning tree, which will have m-1 entries (where m = the number of rows = the number of nodes in network). And the last element of the basis would normally come from a basic "artificial variable" - which seems it can be found by looking for the row with _pi[i] = 0.<br>
<br>I am running against a Dimacs format file. Net simplex solves the problem fine, but at the end, I print out the number of arcs with STATE_TREE (basic) and it is not what I had expected. There are 256 nodes in the network, but only 252 links that are set to basic. <br>
<br>Am I interpreting the basis incorrectly? Or might there be a bug?<br><br>Let me know if you want the dimacs file to repeat the issue I am seeing.<br><br>Thanks,<br>Matt<br>
</blockquote></div><br>