<div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div class="im"><br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Kipp - the example you sent finds the optimal solution after a few <br>
</blockquote>
<br></div>
How does it know the solution is optimal, if integrality does not imply optimality? See me example below.</blockquote><div><br><br>Ahh... this is where the DecompApp::APPisUserFeasible<span class="p"> function comes in.</span> <br>
<br>By default, DIP assumes, that if problem is LP feasible to the linear system and IP feasible, then it is feasible. In the case where the user knows something that DIP does not (e.g., that the linear system does not define the entire valid constraint system, as in TSP), then they must provide a derivation of this function APPisUserFeasible. Then, DIP will check LP feasible, IP feasible and lastly, APPisUserFeasible before declaring a point a feasible solution. <br>
<br>See TSP for an example of its usage:<br> <a href="https://projects.coin-or.org/Dip/browser/trunk/Dip/examples/TSP/TSP_DecompApp.cpp">https://projects.coin-or.org/Dip/browser/trunk/Dip/examples/TSP/TSP_DecompApp.cpp</a><br>
<br><br></div></div><br>