[Cbc] [COIN-CBC] Best way to obtain all global minima

[Pete Janes]

Dear all,

I would like to modify CBC so that it can find all solutions to an 
optimization
problem, subject to the condition that solutions are some distance from 
each other,
as there can possibly be infinitely many solutions. I have been trying 
to do this
by modifying the branch and bound code in CbcModel.cpp.

The main idea is to continue the branch and bound procedure even though 
the best
possible solution has converged towards the best solution, then 
identifying the nodes
which contains a feasible point which equals the best solution.

This seems to be simple in theory but I am having problems implementing 
it. There seems
to be certain areas of CBC which I have overlooked. Does anyone have any 
suggestions
on how best to do this?

Kind regards,

Pete Janes

-- 
Mr Pete Janes
Department of Computer Science
ANU College of Engineering and Computer Science
CSIT Building 108, North Rd,
The Australian National University, Canberra ACT 0200 AUSTRALIA
T: +61 2 6125 1451
E: pete.janes at anu.edu.au