From coin-trac at coin-or.org Fri Feb 3 09:23:26 2012 From: coin-trac at coin-or.org (Bonmin) Date: Fri, 03 Feb 2012 14:23:26 -0000 Subject: [Bonmin-tickets] [Bonmin] #39: Emergency mode-Restore phase fail In-Reply-To: <054.20fa6ba3d6b00fa6a18bfb1edfe5a73b@coin-or.org> References: <054.20fa6ba3d6b00fa6a18bfb1edfe5a73b@coin-or.org> Message-ID: <069.94f2e844844772993728a917d4eb84d0@coin-or.org> #39: Emergency mode-Restore phase fail -----------------------+----------------------- Reporter: camil-va | Owner: pbonami Type: defect | Status: closed Priority: major | Component: External Version: stable/0.1 | Resolution: wontfix Keywords: | -----------------------+----------------------- Changes (by pbonami): * status: new => closed * resolution: => wontfix -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Fri Feb 3 09:24:07 2012 From: coin-trac at coin-or.org (Bonmin) Date: Fri, 03 Feb 2012 14:24:07 -0000 Subject: [Bonmin-tickets] [Bonmin] #14: Bonmin breaking with Exit Code 35584 In-Reply-To: <054.b7b76a20d42033af191bb2bb6881df31@coin-or.org> References: <054.b7b76a20d42033af191bb2bb6881df31@coin-or.org> Message-ID: <069.4ef8b1b68b025e53f58fd596e34ada71@coin-or.org> #14: Bonmin breaking with Exit Code 35584 -----------------------+---------------------- Reporter: rjwjames | Owner: pbonami Type: defect | Status: closed Priority: major | Component: Bonmin Version: stable/0.1 | Resolution: wontfix Keywords: | -----------------------+---------------------- Changes (by pbonami): * status: new => closed * resolution: => wontfix -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Fri Feb 3 09:25:06 2012 From: coin-trac at coin-or.org (Bonmin) Date: Fri, 03 Feb 2012 14:25:06 -0000 Subject: [Bonmin-tickets] [Bonmin] #49: Thirdparty mumps - Visual project - missing file In-Reply-To: <053.4b84655f0753489370824731a2f3d339@coin-or.org> References: <053.4b84655f0753489370824731a2f3d339@coin-or.org> Message-ID: <068.b034070ee5e5c915081f9cdb1254ae93@coin-or.org> #49: Thirdparty mumps - Visual project - missing file --------------------+---------------------- Reporter: ycollet | Owner: pbonami Type: defect | Status: closed Priority: minor | Component: Bonmin Version: 1.5 | Resolution: wontfix Keywords: | --------------------+---------------------- Changes (by pbonami): * status: assigned => closed * resolution: => wontfix -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Wed Feb 22 10:13:34 2012 From: coin-trac at coin-or.org (Bonmin) Date: Wed, 22 Feb 2012 15:13:34 -0000 Subject: [Bonmin-tickets] [Bonmin] #45: Problem with creating the Makefiles for Bonmin In-Reply-To: <055.edf155d9b3aaedbc12a90d0f0a3f150b@coin-or.org> References: <055.edf155d9b3aaedbc12a90d0f0a3f150b@coin-or.org> Message-ID: <070.eba51a7a32819cca5199cee2e96a5039@coin-or.org> #45: Problem with creating the Makefiles for Bonmin ----------------------+----------------------- Reporter: GSandoval | Owner: pbonami Type: defect | Status: assigned Priority: major | Component: Bonmin Version: 1.5 | Resolution: Keywords: | ----------------------+----------------------- Comment (by stefan): Can you post your config.log file from the directory where configure failed here? See also instructions [https://projects.coin-or.org/BuildTools/wiki/user- troubleshooting here]. -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Wed Feb 22 10:25:28 2012 From: coin-trac at coin-or.org (Bonmin) Date: Wed, 22 Feb 2012 15:25:28 -0000 Subject: [Bonmin-tickets] [Bonmin] #50: bb.bestBound() gives wrong value when stopped on gaplimit Message-ID: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> #50: bb.bestBound() gives wrong value when stopped on gaplimit -------------------+--------------------- Reporter: stefan | Owner: pbonami Type: defect | Status: new Priority: major | Component: Bonmin Version: 1.5 | Keywords: -------------------+--------------------- Hi, when I run Bonmin (or Couenne) with a positive value for the {{{bonmin.allowable_fraction_gap}}} parameter, and then use {{{bb.bestBound()}}} to retrieve the lower bound at termination, then I get the value of the incumbent solution. If, however, I use {{{bb.model().getBestPossibleObjValue()}}}, then I seem to get a correct bound. The problem seems to be in the code (snippets) {{{ if (model_.status() == 0) if (bestSolution_) { status = TMINLP::SUCCESS; mipStatus_ = FeasibleOptimal; } double Bab::bestBound() { if (mipStatus_ == FeasibleOptimal) return bestObj_; else if (mipStatus_ == ProvenInfeasible) return 1e200; else return bestBound_; } }}} and that **Cbc claims to have solved a model to optimality even if it stopped at a gap limit**. Therefore, in my Cbc interface, I compare model().getBestPossibleObjValue() with model().getObjValue() to see if the gap has been closed. Writing this, makes me think that this is actually a Cbc bug... Stefan -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Tue Feb 28 12:04:56 2012 From: coin-trac at coin-or.org (Bonmin) Date: Tue, 28 Feb 2012 17:04:56 -0000 Subject: [Bonmin-tickets] [Bonmin] #50: bb.bestBound() gives wrong value when stopped on gaplimit In-Reply-To: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> References: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> Message-ID: <067.569dabe40ca74e55d5569d706a15bcaf@coin-or.org> #50: bb.bestBound() gives wrong value when stopped on gaplimit -------------------+---------------------- Reporter: stefan | Owner: pbonami Type: defect | Status: new Priority: major | Component: Bonmin Version: 1.5 | Resolution: Keywords: | -------------------+---------------------- Comment (by pbonami): Hi Stefan, If you happen to have an example with this behavior under your nose. Would you care to tell me what is the value in bestBound_ when the problem is FeasibleOptimal. Wouldn't that always be a good value to return? > Cbc claims to have solved a model to optimality even if it stopped at a gap limit. I guess I understand that behavior Thanks! Pierre -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Tue Feb 28 12:10:54 2012 From: coin-trac at coin-or.org (Bonmin) Date: Tue, 28 Feb 2012 17:10:54 -0000 Subject: [Bonmin-tickets] [Bonmin] #50: bb.bestBound() gives wrong value when stopped on gaplimit In-Reply-To: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> References: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> Message-ID: <067.30ca6b6e7321810fcb34573885fca985@coin-or.org> #50: bb.bestBound() gives wrong value when stopped on gaplimit -------------------+---------------------- Reporter: stefan | Owner: pbonami Type: defect | Status: new Priority: major | Component: Bonmin Version: 1.5 | Resolution: Keywords: | -------------------+---------------------- Comment (by pbonami): Hi again, From looking at the simple example I think that the function bestBound() should always return bestBound_. Would that be OK with you? Best, Pierre -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming From coin-trac at coin-or.org Wed Feb 29 04:56:33 2012 From: coin-trac at coin-or.org (Bonmin) Date: Wed, 29 Feb 2012 09:56:33 -0000 Subject: [Bonmin-tickets] [Bonmin] #50: bb.bestBound() gives wrong value when stopped on gaplimit In-Reply-To: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> References: <052.8e403f6dfa533de97da441b28e46ce1f@coin-or.org> Message-ID: <067.0b2eee3c8cac5f65843ae21bee221c9c@coin-or.org> #50: bb.bestBound() gives wrong value when stopped on gaplimit -------------------+---------------------- Reporter: stefan | Owner: pbonami Type: defect | Status: new Priority: major | Component: Bonmin Version: 1.5 | Resolution: Keywords: | -------------------+---------------------- Comment (by stefan): Hi, yes, the value in bestBound_ looks good. I don't know about the case when the model is infeasible, but if it's feasible, then bestBound_ should contain the right value. Stefan -- Ticket URL: Bonmin Basic Open-source Nonlinear Mixed INteger programming