I'm using SBB to solve a convex MINLP problem. I need to set the Nodetable limit. Becuse GAMS say Nodetable limit exceeded.
Use option memnodes to reset to > 10000. How I can fixed it can you help me?
Thanks.
İpek
Use option memnodes to reset to > 10000.
Re: Use option memnodes to reset to > 10000.
You need to create an SBB option file (as documented here https://www.gams.com/latest/docs/UG_Sol ... PTION_FILE). You can create this file even from within GAMS in various ways, e.g.
Hope this helps,
Michael
Code: Select all
$onecho > sbb.opt
memnodes 1e6
$offecho
mymodel.optfile = 1;
option minlp=sbb;
solve mymodel min obj using minlp;
Michael
Re: Use option memnodes to reset to > 10000.
Dear Micheal, thank you so so much.
Can I ask you one more thing. The k29 constraint needs a value of 69.975 for node 0. but 34.357 in solution. I could not find the reason for this.
Can I ask you one more thing. The k29 constraint needs a value of 69.975 for node 0. but 34.357 in solution. I could not find the reason for this.
bussieck wrote: ↑3 years ago You need to create an SBB option file (as documented here https://www.gams.com/latest/docs/UG_Sol ... PTION_FILE). You can create this file even from within GAMS in various ways, e.g.
Hope this helps,Code: Select all
$onecho > sbb.opt memnodes 1e6 $offecho mymodel.optfile = 1; option minlp=sbb; solve mymodel min obj using minlp;
Michael
Re: Use option memnodes to reset to > 10000.
I have no idea what you are asking about. BTW, the model is an MIQCP and can be solved to global optimality in seconds by various global solvers, e.g. SCIP (see attached log).
-Michael
-Michael
Re: Use option memnodes to reset to > 10000.
qe (0) value should be = 69.975. However, it does not give correct initial values.
In the solution(see attached log) qe (0) = 34.357 I didnt understand why?
Re: Use option memnodes to reset to > 10000.
From what you sent (an lst file) it is hard to verify the solution reported by SBB. I see no obvious reason why qe(0) should be 69.975. There are multiple ways to verify that a solver reports a solution correctly, e.g. Examiner (https://www.gams.com/latest/docs/S_EXAMINER.html) or just solve a second time, turn on "option limrow=1e9;" and look at the equation listing if you see infeasibilities.
-Michael
-Michael
Re: Use option memnodes to reset to > 10000.
Dear Micheal, thank you so much for your help. I will try it now.bussieck wrote: ↑3 years ago From what you sent (an lst file) it is hard to verify the solution reported by SBB. I see no obvious reason why qe(0) should be 69.975. There are multiple ways to verify that a solver reports a solution correctly, e.g. Examiner (https://www.gams.com/latest/docs/S_EXAMINER.html) or just solve a second time, turn on "option limrow=1e9;" and look at the equation listing if you see infeasibilities.
-Michael
I did not send it again because the file was attached in the previous message. Now in the attachment again.
Re: Use option memnodes to reset to > 10000.
The model does not help. I won't occupy my machine for >2h to get to your feasible point. You can save a solution into a GDX file (and share that) via the savepoint option and share this.
-Michael
-Michael