solutiontarget cplex

Archiver · May 21, 2015, 2:35pm

Hello,
I want to solve the linear relaxation of a mixed integer second-order conic problem with CPLEX.
So, the model type is rmiqcp; for letting it solve with CPLEX regardless the non-convexity, i should set the solution target parameter to 3. So, this is what I did in the file cplex.opt.
But the value of the parameters seems not recognised.

These are the options and solve statements:

option rmiqcp=cplex;
garintmiqcp.optfile=1;
Solve garintmiqcp min obj using rmiqcp;

This is the log:

Cplex 12.6.1.0

Reading parameter(s) from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”

*comment

tilim 3600
solutiontarget 3
threads 1
parallelmode 1
Finished reading from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”
Reading data…
Starting Cplex…
Space for names approximately 0.01 Mb
Use option ‘names no’ to turn use of names off
CPLEX Error 5002: Q in ‘QCP_row_for_constraint1qcp(1)’ is not positive semi-definite.
Presolve time = 0.00 sec. (0.00 ticks)
QCP status(0):
Cplex Time: 0.00sec (det. 0.00 ticks)
*** CPLEX Error 5002: Q in %s is not positive semi-definite.
Try Cplex option SolutionTarget=3 to process non-convex models

— Restarting execution
— modGaroneIntCplex_1h.gms(216) 2 Mb
— Reading solution for model garintmiqcp
— modGaroneIntCplex_1h.gms(221) 3 Mb
— Putfile output /home/claudio/CVTSPrun/GAMSCVTSP/SD_11_1.out
*** Status: Normal completion
— Job modGaroneIntCplex_1h.gms Stop 05/21/15 12:26:44 elapsed 0:02:48.308

What is wrong here?
Claudio

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.

Archiver · May 21, 2015, 2:42pm

I think the documentation (http://gams.com/help/infocenter/topic/gams.doc.solver/cplex/index.html#CPLEXsolutiontarget) says it all: “This parameter affects the behavior only when CPLEX uses the barrier algorithm without crossover to solve a nonconvex continuous quadratic model (QP); that is, the variables of the model are continuous, the objective function includes a quadratic term, and the objective function is not positive semi-definite (PSD).”

You solve an SOCP/QCP problem not a QP problem.

Hope this helps,
Michael Bussieck - GAMSWorld Coordinator

On Thursday, May 21, 2015 at 8:36:32 AM UTC-4, Claudio Gambella wrote:

Hello,
I want to solve the linear relaxation of a mixed integer second-order conic problem with CPLEX.
So, the model type is rmiqcp; for letting it solve with CPLEX regardless the non-convexity, i should set the solution target parameter to 3. So, this is what I did in the file cplex.opt.
But the value of the parameters seems not recognised.

These are the options and solve statements:

option rmiqcp=cplex;
garintmiqcp.optfile=1;
Solve garintmiqcp min obj using rmiqcp;

This is the log:

Cplex 12.6.1.0

Reading parameter(s) from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”

*comment

tilim 3600
solutiontarget 3
threads 1
parallelmode 1
Finished reading from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”
Reading data…
Starting Cplex…
Space for names approximately 0.01 Mb
Use option ‘names no’ to turn use of names off
CPLEX Error 5002: Q in ‘QCP_row_for_constraint1qcp(1)’ is not positive semi-definite.
Presolve time = 0.00 sec. (0.00 ticks)
QCP status(0):
Cplex Time: 0.00sec (det. 0.00 ticks)
*** CPLEX Error 5002: Q in %s is not positive semi-definite.
Try Cplex option SolutionTarget=3 to process non-convex models

— Restarting execution
— modGaroneIntCplex_1h.gms(216) 2 Mb
— Reading solution for model garintmiqcp
— modGaroneIntCplex_1h.gms(221) 3 Mb
— Putfile output /home/claudio/CVTSPrun/GAMSCVTSP/SD_11_1.out
*** Status: Normal completion
— Job modGaroneIntCplex_1h.gms Stop 05/21/15 12:26:44 elapsed 0:02:48.308

What is wrong here?
Claudio

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.

Archiver · May 21, 2015, 2:59pm

Reply-to: gamsworld@googlegroups.com

I see the explanation of the parameter solution target.
But then I don’t know which is the model type of the problem according to GAMS.
I can solve the mixed integer model as miqcp with cplex without setting any particular parameters.

Il giorno giovedÃ¬ 21 maggio 2015 14:42:29 UTC+2, Michael Bussieck ha scritto:

I think the documentation (http://gams.com/help/infocenter/topic/gams.doc.solver/cplex/index.html#CPLEXsolutiontarget) says it all: “This parameter affects the behavior only when CPLEX uses the barrier algorithm without crossover to solve a nonconvex continuous quadratic model (QP); that is, the variables of the model are continuous, the objective function includes a quadratic term, and the objective function is not positive semi-definite (PSD).”

You solve an SOCP/QCP problem not a QP problem.

Hope this helps,
Michael Bussieck - GAMSWorld Coordinator

On Thursday, May 21, 2015 at 8:36:32 AM UTC-4, Claudio Gambella wrote:

Hello,
I want to solve the linear relaxation of a mixed integer second-order conic problem with CPLEX.
So, the model type is rmiqcp; for letting it solve with CPLEX regardless the non-convexity, i should set the solution target parameter to 3. So, this is what I did in the file cplex.opt.
But the value of the parameters seems not recognised.

These are the options and solve statements:

option rmiqcp=cplex;
garintmiqcp.optfile=1;
Solve garintmiqcp min obj using rmiqcp;

This is the log:

Cplex 12.6.1.0

Reading parameter(s) from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”

*comment

tilim 3600
solutiontarget 3
threads 1
parallelmode 1
Finished reading from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”
Reading data…
Starting Cplex…
Space for names approximately 0.01 Mb
Use option ‘names no’ to turn use of names off
CPLEX Error 5002: Q in ‘QCP_row_for_constraint1qcp(1)’ is not positive semi-definite.
Presolve time = 0.00 sec. (0.00 ticks)
QCP status(0):
Cplex Time: 0.00sec (det. 0.00 ticks)
*** CPLEX Error 5002: Q in %s is not positive semi-definite.
Try Cplex option SolutionTarget=3 to process non-convex models

— Restarting execution
— modGaroneIntCplex_1h.gms(216) 2 Mb
— Reading solution for model garintmiqcp
— modGaroneIntCplex_1h.gms(221) 3 Mb
— Putfile output /home/claudio/CVTSPrun/GAMSCVTSP/SD_11_1.out
*** Status: Normal completion
— Job modGaroneIntCplex_1h.gms Stop 05/21/15 12:26:44 elapsed 0:02:48.308

What is wrong here?
Claudio

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.

Archiver · May 23, 2015, 2:38pm

Cplex may do some smart reformulations when the variables in the Q part are all binary. So Cplex accepts the discrete (MIQCP) but not the relaxed problem.

Hope this helps,
Michael Bussieck - GAMSWorld Coordinator

On Thursday, May 21, 2015 at 9:21:38 AM UTC-4, Claudio Gambella wrote:

I see the explanation of the parameter solution target.
But then I don’t know which is the model type of the problem according to GAMS.
I can solve the mixed integer model as miqcp with cplex without setting any particular parameters.

Il giorno giovedÃ¬ 21 maggio 2015 14:42:29 UTC+2, Michael Bussieck ha scritto:

I think the documentation (http://gams.com/help/infocenter/topic/gams.doc.solver/cplex/index.html#CPLEXsolutiontarget) says it all: “This parameter affects the behavior only when CPLEX uses the barrier algorithm without crossover to solve a nonconvex continuous quadratic model (QP); that is, the variables of the model are continuous, the objective function includes a quadratic term, and the objective function is not positive semi-definite (PSD).”

You solve an SOCP/QCP problem not a QP problem.

Hope this helps,
Michael Bussieck - GAMSWorld Coordinator

On Thursday, May 21, 2015 at 8:36:32 AM UTC-4, Claudio Gambella wrote:

Hello,
I want to solve the linear relaxation of a mixed integer second-order conic problem with CPLEX.
So, the model type is rmiqcp; for letting it solve with CPLEX regardless the non-convexity, i should set the solution target parameter to 3. So, this is what I did in the file cplex.opt.
But the value of the parameters seems not recognised.

These are the options and solve statements:

option rmiqcp=cplex;
garintmiqcp.optfile=1;
Solve garintmiqcp min obj using rmiqcp;

This is the log:

Cplex 12.6.1.0

Reading parameter(s) from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”

*comment

tilim 3600
solutiontarget 3
threads 1
parallelmode 1
Finished reading from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”
Reading data…
Starting Cplex…
Space for names approximately 0.01 Mb
Use option ‘names no’ to turn use of names off
CPLEX Error 5002: Q in ‘QCP_row_for_constraint1qcp(1)’ is not positive semi-definite.
Presolve time = 0.00 sec. (0.00 ticks)
QCP status(0):
Cplex Time: 0.00sec (det. 0.00 ticks)
*** CPLEX Error 5002: Q in %s is not positive semi-definite.
Try Cplex option SolutionTarget=3 to process non-convex models

— Restarting execution
— modGaroneIntCplex_1h.gms(216) 2 Mb
— Reading solution for model garintmiqcp
— modGaroneIntCplex_1h.gms(221) 3 Mb
— Putfile output /home/claudio/CVTSPrun/GAMSCVTSP/SD_11_1.out
*** Status: Normal completion
— Job modGaroneIntCplex_1h.gms Stop 05/21/15 12:26:44 elapsed 0:02:48.308

What is wrong here?
Claudio

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.

Archiver · May 25, 2015, 10:22am

Reply-to: gamsworld@googlegroups.com

Yes, this seems the situation.
Thank you for the support.
Claudio

Il giorno sabato 23 maggio 2015 14:38:41 UTC+2, Michael Bussieck ha scritto:

Cplex may do some smart reformulations when the variables in the Q part are all binary. So Cplex accepts the discrete (MIQCP) but not the relaxed problem.

Hope this helps,
Michael Bussieck - GAMSWorld Coordinator

On Thursday, May 21, 2015 at 9:21:38 AM UTC-4, Claudio Gambella wrote:

I see the explanation of the parameter solution target.
But then I don’t know which is the model type of the problem according to GAMS.
I can solve the mixed integer model as miqcp with cplex without setting any particular parameters.

Il giorno giovedÃ¬ 21 maggio 2015 14:42:29 UTC+2, Michael Bussieck ha scritto:

I think the documentation (http://gams.com/help/infocenter/topic/gams.doc.solver/cplex/index.html#CPLEXsolutiontarget) says it all: “This parameter affects the behavior only when CPLEX uses the barrier algorithm without crossover to solve a nonconvex continuous quadratic model (QP); that is, the variables of the model are continuous, the objective function includes a quadratic term, and the objective function is not positive semi-definite (PSD).”

You solve an SOCP/QCP problem not a QP problem.

Hope this helps,
Michael Bussieck - GAMSWorld Coordinator

On Thursday, May 21, 2015 at 8:36:32 AM UTC-4, Claudio Gambella wrote:

Hello,
I want to solve the linear relaxation of a mixed integer second-order conic problem with CPLEX.
So, the model type is rmiqcp; for letting it solve with CPLEX regardless the non-convexity, i should set the solution target parameter to 3. So, this is what I did in the file cplex.opt.
But the value of the parameters seems not recognised.

These are the options and solve statements:

option rmiqcp=cplex;
garintmiqcp.optfile=1;
Solve garintmiqcp min obj using rmiqcp;

This is the log:

Cplex 12.6.1.0

Reading parameter(s) from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”

*comment

tilim 3600
solutiontarget 3
threads 1
parallelmode 1
Finished reading from “/home/claudio/CVTSPrun/GAMSCVTSP/cplex.opt”
Reading data…
Starting Cplex…
Space for names approximately 0.01 Mb
Use option ‘names no’ to turn use of names off
CPLEX Error 5002: Q in ‘QCP_row_for_constraint1qcp(1)’ is not positive semi-definite.
Presolve time = 0.00 sec. (0.00 ticks)
QCP status(0):
Cplex Time: 0.00sec (det. 0.00 ticks)
*** CPLEX Error 5002: Q in %s is not positive semi-definite.
Try Cplex option SolutionTarget=3 to process non-convex models

— Restarting execution
— modGaroneIntCplex_1h.gms(216) 2 Mb
— Reading solution for model garintmiqcp
— modGaroneIntCplex_1h.gms(221) 3 Mb
— Putfile output /home/claudio/CVTSPrun/GAMSCVTSP/SD_11_1.out
*** Status: Normal completion
— Job modGaroneIntCplex_1h.gms Stop 05/21/15 12:26:44 elapsed 0:02:48.308

What is wrong here?
Claudio

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.