Dear Friends,

I am using BARON and find a solution but it says some variables or

expressions are unbounded and the global solution is not guaranteed. I

look at BARON manual and there is

a comment on that

***User did not provide appropriate variable bounds ***. The user will

need to read the BARON

output (in le sum.scr in the gamskeep directory) for likely pointers

to variables and expressions with

missing bounds.

But i could not find such a gamskeep directory and in gamsdir there is

no subdirectory like 225a, I am using the latest version of GAMS.

I will appreciate any hints,

Thank you very much,

Have a good one,

\

Hi,

You need to add to the GAMS command line (or the IDE parameter box)

the parameter keep=1. This will prevent GAMS from removing the scratch

directory. Now in 225a (or the lastest 225? directory) you will find

sum.dat (not sum.scr, this has been changed some time ago). In this

file you will find the required information.

For example, from a Windows CMD prompt (click on the MSDOS icon in the

GAMS IDE) do:

gamslib chenery

gams chenery nlp baron keep=1

notepad 225a\sum.dat

in the file you find the lines:

Missing bound for variable x39 in relation x39*x16

Now how do you find out what variable x39 represents in your original

GAMS model? You can “solve” the model with convert and look at the

dictionary mapping:

gams chenery nlp convert

notepad dict.txt

this has: “x39 pd”, so the scalar variable x39 represents variable pd

in the chenery model. If you check the chenery model, pd is a free

variable (no bounds) and is used in the constraint dem as pd*p(i).

BTW, a look-up in dict.txt makes the connection x16 = p(services).

I’ll see that the documentation gets an update.

Hope this helps.

Michael Bussieck

GAMSWorld Coordinator

On Jan 2, 6:22 pm, Rah wrote:

Dear Friends,

I am using BARON and find a solution but it says some variables or

expressions are unbounded and the global solution is not guaranteed. I

look at BARON manual and there is

a comment on that

***User did not provide appropriate variable bounds ***. The user will

need to read the BARON

output (in le sum.scr in the gamskeep directory) for likely pointers

to variables and expressions with

missing bounds.

But i could not find such a gamskeep directory and in gamsdir there is

no subdirectory like 225a, I am using the latest version of GAMS.

I will appreciate any hints,

Thank you very much,

Have a good one,

\

Thanks, Michael! It was very helpful for me also. However, i am

solving a model using BARON which hasmissing bounds for some

expressions as well. my sum.dat has the following message:

“Missing bound for intermediate expression bar_var22 in bilinear or

rational composition

Detail of the intermediate expression:

(x1 * x7) * x36”

I undestand that providing better bounds for x1, x7 and/or x36 might

help. Is there any way to provide bound for the expression “bar_var22”

also? by the way, dict.txt does not indicate which expression bar_var

actually is. Is there any way to know that?

many thanks!

On Jan 3, 10:20 am, Michael Bussieck wrote:

Hi,

You need to add to the GAMS command line (or the IDE parameter box)

the parameter keep=1. This will prevent GAMS from removing the scratch

directory. Now in 225a (or the lastest 225? directory) you will find

sum.dat (not sum.scr, this has been changed some time ago). In this

file you will find the required information.

For example, from a Windows CMD prompt (click on the MSDOS icon in the

GAMS IDE) do:

gamslib chenery

gams chenery nlp baron keep=1

notepad 225a\sum.dat

in the file you find the lines:

Missing bound for variable x39 in relation x39*x16

Now how do you find out what variable x39 represents in your original

GAMS model? You can “solve” the model with convert and look at the

dictionary mapping:

gams chenery nlp convert

notepad dict.txt

this has: “x39 pd”, so the scalar variable x39 represents variable pd

in the chenery model. If you check the chenery model, pd is a free

variable (no bounds) and is used in the constraint dem as pd*p(i).

BTW, a look-up in dict.txt makes the connection x16 = p(services).

I’ll see that the documentation gets an update.

Hope this helps.

Michael Bussieck

GAMSWorld Coordinator

On Jan 2, 6:22 pm, Rah wrote:

Dear Friends,

I am using BARON and find a solution but it says some variables or

expressions are unbounded and the global solution is not guaranteed. I

look at BARON manual and there is

a comment on that

***User did not provide appropriate variable bounds ***. The user will

need to read the BARON

output (in le sum.scr in the gamskeep directory) for likely pointers

to variables and expressions with

missing bounds.

But i could not find such a gamskeep directory and in gamsdir there is

no subdirectory like 225a, I am using the latest version of GAMS.

I will appreciate any hints,

Thank you very much,

Have a good one,

\

I don’t think you can explicitely bound an expression, you need to do

that by bounding the variables in the expression. I think, BARON is

telling you that bar_var22 = (x1 * x7) * x36. I can also check with

the developer, but for that I need an example. So you if could mail

the model to support@gams.com and refer to this conversion, I’ll take

care of that.

Michael Bussieck - GAMSWorld Coordinator

On Jan 11, 11:15 am, faha wrote:

Thanks, Michael! It was very helpful for me also. However, i am

solving a model using BARON which hasmissing bounds for some

expressions as well. my sum.dat has the following message:

“Missing bound for intermediate expression bar_var22 in bilinear or

rational composition

Detail of the intermediate expression:

(x1 * x7) * x36”

I undestand that providing better bounds for x1, x7 and/or x36 might

help. Is there any way to provide bound for the expression “bar_var22”

also? by the way, dict.txt does not indicate which expression bar_var

actually is. Is there any way to know that?

many thanks!

On Jan 3, 10:20 am, Michael Bussieck wrote:

Hi,

You need to add to the GAMS command line (or the IDE parameter box)

the parameter keep=1. This will prevent GAMS from removing the scratch

directory. Now in 225a (or the lastest 225? directory) you will find

sum.dat (not sum.scr, this has been changed some time ago). In this

file you will find the required information.

For example, from a Windows CMD prompt (click on the MSDOS icon in the

GAMS IDE) do:

gamslib chenery

gams chenery nlp baron keep=1

notepad 225a\sum.dat

in the file you find the lines:

Missing bound for variable x39 in relation x39*x16

Now how do you find out what variable x39 represents in your original

GAMS model? You can “solve” the model with convert and look at the

dictionary mapping:

gams chenery nlp convert

notepad dict.txt

this has: “x39 pd”, so the scalar variable x39 represents variable pd

in the chenery model. If you check the chenery model, pd is a free

variable (no bounds) and is used in the constraint dem as pd*p(i).

BTW, a look-up in dict.txt makes the connection x16 = p(services).

I’ll see that the documentation gets an update.

Hope this helps.

Michael Bussieck

GAMSWorld Coordinator

On Jan 2, 6:22 pm, Rah wrote:

Dear Friends,

I am using BARON and find a solution but it says some variables or

expressions are unbounded and the global solution is not guaranteed. I

look at BARON manual and there is

a comment on that

***User did not provide appropriate variable bounds ***. The user will

need to read the BARON

output (in le sum.scr in the gamskeep directory) for likely pointers

to variables and expressions with

missing bounds.

But i could not find such a gamskeep directory and in gamsdir there is

no subdirectory like 225a, I am using the latest version of GAMS.

I will appreciate any hints,

Thank you very much,

Have a good one,

\

I have a similar problem:

The sum.dat file tells me the following:

Missing bound for variable x14 in relation x14^3

Missing bound for variable x13 in relation x13^3

However, variables x13 and x14 are bounded (by 200 and 900). How can I find out what relations x14^3 means? Or does this actually mean x14 to the power of 3? (The term x14^3 does appear in my model, but I do not see why it should cause problems. The maximum value for x14 is 900, 900 to the power of 3=729e+6)

–

To view this discussion on the web visit https://groups.google.com/d/msg/gamsworld/-/q-3f9d_NmTkJ.

To post to this group, send email to gamsworld@googlegroups.com.

To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en.