Nabble - lp_solve

Re: Description of Branch-Bound rules and when to use them?

2009-12-09T21:32:45Z

Thanks William.

--- In lp_solve@..., "William H. Patton" <pattonwh@...> wrote:

>
> NO, not at all. Branch and bound is the process that searches for the
> integer solutions INSIDE the convex hull described by the constraints.
> "which non-integer variable is to be selected" means
>
> at this tree node, which column required to be integer BUT not yet integer
> in the current LP solution is chosen to be fixed at a near by integer value.
> This defines the next problem node to be solved.
>
> The next problem may be infeasible or have a poorer solution than the best
> current solution.
>
> If so, this branch of the tree search is abandoned and another choice is
> made. The same happens if it has a better solution - another not yet integer
> is fixed and the search proceeds deeper into the tree.
>
> When all the required to be integer columns achieve integer values a
> feasible terminal node is reached.
>
> The objective value here is compared with the current best integer solution
> and this node is abandoned by marking this branch done. The search continues
> upward to an un-examined branch.
>
> Branch and bound finishes when all nodes have been marked as examind.
>
>
>
> The rules are there - together with more subtle options like variable
> priority assignment - to influence the choice of the next not yet integer
> variable to fix. Different choices can vastly reduce the number of nodes
> actually searched. But ii is only a kind of trial process that can find
> which choice is bes for a bunch of similar problems.
>
>
>
> set_var_weights
> <mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
> set_var_weights.htm> , get_var_priority
> <mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
> get_var_priority.htm> , set_var_branch
> <mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
> set_var_branch.htm> , get_var_branch
> <mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
> get_var_branch.htm> ,
>
>
>
> William
>
> _____
>
> From: lp_solve@... [mailto:lp_solve@...] On Behalf
> Of geoffwhittington
> Sent: Wednesday, December 02, 2009 12:13 AM
> To: lp_solve@...
> Subject: [lp_solve] Re: Description of Branch-Bound rules and when to use
> them?
>
>
>
>
>
> Is it fair to say that the rule is irrelevant for those problems that have
> only integer columns? I say this because of the function description ...
>
> >>>> The set_bb_rule function specifies the branch-and-bound rule for
> choosing which non-integer variable is to be selected. This rule can
> influence solving times considerably. Depending on the model one rule can be
> best and for another model another rule.
> The default is NODE_FIRSTSELECT (0).
>
> --- In lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com, Kjell
> Eikland <kjell.eikland@> wrote:
> >
> > The simple answer is no. You have to look at various textbooks, papers,
> the
> > source code and on-line help to find this kind of info. Usually, the
> > suitability of options is highly problem-dependent. At some point, it may
> > be writen up fully in the help documentation.
> >
> > Kjell
> >
> > -> -----Original Message-----
> > -> From: lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com
> > -> [mailto:lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com]
> On Behalf Of dog4pavlov
> > -> Sent: 12 June 2006 20:31
> > -> To: lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com
> > -> Subject: [lp_solve] Description of Branch-Bound rules and
> > -> when to use them?
> > ->
> > -> Hi All,
> > ->
> > -> Sorry if this has been discussed before.
> > ->
> > -> Is there a place that describes the details of each
> > -> Branch-Bound rule
> > -> choice (such as Greedy, PseudoCost, etc..) and when best to
> > -> use these?
> > ->
> > -> Thanks,
> > ->
> > ->
> > ->
> > ->
> > ->
> > ->
> > ->
> > -> ------------------------ Yahoo! Groups Sponsor
> > -> --------------------~-->
> > -> You can search right from your browser? It's easy and it's
> > -> free. See how.
> > -> http://us.click.
> <http://us.click.yahoo.com/_7bhrC/NGxNAA/yQLSAA/W4wwlB/TM>
> yahoo.com/_7bhrC/NGxNAA/yQLSAA/W4wwlB/TM
> > -> ----------------------------------------------------------
> > -> -------~->
> > ->
> > ->
> > -> Yahoo! Groups Links
> > ->
> > ->
> > ->
> > ->
> > ->
> > ->
> >
>

Re: Access violation

2009-12-09T13:23:08Z

Can't do anything with this. You give too less information.

Peter

From: marwan_abb
Sent: Wednesday, December 09, 2009 22:09
To: lp_solve@...
Subject: [lp_solve] Access violation

I am trying to call lp_solve from visual studio and when i try to run the code i have the following error "0xC0000005: Access violation writing location 0x00000010".
Thanks

Access violation

2009-12-09T13:09:51Z

I am trying to call lp_solve from visual studio and when i try to run the code i have the following error "0xC0000005: Access violation writing location 0x00000010".
Thanks

Re: conditional statements

2009-12-09T08:51:21Z

Yes,

by using integer/binary variables.

Peter

On Wed, Dec 9, 2009 at 5:20 PM, carlogiacomello
<carlogiacomello@...>wrote:

>
>
> Hi everyone,
>
> I wonder if there are some techniques to define conditional statements in a
> linear problem. I would like to set some constraints like the following
> ones:
>
> IF ( X>0)
> Y = A X + B
> ELSE
> Y = C X + D
>
> Thank you.
>
> Regards,
> Carlo
>
>
>

conditional statements

2009-12-09T08:20:03Z

Hi everyone,

I wonder if there are some techniques to define conditional statements in a linear problem. I would like to set some constraints like the following ones:

IF ( X>0)
Y = A X + B
ELSE
Y = C X + D

Thank you.

Regards,
Carlo

RE: Unbounded solution in LPSolve

2009-12-08T13:10:32Z

Presolve columns can detect this.

_____

From: lp_solve@... [mailto:lp_solve@...] On Behalf
Of Jasper de Koning
Sent: Tuesday, December 08, 2009 12:38 PM
To: lp_solve@...
Subject: [lp_solve] Unbounded solution in LPSolve

Hello,

I encountered an issue while running LPSolve 5.5.0.14; it finds a solution
for an unbounded problem. Below is an example:

/* Objective function */
max: +C1 + C2 +C3;

/* Constraints */
r1: +C1 + C2 = 10;
r2: +C1 + 2*C2 <= 15;
r3: +C1 >= 0;
r4: +C2 >= 0;

This problem is unbounded, because C3 is a free variable. When I add one
extra restriction to the model: "r5: +C3 >= 0;" lpsolve returns that the
model is unbounded (as I would expect).

Regards,
Jasper de Koning

Unbounded solution in LPSolve

2009-12-08T10:37:50Z

Hello,

I encountered an issue while running LPSolve 5.5.0.14; it finds a solution
for an unbounded problem. Below is an example:

/* Objective function */
max: +C1 + C2 +C3;

/* Constraints */
r1: +C1 + C2 = 10;
r2: +C1 + 2*C2 <= 15;
r3: +C1 >= 0;
r4: +C2 >= 0;

This problem is unbounded, because C3 is a free variable. When I add one
extra restriction to the model: "r5: +C3 >= 0;" lpsolve returns that the
model is unbounded (as I would expect).

Regards,
Jasper de Koning

Re: header file cannot be found

2009-12-08T00:50:32Z

Thanks a lot Peter i will try it.

Marwan Abbadi

--- On Mon, 12/7/09, Peter Notebaert <_peno_@...> wrote:

From: Peter Notebaert <_peno_@...>
Subject: Re: [lp_solve] header file cannot be found
To: lp_solve@...
Date: Monday, December 7, 2009, 4:34 PM

That is a unix header file.

The reason for the problem is that you have not
defined WIN32.

So you must define WIN32 in the project and it
should compile.

Peter

From: marwan_abb
Sent: Tuesday, December 08, 2009 00:06
To: lp_solve@yahoogroup s.com
Subject: [lp_solve] header file cannot be found

I am trying to call lp_solve from visual studio 2005, while adding necessary
headers and run the code the following error keeps appearing "fatal error C1083:
Cannot open include file: 'dlfcn.h': No such file or directory", i tried to
manually locate the header i couldn't find it.

Re: header file cannot be found

2009-12-07T16:34:33Z

That is a unix header file.

The reason for the problem is that you have not defined WIN32.

So you must define WIN32 in the project and it should compile.

Peter

From: marwan_abb
Sent: Tuesday, December 08, 2009 00:06
To: lp_solve@...
Subject: [lp_solve] header file cannot be found

I am trying to call lp_solve from visual studio 2005, while adding necessary headers and run the code the following error keeps appearing "fatal error C1083: Cannot open include file: 'dlfcn.h': No such file or directory", i tried to manually locate the header i couldn't find it.

header file cannot be found

2009-12-07T15:06:53Z

I am trying to call lp_solve from visual studio 2005, while adding necessary headers and run the code the following error keeps appearing "fatal error C1083: Cannot open include file: 'dlfcn.h': No such file or directory", i tried to manually locate the header i couldn't find it.

how to use lpsolve in c++

2009-12-04T19:53:36Z

hi! I am new in LPSOLVE and I want to call LPsolve package in VS 2005, since I am going to count my routine, which is written in c/c++, but I don't know how to call Lpsolve, which is the key package for me, I have downloaded the file lp_solve_5.5.0.15_dev_win32, and my OS is windows XP, the head file like #include "lp_lib.h" is also in my c programm, but the error is that it can't open the file"lp_lib.h"; what should I do after I decompressed files lp_solve_5.5.0.15_dev_win32.zip. thanks! with best wishes yours sincerely lei

Re: MIP

2009-12-04T10:07:08Z

Got it figured out!! Had too many zeros in some constraints. Works perfect!

Leigh

Leigh A wrote:

Hi,

I want to solve a problem where the solution variables have both integers and non-integers but I don't know hwo to setup for some solution variables to be set as integers.

This is my code:

package MIP;

import lpsolve.*;

public class MIP {

public static void main(String[] args) {
try {
// Create a problem with 4 variables and 0 constraints
LpSolve solver = LpSolve.makeLp(0, 24);

// add constraints
solver.strAddConstraint("1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 20000);
solver.strAddConstraint("1 1 1 1 1 1 1 1 1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 1 0 0 -1500 0 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 1 0 0 -2500 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 1 0 0 -5500 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -1 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -1 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -1 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 -1 -1 -1 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -2500 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -3500 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -5500", LpSolve.LE, 0);

// set objective function
solver.strSetObjFn("108.57 105.94 83.28 64.08 94.37 51.59 82.18 139.99 96.33 -9.8 -7.27 -5.79 -5388 -7080 -10187 -15182 -16942 -19186 -12.65 -10.04 -8.34 -27535 -33784 -37999");
solver.setMaxim();

// solve the problem
solver.solve();

// print solution
System.out.println("Value of objective function: " + solver.getObjective());
double[] var = solver.getPtrVariables();
for (int i = 0; i < var.length; i++) {
System.out.println("Value of var[" + i + "] = " + var[i]);
}

// delete the problem and free memory
solver.deleteLp();
}
catch (LpSolveException e) {
e.printStackTrace();
}
}

}

The answer:

Value of objective function: 2376389.090909091
Value of var[0] = 0.0
Value of var[1] = 0.0
Value of var[2] = 0.0
Value of var[3] = 0.0
Value of var[4] = 0.0
Value of var[5] = 0.0
Value of var[6] = 0.0
Value of var[7] = 20000.0
Value of var[8] = 0.0
Value of var[9] = 0.0
Value of var[10] = 0.0
Value of var[11] = 20000.0
Value of var[12] = 0.0 // if this comes into solution it needs to be an integer
Value of var[13] = 0.0 // if this comes into solution it needs to be an integer
Value of var[14] = 3.6363636363636367 // if this comes into solution it needs to be an integer
Value of var[15] = 0.0 // if this comes into solution it needs to be an integer
Value of var[16] = 0.0 // if this comes into solution it needs to be an integer
Value of var[17] = 3.636363636363636 //if this comes into solution it needs to be an integer
Value of var[18] = 0.0
Value of var[19] = 19999.999999999996
Value of var[20] = 0.0
Value of var[21] = 0.0 //if this comes into solution it needs to be an integer
Value of var[22] = 0.0 //if this comes into solution it needs to be an integer
Value of var[23] = 0.0 //if this comes into solution it needs to be an integer

Model name: '' - run #1
Objective: Maximize(R0)

Thanks

Leigh A.

MIP

2009-12-03T13:01:32Z

Hi,

I want to solve a problem where the solution variables have both integers and non-integers but I don't know hwo to setup for some solution variables to be set as integers.

This is my code:

package MIP;

import lpsolve.*;

public class MIP {

public static void main(String[] args) {
try {
// Create a problem with 4 variables and 0 constraints
LpSolve solver = LpSolve.makeLp(0, 24);

// add constraints
solver.strAddConstraint("1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 20000);
solver.strAddConstraint("1 1 1 1 1 1 1 1 1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 1 0 0 -1500 0 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 1 0 0 -2500 0 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 1 0 0 -5500 0 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -1 0 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -1 0 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -1 0 0 0 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 -1 -1 -1 0 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -2500 0 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -3500 0", LpSolve.LE, 0);
solver.strAddConstraint("0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 -5500", LpSolve.LE, 0);

// set objective function
solver.strSetObjFn("108.57 105.94 83.28 64.08 94.37 51.59 82.18 139.99 96.33 -9.8 -7.27 -5.79 -5388 -7080 -10187 -15182 -16942 -19186 -12.65 -10.04 -8.34 -27535 -33784 -37999");
solver.setMaxim();

// solve the problem
solver.solve();

// print solution
System.out.println("Value of objective function: " + solver.getObjective());
double[] var = solver.getPtrVariables();
for (int i = 0; i < var.length; i++) {
System.out.println("Value of var[" + i + "] = " + var[i]);
}

// delete the problem and free memory
solver.deleteLp();
}
catch (LpSolveException e) {
e.printStackTrace();
}
}

}

The answer:

Value of objective function: 2376389.090909091
Value of var[0] = 0.0
Value of var[1] = 0.0
Value of var[2] = 0.0
Value of var[3] = 0.0
Value of var[4] = 0.0
Value of var[5] = 0.0
Value of var[6] = 0.0
Value of var[7] = 20000.0
Value of var[8] = 0.0
Value of var[9] = 0.0
Value of var[10] = 0.0
Value of var[11] = 20000.0
Value of var[12] = 0.0 // if this comes into solution it needs to be an integer
Value of var[13] = 0.0 // if this comes into solution it needs to be an integer
Value of var[14] = 3.6363636363636367 // if this comes into solution it needs to be an integer
Value of var[15] = 0.0 // if this comes into solution it needs to be an integer
Value of var[16] = 0.0 // if this comes into solution it needs to be an integer
Value of var[17] = 3.636363636363636 //if this comes into solution it needs to be an integer
Value of var[18] = 0.0
Value of var[19] = 19999.999999999996
Value of var[20] = 0.0
Value of var[21] = 0.0 //if this comes into solution it needs to be an integer
Value of var[22] = 0.0 //if this comes into solution it needs to be an integer
Value of var[23] = 0.0 //if this comes into solution it needs to be an integer

Model name: '' - run #1
Objective: Maximize(R0)

Thanks

Leigh A.

Re: Will LPSolve work on Windows 7

2009-12-03T10:49:54Z

I'm using windows 7 and LPSolve IDE works just fine.

--------------------------------------------------
Binh Nguyen
Computer Science Department
Rensselaer Polytechnic Institute
Troy, NY, 12180
--------------------------------------------------

On Thu, Dec 3, 2009 at 1:03 PM, tranlo <tranlo@...> wrote:

>
>
> Hi,
>
> I just ordered a new monster desktop (12GB RAM, iCore 920 for $999 from
> Costco.com) but it comes with Windows 7.
>
> I should get it in about 7 days, but I was wondering will the LPSolve IDE
> will still work on Windows 7. Has anyone been able to install it and run it
> on Windows 7?
>
> Thanks!
>
> Laurent
>
> PS: Peter, thank you so much for this wonderful free solver!
>
>
>

Will LPSolve work on Windows 7

2009-12-03T10:03:09Z

Hi,

I just ordered a new monster desktop (12GB RAM, iCore 920 for $999 from Costco.com) but it comes with Windows 7.

I should get it in about 7 days, but I was wondering will the LPSolve IDE will still work on Windows 7. Has anyone been able to install it and run it on Windows 7?

Thanks!

Laurent

PS: Peter, thank you so much for this wonderful free solver!

Re: Lpsolve with Matlab

2009-12-02T14:09:41Z

Thank you peter =)

________________________________
From: Peter Notebaert <_peno_@...>
To: lp_solve@...
Sent: Wed, December 2, 2009 9:37:44 PM
Subject: Re: [lp_solve] Lpsolve with Matlab

That can be done with
set_break_at_ value:

mxlpsolve('set_break_at_ value ' ,
lp, obj1*1.15);

Peter

From: Tengku Joe
Sent: Wednesday, December 02, 2009 19:56
To: lp_solve@yahoogroup s.com
Subject: [lp_solve] Lpsolve with Matlab

Hi guys,
I’m using Lpsolve with Matlab in order to solve my
problem.
The final script should be:
lp_handle
=
lp_maker(f,a, b,e,vlb,vub, xint,scalemode, setminim) ;
solvestat
= mxlpsolve('solve' , lp);
obj= mxlpsolve('get_ objective' ,
lp);
x = mxlpsolve('get_ variables' ,
lp);
mxlpsolve('delete_ lp', lp);

Let say for
the first run I
use real number and I get obj1. Then I want to run it for the
second time
with integer and try to get obj2. However I want to set up a
condition for
the second run where the program will stop and give the answer if
the result
is obj2 <=obj1+15%.

How can I do
that?

I really
appreciate
your help. Thank you.

-suli-

Re: Lpsolve with Matlab

2009-12-02T13:37:44Z

That can be done with set_break_at_value:

mxlpsolve('set_break_at_value', lp, obj1*1.15);

Peter

From: Tengku Joe
Sent: Wednesday, December 02, 2009 19:56
To: lp_solve@...
Subject: [lp_solve] Lpsolve with Matlab

Hi guys,
I’m using Lpsolve with Matlab in order to solve my problem.
The final script should be:
lp_handle =
lp_maker(f,a,b,e,vlb,vub,xint,scalemode,setminim);
solvestat = mxlpsolve('solve', lp);
obj= mxlpsolve('get_objective', lp);
x = mxlpsolve('get_variables', lp);
mxlpsolve('delete_lp', lp);

Let say for
the first run I use real number and I get obj1. Then I want to run it for the
second time with integer and try to get obj2. However I want to set up a
condition for the second run where the program will stop and give the answer if
the result is obj2 <=obj1+15%.

How can I do
that?

I really appreciate
your help. Thank you.

-suli-

Lpsolve with Matlab

2009-12-02T10:56:50Z

Hi guys,
I’m using Lpsolve with Matlab in order to solve my problem.
The final script should be:
lp_handle =
lp_maker(f,a,b,e,vlb,vub,xint,scalemode,setminim);
solvestat = mxlpsolve('solve', lp);
obj= mxlpsolve('get_objective', lp);
x = mxlpsolve('get_variables', lp);
mxlpsolve('delete_lp', lp);

Let say for
the first run I use real number and I get obj1. Then I want to run it for the
second time with integer and try to get obj2. However I want to set up a
condition for the second run where the program will stop and give the answer if
the result is obj2 <=obj1+15%.

How can I do
that?

I really appreciate
your help. Thank you.

-suli-

RE: Re: Description of Branch-Bound rules and when to use them?

2009-12-02T08:21:56Z

NO, not at all. Branch and bound is the process that searches for the
integer solutions INSIDE the convex hull described by the constraints.
"which non-integer variable is to be selected" means

at this tree node, which column required to be integer BUT not yet integer
in the current LP solution is chosen to be fixed at a near by integer value.
This defines the next problem node to be solved.

The next problem may be infeasible or have a poorer solution than the best
current solution.

If so, this branch of the tree search is abandoned and another choice is
made. The same happens if it has a better solution - another not yet integer
is fixed and the search proceeds deeper into the tree.

When all the required to be integer columns achieve integer values a
feasible terminal node is reached.

The objective value here is compared with the current best integer solution
and this node is abandoned by marking this branch done. The search continues
upward to an un-examined branch.

Branch and bound finishes when all nodes have been marked as examind.

The rules are there - together with more subtle options like variable
priority assignment - to influence the choice of the next not yet integer
variable to fix. Different choices can vastly reduce the number of nodes
actually searched. But ii is only a kind of trial process that can find
which choice is bes for a bunch of similar problems.

set_var_weights
<mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
set_var_weights.htm> , get_var_priority
<mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
get_var_priority.htm> , set_var_branch
<mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
set_var_branch.htm> , get_var_branch
<mk:@MSITStore:F:\lpsolve-55\_lp_solveSRCS\5.5.0.14\lp_solve_5.5.0.14.chm::/
get_var_branch.htm> ,

William

_____

From: lp_solve@... [mailto:lp_solve@...] On Behalf
Of geoffwhittington
Sent: Wednesday, December 02, 2009 12:13 AM
To: lp_solve@...
Subject: [lp_solve] Re: Description of Branch-Bound rules and when to use
them?

Is it fair to say that the rule is irrelevant for those problems that have
only integer columns? I say this because of the function description ...

>>>> The set_bb_rule function specifies the branch-and-bound rule for
choosing which non-integer variable is to be selected. This rule can
influence solving times considerably. Depending on the model one rule can be
best and for another model another rule.
The default is NODE_FIRSTSELECT (0).

--- In lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com, Kjell
Eikland <kjell.eikland@...> wrote:
>
> The simple answer is no. You have to look at various textbooks, papers,
the
> source code and on-line help to find this kind of info. Usually, the
> suitability of options is highly problem-dependent. At some point, it may
> be writen up fully in the help documentation.
>
> Kjell
>
> -> -----Original Message-----
> -> From: lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com
> -> [mailto:lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com]
On Behalf Of dog4pavlov

> -> Sent: 12 June 2006 20:31
> -> To: lp_solve@yahoogroup <mailto:lp_solve%40yahoogroups.com> s.com
> -> Subject: [lp_solve] Description of Branch-Bound rules and
> -> when to use them?
> ->
> -> Hi All,
> ->
> -> Sorry if this has been discussed before.
> ->
> -> Is there a place that describes the details of each
> -> Branch-Bound rule
> -> choice (such as Greedy, PseudoCost, etc..) and when best to
> -> use these?
> ->
> -> Thanks,
> ->
> ->
> ->
> ->
> ->
> ->
> ->
> -> ------------------------ Yahoo! Groups Sponsor
> -> --------------------~-->
> -> You can search right from your browser? It's easy and it's
> -> free. See how.
> -> http://us.click.

<http://us.click.yahoo.com/_7bhrC/NGxNAA/yQLSAA/W4wwlB/TM>
yahoo.com/_7bhrC/NGxNAA/yQLSAA/W4wwlB/TM

> -> ----------------------------------------------------------
> -> -------~->
> ->
> ->
> -> Yahoo! Groups Links
> ->
> ->
> ->
> ->
> ->
> ->
>

Re: lp_solve55 error about alloc

2009-12-02T00:03:03Z

My guess is that you don't free memory.
Your allocated memory and lp_solves memory. The latter via calling delete_lp after you are done with a submodel.

Peter

From: Jack Bryan
Sent: Tuesday, December 01, 2009 22:13
To: lp_solve@...
Subject: [lp_solve] lp_solve55 error about alloc

HI,

I am trying to use lp_solve5.5 to set up a small MILP model.

It is bounded and feasible. I need to solve it with different A,b,c ever time in a loop.
So, I use make_lp to design a new model every time.

After running some loops, I got the following message " alloc of 10000 'INT' failed ".

but the model is still boudned and feasible and give optimal solutions.

And, if I set the loop times very large, such as 16*24*24, the C program stops and give:
-------------------------------------------------------------------------------------
rank 7 in job 578 server.name caused collective abort of all ranks
exit status of rank 7: killed by signal 9
rank 3 in job 578 server.name caused collective abort of all ranks
exit status of rank 3: return code 1
---------------------------------------------------------------------------

I run the C program in a MPI cluster .

why ?

Any help is appreciated.

thanks

Jack

Dec. 1 2009

---------------------------------

Model name: 'milp_model' - run #1
Objective: Minimize(R0)

SUBMITTED
Model size: 8 constraints, 6 variables, 16 non-zeros.
Sets: 0 GUB, 0 SOS.

Using DUAL simplex for phase 1 and PRIMAL simplex for phase 2.
The primal and dual simplex pricing strategy set to 'Devex'.

alloc of 10000 'INT' failed
alloc of 10000 'INT' failed
alloc of 10000 'INT' failed
alloc of 10000 'REAL' failed
alloc of 10000 'INT' failed

Relaxed solution 237566.262512 after 4 iter is B&B base.

Feasible solution 237566.262512 after 4 iter, 0 nodes (gap 0.0%)

Optimal solution 237566.262512 after 4 iter, 0 nodes (gap 0.0%).

Excellent numeric accuracy ||*|| = 8.64864e-13

MEMO: lp_solve version 5.5.0.13 for 32 bit OS, with 64 bit REAL variables.
In the total iteration count 4, 0 (0.0%) were bound flips.
There were 0 refactorizations, 0 triggered by time and 0 by density.
... on average 4.0 major pivots per refactorization.
The largest [LUSOL v2.2.1.0] fact(B) had 9 NZ entries, 1.0x largest basis.
The maximum B&B level was 1, 0.5x MIP order, 1 at the optimal solution.
The constraint matrix inf-norm is 2053.11, with a dynamic range of 2053.11.
Time to load data was 0.006 seconds, presolve used 0.001 seconds,
... 0.009 seconds in simplex solver, in total 0.016 seconds.

-----------------------------------------

--------------------------------------------------------------------------------
Chat with Messenger straight from your Hotmail inbox. Check it out

Re: Description of Branch-Bound rules and when to use them?

2009-12-01T22:12:51Z

Is it fair to say that the rule is irrelevant for those problems that have only integer columns? I say this because of the function description ...

>>>> The set_bb_rule function specifies the branch-and-bound rule for choosing which non-integer variable is to be selected. This rule can influence solving times considerably. Depending on the model one rule can be best and for another model another rule.
The default is NODE_FIRSTSELECT (0).

--- In lp_solve@..., Kjell Eikland <kjell.eikland@...> wrote:

>
> The simple answer is no. You have to look at various textbooks, papers, the
> source code and on-line help to find this kind of info. Usually, the
> suitability of options is highly problem-dependent. At some point, it may
> be writen up fully in the help documentation.
>
> Kjell
>
> -> -----Original Message-----
> -> From: lp_solve@...
> -> [mailto:lp_solve@...] On Behalf Of dog4pavlov
> -> Sent: 12 June 2006 20:31
> -> To: lp_solve@...
> -> Subject: [lp_solve] Description of Branch-Bound rules and
> -> when to use them?
> ->
> -> Hi All,
> ->
> -> Sorry if this has been discussed before.
> ->
> -> Is there a place that describes the details of each
> -> Branch-Bound rule
> -> choice (such as Greedy, PseudoCost, etc..) and when best to
> -> use these?
> ->
> -> Thanks,
> ->
> ->
> ->
> ->
> ->
> ->
> ->
> -> ------------------------ Yahoo! Groups Sponsor
> -> --------------------~-->
> -> You can search right from your browser? It's easy and it's
> -> free. See how.
> -> http://us.click.yahoo.com/_7bhrC/NGxNAA/yQLSAA/W4wwlB/TM
> -> -------------------------------------------------------------
> -> -------~->
> ->
> ->
> -> Yahoo! Groups Links
> ->
> ->
> ->
> ->
> ->
> ->
>

lp_solve55 error about alloc

2009-12-01T13:13:04Z

HI,

I am trying to use lp_solve5.5 to set up a small MILP model.

It is bounded and feasible. I need to solve it with different A,b,c ever time in a loop.
So, I use make_lp to design a new model every time.

After running some loops, I got the following message " alloc of 10000 'INT' failed ".

but the model is still boudned and feasible and give optimal solutions.

And, if I set the loop times very large, such as 16*24*24, the C program stops and give:
-------------------------------------------------------------------------------------
rank 7 in job 578 server.name caused collective abort of all ranks
exit status of rank 7: killed by signal 9
rank 3 in job 578 server.name caused collective abort of all ranks
exit status of rank 3: return code 1
---------------------------------------------------------------------------

I run the C program in a MPI cluster .

why ?

Any help is appreciated.

thanks

Jack

Dec. 1 2009

---------------------------------

Model name: 'milp_model' - run #1
Objective: Minimize(R0)

SUBMITTED
Model size: 8 constraints, 6 variables, 16 non-zeros.
Sets: 0 GUB, 0 SOS.

Using DUAL simplex for phase 1 and PRIMAL simplex for phase 2.
The primal and dual simplex pricing strategy set to 'Devex'.

alloc of 10000 'INT' failed
alloc of 10000 'INT' failed
alloc of 10000 'INT' failed
alloc of 10000 'REAL' failed
alloc of 10000 'INT' failed

Relaxed solution 237566.262512 after 4 iter is B&B base.

Feasible solution 237566.262512 after 4 iter, 0 nodes (gap 0.0%)

Optimal solution 237566.262512 after 4 iter, 0 nodes (gap 0.0%).

Excellent numeric accuracy ||*|| = 8.64864e-13

MEMO: lp_solve version 5.5.0.13 for 32 bit OS, with 64 bit REAL variables.
In the total iteration count 4, 0 (0.0%) were bound flips.
There were 0 refactorizations, 0 triggered by time and 0 by density.
... on average 4.0 major pivots per refactorization.
The largest [LUSOL v2.2.1.0] fact(B) had 9 NZ entries, 1.0x largest basis.
The maximum B&B level was 1, 0.5x MIP order, 1 at the optimal solution.
The constraint matrix inf-norm is 2053.11, with a dynamic range of 2053.11.
Time to load data was 0.006 seconds, presolve used 0.001 seconds,
... 0.009 seconds in simplex solver, in total 0.016 seconds.

-----------------------------------------

_________________________________________________________________
Chat with Messenger straight from your Hotmail inbox.
http://www.microsoft.com/windows/windowslive/hotmail_bl1/hotmail_bl1.aspx?ocid=PID23879::T:WLMTAGL:ON:WL:en-ww:WM_IMHM_4:092009

Re: Multiple Runs

2009-11-25T08:49:46Z

Normally not, but if the model has integer/binary/semi-cont/SOS variables then I would advise to do that.

Peter

From: relic4112003
Sent: Wednesday, November 25, 2009 17:16
To: lp_solve@...
Subject: [lp_solve] Multiple Runs

In my program I perform the following steps:

1 - Load LP from file
2 - Solve LP
3 - Over multiple iterations, I change the upper bounds on the variables for the LP.

Do I need to reset the LP or or do anything special to ensure that everything is being solved correctly?

Thanks.

Multiple Runs

2009-11-25T08:16:42Z

In my program I perform the following steps:

1 - Load LP from file
2 - Solve LP
3 - Over multiple iterations, I change the upper bounds on the variables for the LP.

Do I need to reset the LP or or do anything special to ensure that everything is being solved correctly?

Thanks.

Fwd: Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

2009-11-25T07:07:28Z

For some reason this didn't seem to arrive the first time I sent it.

Begin forwarded message:

> From: Konis Kjell <kjell.konis@...>
> Date: 25 novembre 2009 15:25:52 HNEC
> To: "lp_solve@..." <lp_solve@...>
> Subject: Re: [lp_solve] Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0
>
> The underlying lp_solve used in the lpSolve R package is 5.5.0.7 (at least in the current CRAN package). You can check with the command
>
>> .C("lp_solve_version", major = integer(1), minor = integer(1), release = integer(1), build = integer(1))
>
> when the package is loaded.
>
> In case it might help: there is also the lpSolveAPI package (that I maintain) which mimics the lp_solve C API in R. It contains the current 5.5.0.15 version of lp_solve.
>
> Cheers,
> Kjell

> On 25 nov. 2009, at 14:59, Loris Bennett wrote:
>
>> "Peter Notebaert" <_peno_@...> writes:
>>
>>> Have you used the ccc.osx compilation script? And what version of
>>> lp_solve are you talking about? You should use the latest 5.5.0.15
>>> version.
>>
>> Isn't ccc.osx for Mac OS X? I am running AIX, IBM's unix.
>>
>> I am also not trying to compile lpSolve directly but rather the
>> lpSolve package for R, which contains lpSolve.
>>
>> I mentioned in my posting that the version of lpSolve packaged for R
>> seems to be 5.1.0. Looking again I see that it is at least 5.2. On the
>> CRAN site (http://cran.r-project.org/web/packages/lpSolve/index.html)
>> the lpSolve version is given as 5.5, but I believe this to be incorrect.
>>
>> I have had a look at version 5.5.0 of lpSolve and see that in lib_lp.h
>>
>> #if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
>> # define _isnan(x) FALSE
>> #endif
>>
>> has changed to
>>
>> #if defined _WIN32 && !defined __GNUC__
>> # define _isnan(x) FALSE
>> #endif
>>
>> This compiles fine on my AIX using the IBM compiler xlc.
>>
>> Thanks,
>>
>> Loris
>>
>>> Peter
>>>
>>> From: Loris Bennett
>>> Sent: Wednesday, November 25, 2009 09:15
>>> To: lp_solve@...
>>> Subject: [lp_solve] Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0
>>>
>>>
>>>
>>> Hi,
>>>
>>> I have been trying to install lpsolve on AIX 5.3 with xlc 9.0 for R
>>> 2.10.0. The version seems to be v5.1.0. The error I get is
>>>
>>> "/usr/include/math.h", line 1030.18: 1506-275 (S) Unexpected
>>> text integer constant encountered.
>>>
>>> The problem seems to be in lp_lib.h:
>>>
>>> #if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
>>> # define _isnan(x) FALSE
>>> #endif
>>>
>>> The function _isnan() is declared in the system math.h, so if this gets
>>> included after lp_lib.h the error will occur. Commenting the define out
>>> allows the package to be compiled.
>>>
>>> Is this redefinition of _isnan needed? Will I be OK if I just comment it
>>> out?
>>>
>>> Regards
>>>
>>> Loris Bennett
>>> --
>>> Dr. Loris Bennett
>>> Computer Centre
>>> Freie Universität Berlin
>>> Berlin, Germany
>>>
>>>
>>>
>>>
>>>
>>
>> --
>> Dr. Loris Bennett
>> Computer Centre
>> Freie Universität Berlin
>> Berlin, Germany
>>
>>
>

------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/lp_solve/

<*> Your email settings:
Individual Email | Traditional

<*> To change settings online go to:
http://groups.yahoo.com/group/lp_solve/join
(Yahoo! ID required)

<*> To change settings via email:
lp_solve-digest@...
lp_solve-fullfeatured@...

<*> To unsubscribe from this group, send an email to:
lp_solve-unsubscribe@...

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

Re: Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

2009-11-25T06:25:52Z

The underlying lp_solve used in the lpSolve R package is 5.5.0.7 (at least in the current CRAN package). You can check with the command

> .C("lp_solve_version", major = integer(1), minor = integer(1), release = integer(1), build = integer(1))

when the package is loaded.

In case it might help: there is also the lpSolveAPI package (that I maintain) which mimics the lp_solve C API in R. It contains the current 5.5.0.15 version of lp_solve.

Cheers,
Kjell

On 25 nov. 2009, at 14:59, Loris Bennett wrote:

> "Peter Notebaert" <_peno_@...> writes:
>
> > Have you used the ccc.osx compilation script? And what version of
> > lp_solve are you talking about? You should use the latest 5.5.0.15
> > version.
>
> Isn't ccc.osx for Mac OS X? I am running AIX, IBM's unix.
>
> I am also not trying to compile lpSolve directly but rather the
> lpSolve package for R, which contains lpSolve.
>
> I mentioned in my posting that the version of lpSolve packaged for R
> seems to be 5.1.0. Looking again I see that it is at least 5.2. On the
> CRAN site (http://cran.r-project.org/web/packages/lpSolve/index.html)
> the lpSolve version is given as 5.5, but I believe this to be incorrect.
>
> I have had a look at version 5.5.0 of lpSolve and see that in lib_lp.h
>
> #if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
> # define _isnan(x) FALSE
> #endif
>
> has changed to
>
> #if defined _WIN32 && !defined __GNUC__
> # define _isnan(x) FALSE
> #endif
>
> This compiles fine on my AIX using the IBM compiler xlc.
>
> Thanks,
>
> Loris
>
> > Peter
> >
> > From: Loris Bennett
> > Sent: Wednesday, November 25, 2009 09:15
> > To: lp_solve@...
> > Subject: [lp_solve] Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0
> >
> >
> >
> > Hi,
> >
> > I have been trying to install lpsolve on AIX 5.3 with xlc 9.0 for R
> > 2.10.0. The version seems to be v5.1.0. The error I get is
> >
> > "/usr/include/math.h", line 1030.18: 1506-275 (S) Unexpected
> > text integer constant encountered.
> >
> > The problem seems to be in lp_lib.h:
> >
> > #if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
> > # define _isnan(x) FALSE
> > #endif
> >
> > The function _isnan() is declared in the system math.h, so if this gets
> > included after lp_lib.h the error will occur. Commenting the define out
> > allows the package to be compiled.
> >
> > Is this redefinition of _isnan needed? Will I be OK if I just comment it
> > out?
> >
> > Regards
> >
> > Loris Bennett
> > --
> > Dr. Loris Bennett
> > Computer Centre
> > Freie Universität Berlin
> > Berlin, Germany
> >
> >
> >
> >
> >
>
> --
> Dr. Loris Bennett
> Computer Centre
> Freie Universität Berlin
> Berlin, Germany
>
>

Re: Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

2009-11-25T06:24:26Z

Sorry, I ment ccc.aix

The lpsolve link to R is made by someone else not contributing his code the group...
You should contact him.

Peter

From: Loris Bennett
Sent: Wednesday, November 25, 2009 14:59
To: lp_solve@...
Subject: [lp_solve] Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

"Peter Notebaert" <_peno_@...> writes:

> Have you used the ccc.osx compilation script? And what version of
> lp_solve are you talking about? You should use the latest 5.5.0.15
> version.

Isn't ccc.osx for Mac OS X? I am running AIX, IBM's unix.

I am also not trying to compile lpSolve directly but rather the
lpSolve package for R, which contains lpSolve.

I mentioned in my posting that the version of lpSolve packaged for R
seems to be 5.1.0. Looking again I see that it is at least 5.2. On the
CRAN site (http://cran.r-project.org/web/packages/lpSolve/index.html)
the lpSolve version is given as 5.5, but I believe this to be incorrect.

I have had a look at version 5.5.0 of lpSolve and see that in lib_lp.h

#if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
# define _isnan(x) FALSE
#endif

has changed to

#if defined _WIN32 && !defined __GNUC__
# define _isnan(x) FALSE
#endif

This compiles fine on my AIX using the IBM compiler xlc.

Thanks,

Loris

> Peter
>
> From: Loris Bennett
> Sent: Wednesday, November 25, 2009 09:15
> To: lp_solve@...
> Subject: [lp_solve] Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0
>
>
>
> Hi,
>
> I have been trying to install lpsolve on AIX 5.3 with xlc 9.0 for R
> 2.10.0. The version seems to be v5.1.0. The error I get is
>
> "/usr/include/math.h", line 1030.18: 1506-275 (S) Unexpected
> text integer constant encountered.
>
> The problem seems to be in lp_lib.h:
>
> #if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
> # define _isnan(x) FALSE
> #endif
>
> The function _isnan() is declared in the system math.h, so if this gets
> included after lp_lib.h the error will occur. Commenting the define out
> allows the package to be compiled.
>
> Is this redefinition of _isnan needed? Will I be OK if I just comment it
> out?
>
> Regards
>
> Loris Bennett
> --
> Dr. Loris Bennett
> Computer Centre
> Freie Universität Berlin
> Berlin, Germany
>
>
>
>
>

--
Dr. Loris Bennett
Computer Centre
Freie Universität Berlin
Berlin, Germany

Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

2009-11-25T05:59:11Z

"Peter Notebaert" <_peno_@...> writes:

> Have you used the ccc.osx compilation script? And what version of
> lp_solve are you talking about? You should use the latest 5.5.0.15
> version.

Isn't ccc.osx for Mac OS X? I am running AIX, IBM's unix.

I am also not trying to compile lpSolve directly but rather the
lpSolve package for R, which contains lpSolve.

I mentioned in my posting that the version of lpSolve packaged for R
seems to be 5.1.0. Looking again I see that it is at least 5.2. On the
CRAN site (http://cran.r-project.org/web/packages/lpSolve/index.html)
the lpSolve version is given as 5.5, but I believe this to be incorrect.

I have had a look at version 5.5.0 of lpSolve and see that in lib_lp.h

#if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
# define _isnan(x) FALSE
#endif

has changed to

#if defined _WIN32 && !defined __GNUC__
# define _isnan(x) FALSE
#endif

This compiles fine on my AIX using the IBM compiler xlc.

Thanks,

Loris

--
Dr. Loris Bennett
Computer Centre
Freie Universität Berlin
Berlin, Germany

Re: LP_Solve not giving me the right answer?

2009-11-25T03:02:13Z

That fixed it for me. Thanks for explaining it too - I was going crazy trying to figure out the reason. :)

--- In lp_solve@..., "peter_notebaert" <_peno_@...> wrote:

>
> That is a result of the peculiarities of the lp-format.
>
> The 'constraints'
>
> S2_1 <= 0;
> S2_0 <= 0;
> S3_0 <= 0;
> S4_1 <= 0;
> S4_0 <= 0;
>
> Are really seen as bounds on variables. This because there is only one variable in it and there is no label before it.
>
> So the solver puts an upper bound on 0 on these variables. It does not make equations of these.
>
> Ben then later on you specify the variables as bin.
>
> And bin means that they are integer AND an upper bound of 1.
>
> So the previous upper bound of 0 is overruled by the upper bound of 1 by the bin statement.
>
> To solve this, put a label before your constraints to make sure they are interpreted as constraint and not as bound:
>
> r1: S2_1 <= 0;
> r2: S2_0 <= 0;
> r3: S3_0 <= 0;
> r4: S4_1 <= 0;
> r5: S4_0 <= 0;
>
> Then you will have the soluution that you expect.
>
> You could have seen this also by using the lp_solve command line command to solve the model and write the model back to a file with the -wlp option:
>
> lp_solve model.lp -wlp model1.lp
>
> You would have seen that model1.lp did not have your constraints anymore.
>
> Peter
>
> --- In lp_solve@..., "nrclark1" <nicholas.clark@> wrote:
> >
> > Hi all,
> >
> > Sorry if this double-posted! I'm a newbie to the world of LP, currently working on a task-scheduling problem. I made an LP_Solve model which expresses all of my constraints (I think!), but LP_Solve seems to ignore one or more of them.
> >
> > In my example below, S is a matrix representing the start times of each task (tasks are in rows, timesteps are columns).
> >
> > My model parses OK, but the solution that LP_Solve presents directly violates the constraint that I set on line 17 (S4_0 <= 0). Did I make some kind of weird syntax mistake that I can't find? Your help would be greatly appreciated. :)
> >
> > --------------------------------------------
> >
> > /* Minimization Goal */
> > min: ;
> >
> > /* Uniqueness Constraints */
> > S1_0 + S1_1 + S1_2 + S1_3 + S1_4 + S1_5 + S1_6 + S1_7 + S1_8 + S1_9 = 1;
> > S2_0 + S2_1 + S2_2 + S2_3 + S2_4 + S2_5 + S2_6 + S2_7 + S2_8 + S2_9 = 1;
> > S3_0 + S3_1 + S3_2 + S3_3 + S3_4 + S3_5 + S3_6 + S3_7 + S3_8 + S3_9 = 1;
> > S4_0 + S4_1 + S4_2 + S4_3 + S4_4 + S4_5 + S4_6 + S4_7 + S4_8 + S4_9 = 1;
> > S5_0 + S5_1 + S5_2 + S5_3 + S5_4 + S5_5 + S5_6 + S5_7 + S5_8 + S5_9 = 1;
> > S6_0 + S6_1 + S6_2 + S6_3 + S6_4 + S6_5 + S6_6 + S6_7 + S6_8 + S6_9 = 1;
> >
> > S2_1 <= 0;
> > S2_0 <= 0;
> > S3_0 <= 0;
> > S4_1 <= 0;
> > S4_0 <= 0;
> >
> >
> > /* Dependency Constraints */
> > S2_9 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 - S1_6 - S1_7 <= 0;
> > S2_9 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 - S5_6 - S5_7 <= 0;
> > S2_8 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 - S1_6 <= 0;
> > S2_8 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 - S5_6 <= 0;
> > S2_7 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 <= 0;
> > S2_7 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 <= 0;
> > S2_6 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 <= 0;
> > S2_6 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 <= 0;
> > S2_5 - S1_0 - S1_1 - S1_2 - S1_3 <= 0;
> > S2_5 - S5_0 - S5_1 - S5_2 - S5_3 <= 0;
> > S2_4 - S1_0 - S1_1 - S1_2 <= 0;
> > S2_4 - S5_0 - S5_1 - S5_2 <= 0;
> > S2_3 - S1_0 - S1_1 <= 0;
> > S2_3 - S5_0 - S5_1 <= 0;
> > S2_2 - S1_0 <= 0;
> > S2_2 - S5_0 <= 0;
> >
> > S3_9 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 - S2_7 - S2_8 <= 0;
> > S3_8 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 - S2_7 <= 0;
> > S3_7 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 <= 0;
> > S3_6 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 <= 0;
> > S3_5 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 <= 0;
> > S3_4 - S2_0 - S2_1 - S2_2 - S2_3 <= 0;
> > S3_3 - S2_0 - S2_1 - S2_2 <= 0;
> > S3_2 - S2_0 - S2_1 <= 0;
> > S3_1 - S2_0 <= 0;
> >
> > S4_9 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 - S3_7 - S3_8 <= 0;
> > S4_9 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 - S6_6 - S6_7 <= 0;
> > S4_8 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 - S3_7 <= 0;
> > S4_8 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 - S6_6 <= 0;
> > S4_7 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 <= 0;
> > S4_7 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 <= 0;
> > S4_6 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 <= 0;
> > S4_6 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 <= 0;
> > S4_5 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 <= 0;
> > S4_5 - S6_0 - S6_1 - S6_2 - S6_3 <= 0;
> > S4_4 - S3_0 - S3_1 - S3_2 - S3_3 <= 0;
> > S4_4 - S6_0 - S6_1 - S6_2 <= 0;
> > S4_3 - S3_0 - S3_1 - S3_2 <= 0;
> > S4_3 - S6_0 - S6_1 <= 0;
> > S4_2 - S3_0 - S3_1 <= 0;
> > S4_2 - S6_0 <= 0;
> > S4_1 - S3_0 <= 0;
> >
> >
> > /* Resource Constraints */
> > S2_0 + S3_0 + S4_0 <= 1;
> > S2_1 + S3_1 + S4_1 <= 1;
> > S2_2 + S3_2 + S4_2 <= 1;
> > S2_3 + S3_3 + S4_3 <= 1;
> > S2_4 + S3_4 + S4_4 <= 1;
> > S2_5 + S3_5 + S4_5 <= 1;
> > S2_6 + S3_6 + S4_6 <= 1;
> > S2_7 + S3_7 + S4_7 <= 1;
> > S2_8 + S3_8 + S4_8 <= 1;
> > S2_9 + S3_9 + S4_9 <= 1;
> > S1_0 + S5_0 + S6_0 <= 2;
> > S1_0 + S1_1 + S5_0 + S5_1 + S6_0 + S6_1 <= 2;
> > S1_1 + S1_2 + S5_1 + S5_2 + S6_1 + S6_2 <= 2;
> > S1_2 + S1_3 + S5_2 + S5_3 + S6_2 + S6_3 <= 2;
> > S1_3 + S1_4 + S5_3 + S5_4 + S6_3 + S6_4 <= 2;
> > S1_4 + S1_5 + S5_4 + S5_5 + S6_4 + S6_5 <= 2;
> > S1_5 + S1_6 + S5_5 + S5_6 + S6_5 + S6_6 <= 2;
> > S1_6 + S1_7 + S5_6 + S5_7 + S6_6 + S6_7 <= 2;
> > S1_7 + S1_8 + S5_7 + S5_8 + S6_7 + S6_8 <= 2;
> > S1_8 + S1_9 + S5_8 + S5_9 + S6_8 + S6_9 <= 2;
> >
> > /* Variable Types */
> > bin S1_0, S1_1, S1_2, S1_3, S1_4, S1_5, S1_6, S1_7, S1_8, S1_9;
> > bin S2_0, S2_1, S2_2, S2_3, S2_4, S2_5, S2_6, S2_7, S2_8, S2_9;
> > bin S3_0, S3_1, S3_2, S3_3, S3_4, S3_5, S3_6, S3_7, S3_8, S3_9;
> > bin S4_0, S4_1, S4_2, S4_3, S4_4, S4_5, S4_6, S4_7, S4_8, S4_9;
> > bin S5_0, S5_1, S5_2, S5_3, S5_4, S5_5, S5_6, S5_7, S5_8, S5_9;
> > bin S6_0, S6_1, S6_2, S6_3, S6_4, S6_5, S6_6, S6_7, S6_8, S6_9;
> >
>

Re: LP_Solve not giving me the right answer?

2009-11-25T02:49:43Z

That is a result of the peculiarities of the lp-format.

The 'constraints'

S2_1 <= 0;
S2_0 <= 0;
S3_0 <= 0;
S4_1 <= 0;
S4_0 <= 0;

Are really seen as bounds on variables. This because there is only one variable in it and there is no label before it.

So the solver puts an upper bound on 0 on these variables. It does not make equations of these.

Ben then later on you specify the variables as bin.

And bin means that they are integer AND an upper bound of 1.

So the previous upper bound of 0 is overruled by the upper bound of 1 by the bin statement.

To solve this, put a label before your constraints to make sure they are interpreted as constraint and not as bound:

r1: S2_1 <= 0;
r2: S2_0 <= 0;
r3: S3_0 <= 0;
r4: S4_1 <= 0;
r5: S4_0 <= 0;

Then you will have the soluution that you expect.

You could have seen this also by using the lp_solve command line command to solve the model and write the model back to a file with the -wlp option:

lp_solve model.lp -wlp model1.lp

You would have seen that model1.lp did not have your constraints anymore.

Peter

--- In lp_solve@..., "nrclark1" <nicholas.clark@...> wrote:

>
> Hi all,
>
> Sorry if this double-posted! I'm a newbie to the world of LP, currently working on a task-scheduling problem. I made an LP_Solve model which expresses all of my constraints (I think!), but LP_Solve seems to ignore one or more of them.
>
> In my example below, S is a matrix representing the start times of each task (tasks are in rows, timesteps are columns).
>
> My model parses OK, but the solution that LP_Solve presents directly violates the constraint that I set on line 17 (S4_0 <= 0). Did I make some kind of weird syntax mistake that I can't find? Your help would be greatly appreciated. :)
>
> --------------------------------------------
>
> /* Minimization Goal */
> min: ;
>
> /* Uniqueness Constraints */
> S1_0 + S1_1 + S1_2 + S1_3 + S1_4 + S1_5 + S1_6 + S1_7 + S1_8 + S1_9 = 1;
> S2_0 + S2_1 + S2_2 + S2_3 + S2_4 + S2_5 + S2_6 + S2_7 + S2_8 + S2_9 = 1;
> S3_0 + S3_1 + S3_2 + S3_3 + S3_4 + S3_5 + S3_6 + S3_7 + S3_8 + S3_9 = 1;
> S4_0 + S4_1 + S4_2 + S4_3 + S4_4 + S4_5 + S4_6 + S4_7 + S4_8 + S4_9 = 1;
> S5_0 + S5_1 + S5_2 + S5_3 + S5_4 + S5_5 + S5_6 + S5_7 + S5_8 + S5_9 = 1;
> S6_0 + S6_1 + S6_2 + S6_3 + S6_4 + S6_5 + S6_6 + S6_7 + S6_8 + S6_9 = 1;
>
> S2_1 <= 0;
> S2_0 <= 0;
> S3_0 <= 0;
> S4_1 <= 0;
> S4_0 <= 0;
>
>
> /* Dependency Constraints */
> S2_9 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 - S1_6 - S1_7 <= 0;
> S2_9 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 - S5_6 - S5_7 <= 0;
> S2_8 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 - S1_6 <= 0;
> S2_8 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 - S5_6 <= 0;
> S2_7 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 <= 0;
> S2_7 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 <= 0;
> S2_6 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 <= 0;
> S2_6 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 <= 0;
> S2_5 - S1_0 - S1_1 - S1_2 - S1_3 <= 0;
> S2_5 - S5_0 - S5_1 - S5_2 - S5_3 <= 0;
> S2_4 - S1_0 - S1_1 - S1_2 <= 0;
> S2_4 - S5_0 - S5_1 - S5_2 <= 0;
> S2_3 - S1_0 - S1_1 <= 0;
> S2_3 - S5_0 - S5_1 <= 0;
> S2_2 - S1_0 <= 0;
> S2_2 - S5_0 <= 0;
>
> S3_9 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 - S2_7 - S2_8 <= 0;
> S3_8 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 - S2_7 <= 0;
> S3_7 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 <= 0;
> S3_6 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 <= 0;
> S3_5 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 <= 0;
> S3_4 - S2_0 - S2_1 - S2_2 - S2_3 <= 0;
> S3_3 - S2_0 - S2_1 - S2_2 <= 0;
> S3_2 - S2_0 - S2_1 <= 0;
> S3_1 - S2_0 <= 0;
>
> S4_9 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 - S3_7 - S3_8 <= 0;
> S4_9 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 - S6_6 - S6_7 <= 0;
> S4_8 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 - S3_7 <= 0;
> S4_8 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 - S6_6 <= 0;
> S4_7 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 <= 0;
> S4_7 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 <= 0;
> S4_6 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 <= 0;
> S4_6 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 <= 0;
> S4_5 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 <= 0;
> S4_5 - S6_0 - S6_1 - S6_2 - S6_3 <= 0;
> S4_4 - S3_0 - S3_1 - S3_2 - S3_3 <= 0;
> S4_4 - S6_0 - S6_1 - S6_2 <= 0;
> S4_3 - S3_0 - S3_1 - S3_2 <= 0;
> S4_3 - S6_0 - S6_1 <= 0;
> S4_2 - S3_0 - S3_1 <= 0;
> S4_2 - S6_0 <= 0;
> S4_1 - S3_0 <= 0;
>
>
> /* Resource Constraints */
> S2_0 + S3_0 + S4_0 <= 1;
> S2_1 + S3_1 + S4_1 <= 1;
> S2_2 + S3_2 + S4_2 <= 1;
> S2_3 + S3_3 + S4_3 <= 1;
> S2_4 + S3_4 + S4_4 <= 1;
> S2_5 + S3_5 + S4_5 <= 1;
> S2_6 + S3_6 + S4_6 <= 1;
> S2_7 + S3_7 + S4_7 <= 1;
> S2_8 + S3_8 + S4_8 <= 1;
> S2_9 + S3_9 + S4_9 <= 1;
> S1_0 + S5_0 + S6_0 <= 2;
> S1_0 + S1_1 + S5_0 + S5_1 + S6_0 + S6_1 <= 2;
> S1_1 + S1_2 + S5_1 + S5_2 + S6_1 + S6_2 <= 2;
> S1_2 + S1_3 + S5_2 + S5_3 + S6_2 + S6_3 <= 2;
> S1_3 + S1_4 + S5_3 + S5_4 + S6_3 + S6_4 <= 2;
> S1_4 + S1_5 + S5_4 + S5_5 + S6_4 + S6_5 <= 2;
> S1_5 + S1_6 + S5_5 + S5_6 + S6_5 + S6_6 <= 2;
> S1_6 + S1_7 + S5_6 + S5_7 + S6_6 + S6_7 <= 2;
> S1_7 + S1_8 + S5_7 + S5_8 + S6_7 + S6_8 <= 2;
> S1_8 + S1_9 + S5_8 + S5_9 + S6_8 + S6_9 <= 2;
>
> /* Variable Types */
> bin S1_0, S1_1, S1_2, S1_3, S1_4, S1_5, S1_6, S1_7, S1_8, S1_9;
> bin S2_0, S2_1, S2_2, S2_3, S2_4, S2_5, S2_6, S2_7, S2_8, S2_9;
> bin S3_0, S3_1, S3_2, S3_3, S3_4, S3_5, S3_6, S3_7, S3_8, S3_9;
> bin S4_0, S4_1, S4_2, S4_3, S4_4, S4_5, S4_6, S4_7, S4_8, S4_9;
> bin S5_0, S5_1, S5_2, S5_3, S5_4, S5_5, S5_6, S5_7, S5_8, S5_9;
> bin S6_0, S6_1, S6_2, S6_3, S6_4, S6_5, S6_6, S6_7, S6_8, S6_9;
>

LP_Solve not giving me the right answer?

2009-11-25T02:32:04Z

Hi all,

Sorry if this double-posted! I'm a newbie to the world of LP, currently working on a task-scheduling problem. I made an LP_Solve model which expresses all of my constraints (I think!), but LP_Solve seems to ignore one or more of them.

In my example below, S is a matrix representing the start times of each task (tasks are in rows, timesteps are columns).

My model parses OK, but the solution that LP_Solve presents directly violates the constraint that I set on line 17 (S4_0 <= 0). Did I make some kind of weird syntax mistake that I can't find? Your help would be greatly appreciated. :)

--------------------------------------------

/* Minimization Goal */
min: ;

/* Uniqueness Constraints */
S1_0 + S1_1 + S1_2 + S1_3 + S1_4 + S1_5 + S1_6 + S1_7 + S1_8 + S1_9 = 1;
S2_0 + S2_1 + S2_2 + S2_3 + S2_4 + S2_5 + S2_6 + S2_7 + S2_8 + S2_9 = 1;
S3_0 + S3_1 + S3_2 + S3_3 + S3_4 + S3_5 + S3_6 + S3_7 + S3_8 + S3_9 = 1;
S4_0 + S4_1 + S4_2 + S4_3 + S4_4 + S4_5 + S4_6 + S4_7 + S4_8 + S4_9 = 1;
S5_0 + S5_1 + S5_2 + S5_3 + S5_4 + S5_5 + S5_6 + S5_7 + S5_8 + S5_9 = 1;
S6_0 + S6_1 + S6_2 + S6_3 + S6_4 + S6_5 + S6_6 + S6_7 + S6_8 + S6_9 = 1;

S2_1 <= 0;
S2_0 <= 0;
S3_0 <= 0;
S4_1 <= 0;
S4_0 <= 0;

/* Dependency Constraints */
S2_9 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 - S1_6 - S1_7 <= 0;
S2_9 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 - S5_6 - S5_7 <= 0;
S2_8 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 - S1_6 <= 0;
S2_8 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 - S5_6 <= 0;
S2_7 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 - S1_5 <= 0;
S2_7 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 - S5_5 <= 0;
S2_6 - S1_0 - S1_1 - S1_2 - S1_3 - S1_4 <= 0;
S2_6 - S5_0 - S5_1 - S5_2 - S5_3 - S5_4 <= 0;
S2_5 - S1_0 - S1_1 - S1_2 - S1_3 <= 0;
S2_5 - S5_0 - S5_1 - S5_2 - S5_3 <= 0;
S2_4 - S1_0 - S1_1 - S1_2 <= 0;
S2_4 - S5_0 - S5_1 - S5_2 <= 0;
S2_3 - S1_0 - S1_1 <= 0;
S2_3 - S5_0 - S5_1 <= 0;
S2_2 - S1_0 <= 0;
S2_2 - S5_0 <= 0;

S3_9 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 - S2_7 - S2_8 <= 0;
S3_8 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 - S2_7 <= 0;
S3_7 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 - S2_6 <= 0;
S3_6 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 - S2_5 <= 0;
S3_5 - S2_0 - S2_1 - S2_2 - S2_3 - S2_4 <= 0;
S3_4 - S2_0 - S2_1 - S2_2 - S2_3 <= 0;
S3_3 - S2_0 - S2_1 - S2_2 <= 0;
S3_2 - S2_0 - S2_1 <= 0;
S3_1 - S2_0 <= 0;

S4_9 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 - S3_7 - S3_8 <= 0;
S4_9 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 - S6_6 - S6_7 <= 0;
S4_8 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 - S3_7 <= 0;
S4_8 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 - S6_6 <= 0;
S4_7 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 - S3_6 <= 0;
S4_7 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 - S6_5 <= 0;
S4_6 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 - S3_5 <= 0;
S4_6 - S6_0 - S6_1 - S6_2 - S6_3 - S6_4 <= 0;
S4_5 - S3_0 - S3_1 - S3_2 - S3_3 - S3_4 <= 0;
S4_5 - S6_0 - S6_1 - S6_2 - S6_3 <= 0;
S4_4 - S3_0 - S3_1 - S3_2 - S3_3 <= 0;
S4_4 - S6_0 - S6_1 - S6_2 <= 0;
S4_3 - S3_0 - S3_1 - S3_2 <= 0;
S4_3 - S6_0 - S6_1 <= 0;
S4_2 - S3_0 - S3_1 <= 0;
S4_2 - S6_0 <= 0;
S4_1 - S3_0 <= 0;

/* Resource Constraints */
S2_0 + S3_0 + S4_0 <= 1;
S2_1 + S3_1 + S4_1 <= 1;
S2_2 + S3_2 + S4_2 <= 1;
S2_3 + S3_3 + S4_3 <= 1;
S2_4 + S3_4 + S4_4 <= 1;
S2_5 + S3_5 + S4_5 <= 1;
S2_6 + S3_6 + S4_6 <= 1;
S2_7 + S3_7 + S4_7 <= 1;
S2_8 + S3_8 + S4_8 <= 1;
S2_9 + S3_9 + S4_9 <= 1;
S1_0 + S5_0 + S6_0 <= 2;
S1_0 + S1_1 + S5_0 + S5_1 + S6_0 + S6_1 <= 2;
S1_1 + S1_2 + S5_1 + S5_2 + S6_1 + S6_2 <= 2;
S1_2 + S1_3 + S5_2 + S5_3 + S6_2 + S6_3 <= 2;
S1_3 + S1_4 + S5_3 + S5_4 + S6_3 + S6_4 <= 2;
S1_4 + S1_5 + S5_4 + S5_5 + S6_4 + S6_5 <= 2;
S1_5 + S1_6 + S5_5 + S5_6 + S6_5 + S6_6 <= 2;
S1_6 + S1_7 + S5_6 + S5_7 + S6_6 + S6_7 <= 2;
S1_7 + S1_8 + S5_7 + S5_8 + S6_7 + S6_8 <= 2;
S1_8 + S1_9 + S5_8 + S5_9 + S6_8 + S6_9 <= 2;

/* Variable Types */
bin S1_0, S1_1, S1_2, S1_3, S1_4, S1_5, S1_6, S1_7, S1_8, S1_9;
bin S2_0, S2_1, S2_2, S2_3, S2_4, S2_5, S2_6, S2_7, S2_8, S2_9;
bin S3_0, S3_1, S3_2, S3_3, S3_4, S3_5, S3_6, S3_7, S3_8, S3_9;
bin S4_0, S4_1, S4_2, S4_3, S4_4, S4_5, S4_6, S4_7, S4_8, S4_9;
bin S5_0, S5_1, S5_2, S5_3, S5_4, S5_5, S5_6, S5_7, S5_8, S5_9;
bin S6_0, S6_1, S6_2, S6_3, S6_4, S6_5, S6_6, S6_7, S6_8, S6_9;

Re: Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

2009-11-25T02:13:28Z

Have you used the ccc.osx compilation script? And what version of lp_solve are you talking about? You should use the latest 5.5.0.15 version.

Peter

From: Loris Bennett
Sent: Wednesday, November 25, 2009 09:15
To: lp_solve@...
Subject: [lp_solve] Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

Hi,

I have been trying to install lpsolve on AIX 5.3 with xlc 9.0 for R
2.10.0. The version seems to be v5.1.0. The error I get is

"/usr/include/math.h", line 1030.18: 1506-275 (S) Unexpected
text integer constant encountered.

The problem seems to be in lp_lib.h:

#if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
# define _isnan(x) FALSE
#endif

The function _isnan() is declared in the system math.h, so if this gets
included after lp_lib.h the error will occur. Commenting the define out
allows the package to be compiled.

Is this redefinition of _isnan needed? Will I be OK if I just comment it
out?

Regards

Loris Bennett
--
Dr. Loris Bennett
Computer Centre
Freie Universität Berlin
Berlin, Germany

Error compiling on AIX 5.3 with xlc 9.0 for R 2.10.0

2009-11-25T00:15:47Z

Hi,

I have been trying to install lpsolve on AIX 5.3 with xlc 9.0 for R
2.10.0. The version seems to be v5.1.0. The error I get is

"/usr/include/math.h", line 1030.18: 1506-275 (S) Unexpected
text integer constant encountered.

The problem seems to be in lp_lib.h:

#if !defined _WINDOWS && !defined _WIN32 && !defined WIN32
# define _isnan(x) FALSE
#endif

The function _isnan() is declared in the system math.h, so if this gets
included after lp_lib.h the error will occur. Commenting the define out
allows the package to be compiled.

Is this redefinition of _isnan needed? Will I be OK if I just comment it
out?

Regards

Loris Bennett
--
Dr. Loris Bennett
Computer Centre
Freie Universität Berlin
Berlin, Germany

Re: calling lp-solve using c in visual studio 2005

2009-11-24T05:20:14Z

I hate it when people stay asking the same questions over and over again if there is a reference guide AND examples describing and showing all this.

Peter

--- In lp_solve@..., "marwan_abb" <marwan_abb@...> wrote:
>
> how can i call lp-solve.exe using c in VS 2005, so i can send variables and get back results, what are the nesscary commands in c that i must write, and what are the libraries that i must add.
> thanks a lot
>

calling lp-solve using c in visual studio 2005

2009-11-24T05:00:36Z

how can i call lp-solve.exe using c in VS 2005, so i can send variables and get back results, what are the nesscary commands in c that i must write, and what are the libraries that i must add.
thanks a lot