, it is sufficient to
enclose the corresponding fragment within #ifdef/#endif and add
necessary -D options in the makefile.
Btw, dynamic loading/unloading can be performed in DllMain; see
file glplib03.c; probably it is a most appropraite place for that.
Andrew Makhorin
It would be nice to have this change
I have a data with 3 indices c[i,j,k]. How can I write it in a MathProg
format? I would like to fix k=1, give the matrix c[i,j,1], fix k=2, give
the matrix c[i,j,2], and go on...
Like this:
param c :=
[*,*,1] 1 1 2.3
1 2 4.5
. . .
[*,*,2] 1 1 5.6
1 2 7.8
. . .
to look through the code; probably it will take two or
three days.
Could you please consider changes suggested by Vijay?
Best regards,
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
Does anyone know of a theoretical size limit for a non-integer GLPK
problem?
Theoretically the problem size is limited by 100,000,000 rows and
columns and 500,000,000 non-zero constraint coefficients; this is
because glpk is a 32-bit application in the sense that it uses 32-bit
integers to
I cannot run it, because
my machine has only 256Mb of RAM.)
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
I produced a DLL as well by adding __stdcall __export to the right
function calls. [BTW, I'm using the free Turbo C++ compiler from
Borland.]
BUT, I cannot seem to get the callback function for glp_term_hook to
properly work. Before I begin to setup the problem, I use very similar
VBA
if an error occurs in the GLPK library xfault is called and
executes abort(). This is permissible in GLPSOL.
In other programs using the library this behavior causes trouble:
- The calling program cannot release objects and libraries.
- The user of a GUI program cannot save his prior work.
the problem is due to GLPK being compiled with __cdecl calling
convention instead of __stdcall calling convention.
References:
http://www.geocities.com/yongweiwu/stdcall.htm
http://msdn2.microsoft.com/en-us/library/zxk0tw93.aspx
I have recompiled the DLL with compiler switch /Gz
Are there other manuals on GMPL?
You can try to find some educational materials in the internet.
I could not find out in the GMPL
manual how to get the number of constraints, parameters, variables
etc. read into the model. Also, I could not see how it may be possible
to create intermediary
I could not identify any __stdcall Windows API function that would
allow catching abort() as raised by xfault(). Without such a function
I will not be able to stop the termination of Excel or any VBA
applications using the GLPK library. This is why I am requesting the
hook.
There is the
I can understand that You want to keep the core of GLPK independent
of operating system considerations.
Writing a wrapper seems easy to do. Care has to be given to
glp_term_hook because the hook function will be __stdcall. The wrapper
will have to alloc memory to build a info structure and
Benchmark.mod may be included in glpk's examples if you think that
challenging glpk is interesting.
I have added your model to glpk examples changing its name to huge.mod.
___
Help-glpk mailing list
Help-glpk@gnu.org
when writing a console program in C signal() would be fine.
Unfortunately this function does not exist in other languages like
Visual Basic. Maybe this could be put into the same wrapper needed to
provide __stdcall calling convention.
In principle, if the application program does not pass
My LP problem is one with approximately 400K columns and 400K
rows--of which, the matrix data is almost completely sparse. Based
upon your information Andrew, this is well within the limits of GLPK
(assuming adequate hardware is available). The concept of having to
setup this many columns
*xdlopen(const char *file, int mode);
void *xdlsym(void *handle, const char *name);
int xdlclose(void *handle);
const char *xdlerror(void);
I plan to implement these routines for next several days.
What do you think about that?
Andrew Makhorin
Hopefully the Windows version will be included in 4.28.
I included it, however, at the moment it is disabled.
See the file glplib12.c for details.
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
In general I don't recommend generating instance files like MPS or CPLEX
LP directly. If you don't want to use a modeling language because of
size or other issues, then look at the APIs that solvers have to
directly input an instance. COIN-OR OSI (Open Solver Interface)
provides an API
You generally do this by adding a binary variable (and, of course, this
becomes an integer program rather than a simple linear program, and
therefore more difficult to solve).
For example:
var z, binary;
s.t. condition{f in FOO, b in BAR}: sum{a in BAZ} x[f,b,a] = 5 * z;
Probably correct
Probably correct formulation should be the following:
s.t. condition1{f in FOO, b in BAR}:
sum{a in BAZ} x[f,b,a] = 5 * z;
s.t. condition2{f in FOO, b in BAR}:
sum{a in BAZ} x[f,b,a] = u[f,b,a] * z;
where u[f,b,a] is an upper bound of x[f,b,a].
The second constraint
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
GLPK 4.28 -- Release Information
Release date: Mar 25, 2008
GLPK (GNU Linear Programming Kit) is intended for solving large-scale
linear programming (LP), mixed integer linear programming (MIP), and
other related
I wonder if adding additional constraints of
s.t. condition3{f in FOO, b in BAR, a in BAZ} :
x[f,b,a] = u[f,b,a] * z;
would help in developing an integer solution faster?
If u[f,b,a] is stronger than in lp relaxation, I suppose.
However, in general case, it may significantly depend
I'm trying to build GLPK/Mathprog 4.9 under Visual Studio 2005 for a x64
machine (WinXP Pro x64). My VS2005 Projekt for x86 builds and runs
beautifully. The x64 version also builds, but when I try out the
generated code it crashes with an access violation.
The problem seem to be the setjmp()
if I understand the source code correctly, I can not obtain the number
of iterations having been done by the interior point routine, nor can I
get the path that has been followed, in my program, w/o modifying the
sources.
Am I right?
Yes, currently there is no exit routine to do that.
question: why do you assign penalties to slack variables
initially introduced in the master lp? In your code each slack variable
has a unity column, which itself might be considered as a pattern, so
there would be no difference between slack and pattern variables.
Andrew Makhorin
://citeseer.ist.psu.edu/28583.html
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
I #8217;m experimenting with the integer programming
capabilities of GLPK (Version 4.27).
The function #8220;lpx_intopt #8221; has a parameter that allows for
#8220;all cuts #8221; to be used, which include cover, clique, gomory
and mixed-integer-rounding cuts. But lpx_intopt does not use
Andrew, I do not agree.
With what do you not agree?
I did not say that it is impossible to determine a minimal/irreducible
infeasible system for mip. I only said that this is impractical. It is
always possible to minimize the sum of residuals, however, to obtain
the minimal/irreducible system
i am searching for an expression equal to the mathematicl sign function.
I have a sum in a subject to statement, and i want the returning value of
this sum to be zero if the sum is 0 or to be one if the sum is greater than
0.
Example:
subject to rooms{a in A}: sum{b in B} sign(sum{c in C}
var sx{a in A, b in B}, binary;
/* sx[a,b] is sign(sum{c in C} x[a,b,c]) */
/* in other word, sx[a,b] is logical_or{c in C} x[a,b,c] */
s.t. foo{a in A, b in B}: sum{c in C} x[a,b,c] = card(C) * sx[a,b];
s.t. rooms{a in A}: sum{b in B} sx[a,b] = r;
Incorrect. The first constraint must be
Just a minor note. When using a 'mipgap' value, ios_driver() will
return GLP_ETMLIM if the mipgap tolerance is satisfied. This causes
the message TIME LIMIT EXCEEDED; SEARCH TERMINATED message to print
from within glp_intopt(). Not a big deal, but it was confusing the
first few times I saw
I am a new comer of GLPK and been very satisfied with it! Thank Makhorin
for his great works!
Thank you for your interest in glpk.
From the FAQs (2004?) of the documentation, quadratic programming
solvers seems to be one of the future plan at that time. I wonder can we
see this in the near
Can I do something like this ?
Variables : d, x[i]
Constants : MIN, MAX
Set : I with variable card 1..d
MIN = d = MAX
sum{i in 1..d} x[i]= 5
No, because d is a variable. Probably you should try formulating your
problem as mip.
___
Help-glpk
Yes i need several clauses connected with or. But i do not
understand your idea.
The idea is quite simple.
Let you need to model the following condition:
a1 = x = b1 OR a2 = x = b2 OR ... OR an = x = bn
assuming that regions [a1,b1], [a2,b2], ..., [an,xn] are
pairwise disjoint.
Using
I have written a problem with a lot of binary variables (see the
attached lp matrix). I am working with glpk410 calling the routines
by VB for Excel as follows:
lp = glpk_read_cpxlp(Trim(DireccionOPT probname))
Call glpk_set_real_parm(lp, 313, 2)
'Optimizar
I have installed glpk-4.28 and cant seem to install glpkmex in
MATLAB and i keep getting this error:
makeglpkmex
GLPKMEX - A Matlab interface for GLPK. Script installer.
Version 2.4 compatible with GLPK 4.20 (or higher)
(C) 2001-2007, Nicolo #39; Giorgetti.
Do you want to use
I spent long time to fix this error, but still failed.
spx_invert: the basis matrix is singular
spx_simplex: numerical problems with basis matrix
spx_simplex: sorry, basis recovery procedure not implemented yet
bfd_ftran: the factorization is not valid
Part of the code is shown as
I'm trying to compile glpk as a DLL so I can integrate it with Ruby.
Can anyone share experiences of doing that?
You can use the batch file 'Build_GLPK_with_VC6_MT_DLL.bat' included
in the distribution (see subdirectory 'w32'). Before running it please
read comments within it.
Some linear programming problems have more than one optimal solution. Glpk
returns just one of them, though.
Is there a way to find all of them using glpk?
Please see:
http://lists.gnu.org/archive/html/help-glpk/2006-06/msg00010.html
___
I have a symbolic set J, further I want to define a set B identifying
each unordered combination of different elements of aforementioned set:
set J;
set B := setof{j1 in J, j2 in J : j1 j2} (j1, j2);
display B;
data;
set J := 'A' 'B';
Unfortunately this results in error
Display
I have installed glpk-4.28 and replaced replace glp_ulong by glp_long in
file
glpkcc.cpp but when I run makeglpkmex Matlab send the next message:
GLPKMEX - A Matlab interface for GLPK. Script installer.
Version 2.4 compatible with GLPK 4.20 (or higher)
(C) 2001-2007, Nicolo'
I want to ask a question about problem of installation.
I got the following message when I run makeglpkmex to install glpkmex.
---
GLPKMEX - A Matlab interface for GLPK. Script
/home1/usr2/my_name/gnu4.28/lib/libglpk.a(glplib08.o)(.text+0x1d1):
In function `_glp_lib_xfopen #39;:
/home/usr2/my_name/glpk-4.28/src/glplib08.c:496: undefined reference to
`gzopen #39;
You need either to make zlib available to the linker passing it the
option '-lz', or rebuild glpk
for using SQL queries the database password must be used in the GLPK model.
It is unwise to save passwords in unencrypted files. In business
environments regulation often explicitly forbids such behavior.
The solution you kindly suggested would leave the password on the harddisk.
Being able
I replaced 'lib_mem_usage' by 'glp_mem_usage' but now have the next
message:
GLPKMEX - A Matlab interface for GLPK. Script installer.
Version 2.4 compatible with GLPK 4.20 (or higher)
(C) 2001-2007, Nicolo' Giorgetti.
Do you want to use graphic installer? Y/N [Y]: y
GLPK path...
This relates to GLPK 4.25 (I did not want to upgrade
past 4.25 while the database functionality was being
resolved, for the record 4.28 is current).
I have been trying to find out how deep the
non-negativity condition on structural variables is.
So I took the tutorial problem from the
I am writing a timetabling program in GMPL and I #39;d like to
write some incompatibility constraints of courses generally, e.g. I
want to say that no more than 1 of {A1, A2, A3} can appear in the same
slot, and no more than 1 of {B1, B2} can appear in the same slot
something like..
set
were just created by glp_create_prob.
But in principle your way is normal; glp_erase_prob does the same.
And does this
mean I can rebuild a similar problem of identical size
and have it warm start?
Yes. If you provide appropriate statuses of rows and columns,
you can then use warm start.
Andrew
Recently (about two months ago) glpk was included in the cygwin
distribution.
Please see:
http://cygwin.com/packages/glpk/
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
Hello,I'm working for compute the revised simplex algorithm using BLAS
and LAPACK.
I was thinking than maybe, it's easier to fork you're simplex
algorithm in order to use it using BLAS and LAPACK.
You are working on matrix with lots of zero, and i think the
easiest way with BLAS and LAPACK
Finally, what is the best way to empty a problem
object of input data -- I currently remove all the
rows and cols but is there a more elegant method?
In 4.29 I added api routine glp_erase_prob which
erases the problem object content as if the object
were just created by glp_create_prob.
using the files delivered with GLPK-4.28 in /w32 glpk is compiled
without ODBC support.
I suggest to add a config.h file to directory /w32 defining
#define ODBC_DLNAME odbc32.dll
/* ODBC shared library name if this feature is enabled */
and to change Makefile_VC6 to use compiler flags:
I saw a couple of threads on this topic, but I don't think they
completely answered my question. I am getting the PROBLEM HAS NO
PRIMAL FEASIBLE SOLUTION from glpsol. This doesn't surprise me as I
am currently debugging my input file. I have a MathProg model of my
problem by my data sets
Xypron,
ODBC32.DLL is a dynamic link library and supplied with Windows
implementing the same standardized interface as iODBC or unixODBC do for
Linux.
All compilers supported by GLPK (Visual C , Borland C 5.5 and DevC++)
can compile GLPK with support for ODBC32.DLL. I did this for
For a VLSI application, I #39;m trying to solve a variant of the
warehouse location problem. My problem stems from the fact that there
are ~5K customers and warehouses, which translates into a 5K by 5K
matrix. This leads me to wonder whether GLPK supports some sort of
sparse matrix
to track all such rows/columns because of recursive
nature of the presolving process.
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
I had a line of math-prog code that had a circular reference:
param FromNode{ (train,sd,seq) in TRAIN_SEGS } symbolic
:= Terminal[FromNode[train,sd,seq]];
GMPL, much further down in the model, just stops without any
warning or anything else. It took a while to figure out the circular
I am debugging the code I plan to use to generate cplex-format
input files. I am comparing my generated files to my gold standard
which is a MathProg model. I provide the MathProg model to glpsol and
write out a cplex translation to compare against my directly-generated
cplex file. I am at
I had a line of math-prog code that had a circular reference:
param FromNode{ (train,sd,seq) in TRAIN_SEGS } symbolic
:= Terminal[FromNode[train,sd,seq]];
GMPL, much further down in the model, just stops without any
warning or anything else. It took a while to figure out the circular
I used a C program to form a gmpl program to solve my problem,for it
has a huge number of variables.But some of the answers were missed,for
example(I copied a part of the answers)
3851 wx[w1] B -2
3852 wx[w2] B -2
As I am going to study a multiobjective problem, I need to
explore Pareto Front with weighted sum method. To improve quality and
speed of search it would be useful to get list of sensitivity bounds
for objective coefficients of variables.
If I use
lpx_print_sens_bnds routine I get a file
I'm trying to specify fixed values to some variables in an array using =
expression (see 4.3 Variable statement in GMPL Reference). Something
like this:
set MySet;
set MySubset within MySet;
var X {i in MySet} = if i in MySubset then 0 else binary;
or, similarly:
var X {i
If I keep the presolver on, the basis information from the previous
solve is lost and the 2-phase primal algorithm kicks off from scratch
(instead of the dual simplex using the previous basis). If I turn the
presolver off, I know my problem is bigger than it needs to be, but
the dual simplex
I'm currently writing an interface between glpk and scilab (see
http://www.scilab.org).
My interface is working fine, but for some problems, glpk emits an
assert which made scilab to hangs.
I have seen a preceding post where somebody adviced to catch the ABORT
signal to avoid the hang.
I
Robbie,
The strange behavior of the simplex solver is caused by small
objective coefficients; they are of the order 1e-9, i.e. even less
than the reduced cost tolerance (which is 1e-8 by default). In the
current implementation the objective row is not scaled, so the solver
just think that reduced
Can someone send me some information about :
Harris two-pas ratio test
Please see:
P.M.J.Harris (1973), Pivot Selection Methods of the Devex LP Code,
Mathematical Programming 5, 1-28.
R.Fletcher (1994), Steepest Edge, Degeneracy and Conditioning in LP,
15th Intl. Symp. on Math. Prog.
I
I am using GLPK to solve a problem by column generation.
I have a master problem MP which is a Mixed Integer Problem.
-STEP 1: the dual problem of the relaxed MP is solved iteratively and
new columns are added at each iteration (the new columns are generated
by solving a sub problem),
-
Does anyone know how many integers variables can handle the glpk 4.29?
Please see file doc/bench1.txt included in the glpk distribution.
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
I write to you hoping to find an answer, because google didnt. Well
my question is: Is there a way to solve a Transportation Problem
without writing the model to a file? Like setting up some variables
and sending them to the solver? My transportation problem has sums
from 1 to N and
Andrew, forgiveness for my ignorance but I can not understand the file
bench1.txt
It is a benchmark for the miplib, a set of mip instances, showing
the performance of the glpk mip solver.
For example, the line
dsbmip -3.051981750e+02 opt 233 3283 92.8
says that
I am new with glpk and I try to write a program in C++ using glpk.
But I get the following problem
I write in the code: (LPX* prob with direction min)
s = lpx_simplex(prob);
if (s != LPX_E_OK)
{
std::string name = ;
lpx_write_mps(prob, (char*)name.c_str());
I know I can probably produce a better analysis of the uniformity
of glpk's random number generator with a sample size smaller than 1.75
million integers, but what would be the point of that?
More interestingly is why does mathprog require a 1000 bytes per integer?
Here are some
I have a small integer program whose optimal solution value is 49.
Root relaxation is 48.5454. Since all the variables are integer, one
expects it to stop when a solution with value 49 is found. Instead,
GLPK takes a long time to converge.
I also tried lp-solve, it found an optimal solution
This question is not directly related to the usage of glpk but it
is related to linear programming in general. I could not find the
answer to this question in the Internet so I hope I can find an answer
here.
1) What is the differences between linear programming (LP)
relaxation and
I have a small problem and I want have the dual solution.
But I find that the GLPK and LP-solve give me 2 different dual
solution (the primal solution and the objective value is the same).
Could someone help to explain the reason??
In column 'Marginal' the glpk simplex solver reports
Has there been any thought to some sort of plug-in API in the GLPK code
so that researchers and others can easily add their research or newly
developed LP solution techniques into the GLPK framework?
In principle, there is no need in such plug-in. The application can
obtain all input data
I am trying to use GLPK in a flux-balance analysis context in biology.
Once the linear constraints are defined and maximization of the
objective function is done, I often find that the solution contains too
many changes across the free variable set, and I cannot change that many
variables
the instance, for which you
obtained inexact results.
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
I am trying to use GLPK in a flux-balance analysis context in biology.
Once the linear constraints are defined and maximization of the
objective function is done, I often find that the solution contains too
many changes across the free variable set, and I cannot change that many
variables
Markus,
I tried both your instances that you posted me with glpsol --nopresol
--noscale --min/--max, and in all four cases the solution reported
was sufficiently accurate in the sense that KKT optimality conditions
had zero residual errors.
Could you demonstrate how inexactness looks like?
Max flow and min cost are the basic problems. In the next step the
approach should is used for multi-commodity scenarios. Therefore, we
started with a LP solver. Do you propose an other direction to solve an
mcf?
There exist many specialized network optimization algorithms, and many
of them
Thanks, I changed it and I can kind of reproduce glpsol's behaviour,
but now I pinpointed the problem to this LP:
http://www.mimuw.edu.pl/~mucha/lpQ_bad.txt
glpsol --cpxlp lpQ_bad.txt
give UNBOUNDED solution here. However, both
glpsol --std --cpxlp lpQ_bad.txt
and
glpsol --adv --cpxlp
Nigel,
Thank you very much for your examples.
I think that the mailing list (which is also is part of the glpk
project) is a best place for these examples, since they are mainly
depend on the environment, not on glpk.
Andrew Makhorin
___
Help-glpk
Is this new? GLPK's home page
http://www.gnu.org/software/glpk/#lists says GLPK has two mailing
lists: help-glpk@gnu.org and [EMAIL PROTECTED] and 'The main
discussion list is help-glpk@gnu.org, and is used to discuss all
aspects of GLPK, including its development and porting.'
Can you give
In a specialized single-commodity network flow solver,
the arithmetic is normally exact even if done in floating point.
That is usually not true with multi-commodity flows.
If glpk floating point gives you the correct basis,
glpk exact should give you the exact answer in fairly short order.
there are two general constraints.
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
Did you plan to include this new version of glp_simplex in 4.30 ?
I plan to replace implementation of the primal simplex by a new
version, however, I am not sure about 4.30, because the new routine
is still under development.
___
Help-glpk mailing
Or you could switch columns and rows in your csv-file:
Param, Value
SHIP_PENALTY, 10
SHIP_LTL_PENALTY, 100
...
Note that spaces are *not* ignored within fields, so probably the
csv file should look like this:
Param,Value
SHIP_PENALTY,10
SHIP_LTL_PENALTY,100
...
Could anyone perhaps give me some literature pointers (preferrably
freely available on the net)
on how to achieve this well-conditioned situation, or some examples?
I was playing a bit with
scaling variables and it seems to improve things a bit, but I am
probably reinventing the wheel.
distribution. Thus it will appear in 4.30.
Andrew Makhorin
___
Help-glpk mailing list
Help-glpk@gnu.org
http://lists.gnu.org/mailman/listinfo/help-glpk
With the lpx_load_matrix for example, and other rutines from the
API, so I wont have to read the data from a file, I pass in the code,
for example a structure.
If you can compute all lp components, you can pass the data directly to
the solver using glpk api routines. However, if you just need
Sorry.. I've probably misunderstood this. Let's take -3=5-8,
5=0, 8=0,
but
|-3|5+8
In that context b1[i] and b2[i] cannot be non-zero at the same time,
because either b1[i] or b2[i] (or both) is always non-basic in any
optimal basic solution.
This only works if the objective is the
Bonjour,
Je suis un nouvel utilisateur de GLPK, Merci de m #39;aider à résoudre
ce prblème Le problème: on veut exprimer une contrainte de priorité
entre deux flux. soit deux débits de flux SR1 et SR2 qui sont parmi
les ressources disponibles(variables du problème à optimiser) on ne
veut pas
I am afraid I have to bother you with a stupid question. My ultimate
goal is to be able to compile the example file sample.c. To make that
work on my computer, I understood from some instructions I found that
I need to compile glpk first on my computer (Windows XP, Borland 5
compiler).
I
I downloaded glpk 4.29 from one of the mirrors and tried to compile it
by running the included Build_GLPK_with_BC5.bat file. Here is theoutput:
C:\glpk\glpk-4.29\w32rem Build GLPK with Borland C++ 5.0
C:\glpk\glpk-4.29\w32rem NOTE: Make sure that HOME variable specifies
correct
I don't know why, however, the bcc32 compiler installed on your system
performs C++ compilation rather than C as defined by filename extension.
If I rename '.c' to '.cpp' and run bcc32, I get the same errors as you.
Most probably there is the option -P specified in the file bcc32.cfg
(it is
I wondered if there were some special features in GLPK enabling the
use of a non linear constraitn such :
a = 0 or a 3 .
There is no such feature. However, you can model semicontinuous variables
as follows:
a = 3 * z,
a = M * z,
where z is an auxiliary binary variable, M is an upper bound
I am trying to write a model to maximize overall output when selecting
personnel given a budget, the salary of each potential person and
their output ability.
The problem I have is writing a rule where one person will be
choosen as team leader and their output will be calculated as double.
Few days back, there was a discussion on the mailing list about
sparse and dense constraint coefficient matrices. I thought it would
be nice to be able to visualize matrix A and see how much sparse it
is. This might help verify, understand and debug LP/MIP models created
using MathProg.
Glpk
I just played around these functions, seems to be working fine. For
small matrices the BMP images would be too small. For 10 X 10 matrix,
it would 10 X 10 pixel image.
Did you create these functions and structures like SPM to debug in
initial stages of GLPK development?
Yes, I used them to
give optimal integer solution (x_1 = 2, x_2 = 2). Since there are 3 optimal
integer solutions, it is matter of finding all of them, instead of accepting
first optimal solution. I need to work on this to find
You probably do not loose solution, and do not have numerical
difficulties; just
301 - 400 of 1851 matches
Mail list logo