Thank you everybody, almost every point discussed here is now written on my
blog http://stla.github.io/stlapblog/posts/KantorovichWithJulia.html.
Looks good!
On Friday, May 2, 2014 12:55:49 PM UTC-4, Stéphane Laurent wrote:
Thank you everybody, almost every point discussed here is now written on
my blog http://stla.github.io/stlapblog/posts/KantorovichWithJulia.html.
Right, it works. Thank you.
If I don't call GLPKMathProgInterface, does JuMP use an internal solver ?
Le mardi 22 avril 2014 23:25:07 UTC+2, Carlo Baldassi a écrit :
Note that you can still use GLPK.exact with JuMP, you just need to add
change the m=Model() line to this:
using
On Wednesday, April 23, 2014 3:40:02 PM UTC-4, Stéphane Laurent wrote:
If I don't call GLPKMathProgInterface, does JuMP use an internal solver ?
If a solver isn't specified, JuMP (actually MathProgBase) will search for
an available solver and pick one by default. JuMP does not have an
Miles, I have successfully installed JuMP and GLPKMathProgInterface on
Windows 32-bit.
Your code works very well, this is really awesome !! However the result is
not as precise as the one given by *GLPK.exact*.
using JuMP
mu = [1/7, 2/7, 4/7]
nu = [1/4, 1/4, 1/2]
n = length(mu)
m =
Cool! Glad to hear you got it working. Supporting exact coefficients in
JuMP is technically possible, and I've opened an issue for
it: https://github.com/JuliaOpt/JuMP.jl/issues/162. This will probably
remain on the wishlist for a while.
On Tuesday, April 22, 2014 2:28:01 PM UTC-4, Stéphane
Note that you can still use GLPK.exact with JuMP, you just need to add
change the m=Model() line to this:
using GLPKMathProgInterface
m = Model(solver=GLPKSolverLP(method=:Exact))
while all the rest stays the same.
As an aside, it's really kind of annoying that GLPK.exact uses (basically)
My blog post is updated.
Iain, I have tried your code with the example of my blog. I see the good
result in the output (*3//28*), but I don't understand how to know it is
the good one.
using RationalSimplex
using Base.Test
b = [1//7, 2//7, 4//7, 1//4, 1//4, 1//2]
c = [0//1, 1//1, 1//1,
Stéphane, sorry for the confusion (I should have made myself clearer
before), but the new version of the blog post is not correct, in that the
difference between 1-X and 1.-X has nothing to do with Int64 vs Float64,
but rather with the difference between the operators `-` (minus) and `.-`
Thank you Iain I will try your solver soon I hope. And thank you again
Carlo, I will update my post.
I implemented a version of simplex method for rational numbers - so you
solve it exactly in pure Julia.
https://github.com/IainNZ/RationalSimplex.jl
Not for serious work - just for fun!
On Saturday, April 12, 2014 11:50:26 AM UTC-4, Stéphane Laurent wrote:
Thank you everybody. I have updated
Where's the MathProgBase interface? :)
On Wed, Apr 16, 2014 at 11:07 PM, Iain Dunning iaindunn...@gmail.com wrote:
I implemented a version of simplex method for rational numbers - so you
solve it exactly in pure Julia.
https://github.com/IainNZ/RationalSimplex.jl
Not for serious work - just
Thank you everybody. I have updated my blog
posthttp://stla.github.io/stlapblog/posts/KantorovichWithJulia.html,
especially to include Carlo's comments.
Unfortunately I have some problems to use JuMP (I have opened another
discussion about it). And installing pycddlib on Windows 64bit is a
Thank you for these precious informations. The JuMP package looks very
awesome, I hope to give it a try soon.
There was a Julia age during which BigInt(3)/BigInt(28) was equal to the
BigRational 3/28, why this feature has been removed ?
It would be too long to explain what my R appli here
By the way for another problem I need to get the vertices of the
polyhedron defined by the linear constraints, as with the cddlib library,
do you know how I could get that ?
Enumerating vertices requires a very different algorithm from optimizing
over polyhedra. The best way to do this
Unless I'm blind and just can't find one, it appears there is not yet a
solid high-dimensional computational geometry package for Julia, like
cddlib or the Multi-Parametric Toolbox for Matlab. I imagine a wrapper
around cddlib would be fairly easy to write (perhaps even autogenerated),
or you
There was a Julia age during which BigInt(3)/BigInt(28) was equal to the
BigRational 3/28, why this feature has been removed ?
I don't think that command ever worked like that actually; in order to get
Rational values, you need to use a double-slash // in stead of a single
slash:
big(3) //
Again, thank you for all these answers. Sorry Carlo, I missed the double
slash in your previous answer.
It would be a good opportunity for me to call Python in order to train my
skills in Python in addition to Julia. But what do you suggest me to call
pycddlib with PyCall rather than calling
Again, thank you for all these answers. Sorry Carlo, I missed the double
slash in your previous answer.
It would be a good opportunity for me to call Python in order to train my
skills in Python in addition to Julia. But why do you suggest me to call
pycddlib with PyCall rather than calling
Either approach should work in principle (pycddlib vs. cddlib), but I
suspect that the Python wrapper is higher-level, and will be easier to use
from Julia. For reference, here's a snippet of how you might calculate the
extreme points as rationals, calling pycddlib from Julia (adapted from code
Hello guys,
I hope you'll enjoy this article on my
bloghttp://stla.github.io/stlapblog/posts/KantorovichWithJulia.html
.
If you're able to use GNU MP on your machine, would you be able to find
*3/28* ?
Any other comment is welcomed !
You might be exited when you find out that you can use μ, ν, M₁ and M₂ as
variable names in Julia, because we support UTF-8.
Sometimes it might give more readable code, when you compare it to the math
texts.
Ivar
kl. 19:08:19 UTC+2 onsdag 9. april 2014 skrev Stéphane Laurent følgende:
Hi Stéphane, nice post! I have a number of comments and suggestions which
you may find useful. I accompanied these comments with some demo code, you
can find here https://gist.github.com/carlobaldassi/10312215.
A) Generic to Julia
A.1) as Ivar said, you can use Unicode characters if you want;
About GLPK.exact it is not possible to get the rational number 3/28 instead
of a decimal approximation ?
No, unfortunately. Also, for that to happen/make sense, you'd also need to
be able to pass all the *inputs* as exact rational values, i.e. as 1//7
instead of 1/7. This would be possible if
Todo: right generic simplex implementation - it'd be so easy in Julia!
On Wednesday, April 9, 2014 6:18:26 PM UTC-4, Carlo Baldassi wrote:
About GLPK.exact it is not possible to get the rational number 3/28
instead of a decimal approximation ?
No, unfortunately. Also, for that to
When we have a simplex solver (either in Julia or external) that supports
rational inputs, we could consider making this work with JuMP, but for now
JuMP stores all data as floating-point as well.
Stephane, nice work. LP definitely needs more exposure in the probability
community. Please
Another semi-hacky option here is using a conventional double-precision LP
solver to tell you the active set at the solution the solver considered
optimal (up to whatever its tolerance was set to). Then you can take that
active set and solve the arbitrary-precision version of the constraint
Hello Iain, I don't understand what you mean :
*julia versioninfo()*
*Julia Version 0.2.0*
*Commit 05c6461 (2013-11-16 23:44 UTC)*
*Platform Info:*
* System: Windows (x86_64-w64-mingw32)*
* WORD_SIZE: *
*julia *
*64*
*julia *
* BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH
Unfortunately, from insepcting GLPK source code it seems that whether to
use GNU MP or GLPK bignum is decided at compile time. So it appears that
the Windows binaries which are automatically downloaded by the Julia
package are just not compiled with GNU MP. As far as I can tell, the
options
It looks like those GLPK binaries are Visual Studio builds, and as I
understand it recent versions of GMP are difficult if not impossible to
compile with Visual Studio. In fact GMP was forked into MPIR
(http://www.mpir.org/#about), with one major motivation being MSVC support.
MPIR might be
Thank you for your answers. Unfortunately, (pre)-compiled binaries,
dll, etc, is like Chinese for me. Moreover when you talk about GLP I
don't know if you talk about the C library or the julia Package. Currently
I was just trying GLPK for fun so this is not important. Thank you again.
I strongly recommend you update to Julia 0.2 at least and use the dedicated
GLPK package https://github.com/JuliaOpt/GLPK.jl - I'm sure you'll get much
better support. Version 0.1 of Julia (which had inbuilt GLPK - I had
forgotten!) isn't supported anymore.
How did you get version 0.1?
32 matches
Mail list logo