[Symphony] bicriteria example?

[Ted Ralphs]

Well, it was a combination of four different things that caused the problem.

1. The new pre-processor was messing with the reformulation required
for the use of the Chebyshev norm in order to find non-supported
solutions. For now, I have disabled the pre-processor when the
multicriteria solve capability is invoked.

2. The new procedure for automatically setting the problem granularity
was also conflicting with the setting of granularity for the
multi-criteria case, which has to be done a little differently.

3. The objective coefficient on x_2 in your MPS file should be "0" to
reproduce the example precisely (see attached file).

4. The example code in the slides you were looking at is actually
wrong---because it's a maximization problem, the coefficient in the
second objective for x_2 should be set to -1. In other words, you
should have the following line of code in your main() function:

   sym_set_obj2_coeff(env, 1, -1);

Alternatively, you can actually set the objective sense to
maximization, but then the MPS file should be changed to have
nonnegative coefficients. Sorry about that! The code is now fixed in
stable/5.2 and trunk. If you check out

https://projects.coin-or.org/svn/SYMPHONY/stable/5.2

with subversion and build as before, then create your executable from
the fixed main() function and run it on the attached test.mps, you
should get the expected results. Thanks for pointing this out!

If you prefer to download an archive of the source for a point release
instead of using subversion, I will make a new release of SYMPHONY
over the weekend. I believe you can also get a zip file of the whole
thing by going here:

https://projects.coin-or.org/SYMPHONY/changeset/1770/stable/5.2?old_path=%2F&format=zip

Cheers,

Ted

On Tue, May 4, 2010 at 12:38 PM, Ted Ralphs <ted at lehigh.edu> wrote:
> Something does seem to be broken with the bicriteria solve capability.
> We'll take a look at it and get back to you.
>
> Cheers,
>
> Ted
>
> On Mon, May 3, 2010 at 4:09 PM, Ross Anderson <rpanderson at gmail.com> wrote:
>> Hi all,
>> I'm attempting to reproduce the bicriteria example found in many of the
>> presentations (like the Advanced DIMACS06 tutorial). Here's the problem
>> reproduced:
>> vmax [8x1, x2],
>> subject to 7x1 + x2 <= 56,
>> 28x1 + 9x2 <= 252,
>> 3x1 + 7x2 <= 105,
>> x1, x2 >= 0
>>
>> Here's my MPS file:
>>
>> NAME          BICRIT
>> ROWS
>>  N  COST
>>  L  LIM1
>>  L  LIM2
>>  L  LIM3
>> COLUMNS
>>     XONE      COST                -8   LIM1                 7
>>     XONE      LIM2                 28  LIM3                 3
>>     XTWO      COST                -1   LIM1                 1
>>     XTWO      LIM2                 9   LIM3                 7
>> RHS
>>     RHS1      LIM1                 56  LIM2                252
>>     RHS1      LIM3                105
>> BOUNDS
>>  LI BND1      XONE                 0
>>  LI BND1      XTWO                 0
>> ENDATA
>>
>>
>> and the tutorial code:
>>
>> #include "symphony.h"
>>
>> int main(int argc, char **argv)
>> {
>>
>>    sym_environment *env = sym_open_environment();
>>    sym_parse_command_line(env, argc, argv);
>>    sym_load_problem(env);
>>
>>    sym_set_obj2_coeff(env, 1,1);
>>    sym_mc_solve(env);
>>
>>    sym_close_environment(env);
>>
>>    return(0);
>>
>> }
>>
>>
>> The solver screen outputs  XONE      8.000  (implying XTWO is 0), which is
>> the outcome for \theta in (-infty, 1.333). How do I go about finding the
>> other ranges? I also tried passing mc_find_supported_solutions 1 in a
>> parameter file - no change.
>>
>> Thank you.
>>
>> _______________________________________________
>> Symphony mailing list
>> Symphony at list.coin-or.org
>> http://list.coin-or.org/mailman/listinfo/symphony
>>
>>
>
>
>
> --
> Dr. Ted Ralphs
> Associate Professor, Lehigh University
> (610) 628-1280
> ted 'at' lehigh 'dot' edu
> coral.ie.lehigh.edu/~ted
>



-- 
Dr. Ted Ralphs
Associate Professor, Lehigh University
(610) 628-1280
ted 'at' lehigh 'dot' edu
coral.ie.lehigh.edu/~ted
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test.mps
Type: application/x-mps
Size: 510 bytes
Desc: not available
URL: <http://list.coin-or.org/pipermail/symphony/attachments/20100507/74081ac4/attachment.bin>