GAMS World Forum

It's unclear what np is, but I assume this is an alias to n. Posting complete examples really helps. Here are two solutions. One is a direct on and the other uses more symbols but can easily serve to calculate similar sums for other nodes, not just n4: set l /L1*L6/, n /n1*n5 /, t /t1/; alias (n,np)...

That's obviously a weird behavior. Without more information it is hard to say what's going on.

-Michael

Solvers work with tolerances. A popular tolerance for deciding if a value is "integer" is if abs(x-round(x))<1e-5. Many solvers allow you to change such tolerances. For example, Cplex: epint (see https://www.gams.com/latest/docs/S_CPLE ... CPLEXepint).

-Michael

I guess there is currently no other way than to create (utf-8 encoded) CSV file (from GAMS) and to load these into Excel. We (GAMS) has this on the radar and are working on this.

-Michael

Alternative optimal solutions? Both solutions have objective 2. So you get different (primal) solutions depending on pure chance. This is not uncommon. Just search the internet for "degenerate linear programs".

-Michael

Hi,

You need to understand the difference between compile and execution time in GAMS. See e.g. https://www.gams.com/latest/docs/UG_Gam ... ll_TwoPass. Creating a file at execution time is done via the put facility, see https://www.gams.com/latest/docs/UG_Put.html

-Michael

Hi,

back in the days (12 years ago?) I wrote a BP for the graph coloring algorithm in GAMS. This uses some advanced GAMS magic. You can freely use this, but I can't/won't support this any longer.
-Michael

The barrier algorithm is notoriously difficult to restart. See https://scicomp.stackexchange.com/quest ... warm-start.

-Michael

Viktor, The initialize is a costly function. It runs a GAMS process to create scratch file (the dat files) and initializes the model instance from those. Instead of breaking the sequence 5000k solves into 100k initialize/reset it might be worth seeing where the memory leak comes from. I can reproduc...

There are many ways how to deal with infeasibility. Bruce McCarl Newsletter vol. 40 gives good recommendations on how to deal with infeasible models (see section 3 of https://www.gams.com/fileadmin/communit ... news40.pdf).

-Michael