[Fwd: Re: contribution to GLPK]

-------- Forwarded Message --------
From: Vuk <[email protected]>
To: Andrew Makhorin <[email protected]>
Subject: Re: contribution to GLPK
Date: Sat, 22 Jul 2017 02:15:58 +0200 (CEST)

Hi Andrew,

3 models are ready, I think. Next step should be involving mailing list 
members as jury?

Thanks & best regards,
Peter


On Fri, 21 Jul 2017, Vuk wrote:

> Hi Andrew,
>
> Thanks for your recommendations. Fixing is in progress with max load.
>
> My question is: all four .mod models are ready to post onto the glpk mailing 
> list in order to be discussed?
>
> Tibor's cc-ing will be important again, because in some weeks/months, I will 
> prepare and contribute my power plant scheduler with fake data. Until that, 
> Tibor will be removed from cc recipients.
>
> There is an embedding of a special infinite graph into the 6th dimensional 
> euclydian grid, which was found by a low level .lp model generated by a 
> selfmade python script. This graph is important, because this 
> construction+existence has a corollary: all 256 elementary cellular automata 
> can be embedded into E^n grids if n >= 6, not just the amphirical ones. The 
> embedded automata is outer-totalistic and has 3 states only, including 
> background. I will publish it later, but these results were checked both by 
> Tibor et al and independently, the conwaylife.com forums community.
>
> Besides, I haven't found any patterns to be followed like indentation, 
> maximal number of columns, other usual "coding styles".
> Well, self-explanatory variable names + descriptions were missing, as you 
> pinpointed. Remark styles will be guessed based on examples delivered in 
> debian/ubuntu builds - it is also important when merging/splitting 
> descriptions.
>
> Also, shall I explain why cos() magic is used for example when the usage of 
> Hamming-distance was hidden in toto.mod? (lacking type casting support in 
> *MPL?)
>
> My goal partly is: showing that even _high(er) level_ mathematical 
> programming/modeling is a powerful development solution.
>
> Best Regards,
> Peter
>
> On Fri, 21 Jul 2017, Andrew Makhorin wrote:
>
>> Hi Peter,
>> 
>> Thank you for your contribution.
>> 
>> Could you please provide your example models with more comments?
>> If someone looks for an example, he usually is not interested in a
>> particular solution of a problem, with which he might not be familiar.
>> In our case the purpose of an example model is to show how this or that
>> problem can be formalized and how it can be then formulated in terms of
>> lp/mip. Do you agree? So please add to your examples a brief
>> description of the problem being solved (just a reference to a website
>> is not sufficient), and also please add descriptions of all variables
>> and constraints; otherwise your example models are completely unclear
>> and unreadable. Please look at models glpk/examples/pentomino.mod and
>> glpk/examples/sudoku.mod as for good examples. Besides, please provide
>> only one version of data section included into .mod file.
>> 
>> Best regards,
>> 
>> Andrew Makhorin
>> 
>> 
>> On Thu, 2017-07-20 at 23:26 +0200, Vuk wrote:
>>> Greetings!
>>> 
>>> After a short break, here you are, some models that I would like to check
>>> whether it is worth to include them into the next release of GLPK.
>>> 
>>> There will also be later an example pure LP model of scheduling power
>>> plants based on my thesis. Its matrix is a network matrix, basically, it
>>> is a matroid. Proven since 2006 (I proved the total unimodularity of the
>>> matrix in 2004 based on colourings and works of Ghouila-Houri (1962) et
>>> al.)
>>> 
>>> The current list:
>>> 1. cellular automata class embedding (Wolfram into margolus)
>>> 2. tiling with suspected rectifiable polyominoes
>>> 3. Conway's Game of Life garden of eden tester
>>> 4. covering code searcher
>>> 
>>> Regards,
>>> Peter, NASZVADI
>>> +36 30 5984 790
>>> 
>>> (added Tibor Illes, he supervised my thesis. If I could be wrong, he can
>>> correct me:)
>>> 
>> 
>> 
>> 
>

_______________________________________________
Help-glpk mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/help-glpk