Dear Dmitry, On Tue, Aug 18, 2009 at 12:21 PM, Torsten Anders < [email protected]> wrote:
> Dear Dmitry, > > Is it possible to create and load Oz program source fragment at the >> run-time? >> > > > Oz is a highly dynamic language. You can construct (and then solve) > constraint problems depending on user input. Depending on how much > flexibility you need there are different approaches (e.g., in an extreme > case, you can feed an Oz compiler arbitrary Oz code that is generated > dynamically depending on user input). > My advice is not to generate Oz code depending on user input, but to make the search script generic instead. Use some conventional data structures to specify the problem you want to solve, and let the script post the propagators following the data specifying the problem. The performance overhead is really small, because the "interpretation" of the problem by the script is only done at the very beginning of the constraint solving. Cheers, Raphael > Best > Torsten > > > On Aug 18, 2009, at 10:23 AM, Dmitry Negius wrote: > > In my program I need to implement answers quantitative questions from >> user. >> Domain variables are not known in the design time and will be obtained at >> the runt-time from the user questions and it will be very many variables >> in the constraints system. Constraints system will be constructed at the >> run-time. Is it possible to create and load Oz program source fragment at >> the run-time? >> >> 2009/8/18 Torsten Anders <[email protected]> >> Dear Dmitry, >> >> I don't quite understand your system, do you mean a constraint problem >> like the following? >> >> {Browse >> {SearchOne >> proc {$ Sol} >> %% variable domain declaration >> A = {FD.decl} >> B = {FD.int 0#1} >> in >> Sol = unit(a:A b:B) >> %% constraint application >> A <: 1000 >> {FD.impl (A >: 255) B 1} >> {FD.impl {FD.nega B} {FD.nega (A >: 255)} 1} >> %% and so forth >> %% >> %% distribution strategy (variable ordering) >> {FD.distribute ff Sol} >> end}} >> >> Best >> Torsten >> >> >> On Aug 17, 2009, at 6:58 PM, Dmitry Negius wrote: >> Excuse me but I want to ask: "Does mozart-oz solves logicaly-algebral >> constraints - the mix of logical and algebral operations over variables? >> System NUT for example solves only algebral constraints in the form of >> algebraic equations (formula = formula). >> >> For my AI project I need to solve constraints like this: >> A>255 => B. >> not B. >> ?- A<1000. >> >> Of couse this is mo intelligent then to solve only algebraic constraints >> and has better application. >> >> Manually this constaints system can be solved in the next way: >> >> A>255 => B >> ------------------- >> not B => not (A>255) >> ------------------------------- >> not B => (A=<255) not B >> ---------------------------------------------------- >> A=<255 >> 255<1000 >> ------------------------- >> A<1000 >> >> >> >> -- >> Torsten Anders >> Interdisciplinary Centre for Computer Music Research >> University of Plymouth >> Office: +44-1752-586219 >> Private: +44-1752-558917 >> http://strasheela.sourceforge.net >> http://www.torsten-anders.de >> > > > _________________________________________________________________________________ > mozart-users mailing list > [email protected] > http://www.mozart-oz.org/mailman/listinfo/mozart-users >
_________________________________________________________________________________ mozart-users mailing list [email protected] http://www.mozart-oz.org/mailman/listinfo/mozart-users
