You can subscribe to this list here.
2010 
_{Jan}

_{Feb}

_{Mar}

_{Apr}

_{May}

_{Jun}

_{Jul}

_{Aug}

_{Sep}
(1) 
_{Oct}
(8) 
_{Nov}

_{Dec}


2011 
_{Jan}

_{Feb}

_{Mar}
(5) 
_{Apr}
(4) 
_{May}
(2) 
_{Jun}

_{Jul}
(7) 
_{Aug}
(34) 
_{Sep}
(11) 
_{Oct}
(25) 
_{Nov}
(9) 
_{Dec}
(19) 
2012 
_{Jan}
(3) 
_{Feb}

_{Mar}
(7) 
_{Apr}
(13) 
_{May}
(13) 
_{Jun}
(24) 
_{Jul}
(16) 
_{Aug}
(2) 
_{Sep}

_{Oct}
(14) 
_{Nov}
(34) 
_{Dec}
(9) 
2013 
_{Jan}
(38) 
_{Feb}
(18) 
_{Mar}
(17) 
_{Apr}
(17) 
_{May}
(36) 
_{Jun}
(3) 
_{Jul}
(4) 
_{Aug}
(25) 
_{Sep}
(25) 
_{Oct}
(15) 
_{Nov}
(29) 
_{Dec}
(9) 
2014 
_{Jan}
(40) 
_{Feb}
(84) 
_{Mar}
(25) 
_{Apr}
(47) 
_{May}
(28) 
_{Jun}
(21) 
_{Jul}
(19) 
_{Aug}
(27) 
_{Sep}
(16) 
_{Oct}
(68) 
_{Nov}
(33) 
_{Dec}
(15) 
2015 
_{Jan}
(15) 
_{Feb}
(16) 
_{Mar}
(2) 
_{Apr}
(11) 
_{May}
(16) 
_{Jun}
(65) 
_{Jul}
(26) 
_{Aug}
(22) 
_{Sep}
(18) 
_{Oct}
(40) 
_{Nov}
(14) 
_{Dec}
(18) 
2016 
_{Jan}
(8) 
_{Feb}
(38) 
_{Mar}
(11) 
_{Apr}
(47) 
_{May}
(12) 
_{Jun}
(5) 
_{Jul}
(10) 
_{Aug}
(3) 
_{Sep}
(30) 
_{Oct}
(32) 
_{Nov}
(29) 
_{Dec}
(10) 
2017 
_{Jan}
(5) 
_{Feb}
(9) 
_{Mar}
(38) 
_{Apr}
(6) 
_{May}

_{Jun}

_{Jul}

_{Aug}

_{Sep}

_{Oct}

_{Nov}

_{Dec}

S  M  T  W  T  F  S 


1
(2) 
2
(1) 
3

4

5

6

7

8

9

10

11

12
(6) 
13
(2) 
14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30
(3) 
31




From: Roland Kaminski <kaminski@cs...>  20121030 20:41:53

On Tuesday, October 30, 2012 08:15:28 PM Martin wrote: > On Tue, 20121030 at 12:34 0700, Ke Wang wrote: > > Hi there: > > > > I had trouble running Clingo, after I downloaded it and unzipped it to > > a folder to which I cd to and invoked the command, here is what > > happens: > > > > ke@...:~/Downloads/clingo3.0.4x86linux$ clingo version > > The program 'clingo' is currently not installed. To run 'clingo' > > please ask your administrator to install the package 'gringo' > > > > After I have done the same thing with Gringo > > ke@...:~/Downloads/gringo3.0.4x86linux$ gringo help > > The program 'gringo' is currently not installed. To run 'gringo' > > please ask your administrator to install the package 'gringo'. > > > > I was wondering what it means by they are not installed and how can I > > get them installed. > > > > Thanks very much. > (you didn't say which operating system you're using) Likely, it is an Ubuntu (or Debian). You could try: sudo aptitude install gringo clasp This command installs the binaries (together with clasp) in a systemwide accessible location. Of course this requires administrative privileges. Regards, Roland 
From: Martin <mjb@cs...>  20121030 20:15:41

On Tue, 20121030 at 12:34 0700, Ke Wang wrote: > Hi there: > > I had trouble running Clingo, after I downloaded it and unzipped it to > a folder to which I cd to and invoked the command, here is what > happens: > > ke@...:~/Downloads/clingo3.0.4x86linux$ clingo version > The program 'clingo' is currently not installed. To run 'clingo' > please ask your administrator to install the package 'gringo' > > After I have done the same thing with Gringo > ke@...:~/Downloads/gringo3.0.4x86linux$ gringo help > The program 'gringo' is currently not installed. To run 'gringo' > please ask your administrator to install the package 'gringo'. > > I was wondering what it means by they are not installed and how can I > get them installed. > > Thanks very much. Is . in your PATH? I.E. when you run: echo $PATH is . (the current directory) in the list of directories given. Most systems tend not to include it by default (you didn't say which operating system you're using) for security reasons. If not you can either add it or you can try: ./clingo help ./gringo help HTH Cheers,  Martin 
From: Ke Wang <kewangad@gm...>  20121030 19:34:15

Hi there: I had trouble running Clingo, after I downloaded it and unzipped it to a folder to which I cd to and invoked the command, here is what happens: ke@...:~/Downloads/clingo3.0.4x86linux$ clingo version The program 'clingo' is currently not installed. To run 'clingo' please ask your administrator to install the package 'gringo' After I have done the same thing with Gringo ke@...:~/Downloads/gringo3.0.4x86linux$ gringo help The program 'gringo' is currently not installed. To run 'gringo' please ask your administrator to install the package 'gringo'. I was wondering what it means by they are not installed and how can I get them installed. Thanks very much. 
From: Martin <mjb@cs...>  20121013 21:57:52

On Fri, 20121012 at 15:32 0400, Ankesh wrote: > Kind of real number support we need: > > > 1) Any input disjunctive logic program (i.e. CCEC description) is > expected to have a finite number of real numbers as constants. > > > 2) The input real numbers combine according to the arithmetic > equations (involving operators +, , *, /) defined in the CCEC > description. Real numbers generated using those equations are finitely > many (ignoring the nuances relating floatingpoint arithmetic), and > can be predetermined. Ummm... floating point != real. If you want real semantics then *do not* use floating point unless you can prove that it is really OK. These proofs are often difficult. > 3) With a little bit of preprocessing, we can restrict the domain of > real numbers, in the input disjunctive logic program, to finitely many > constants. As such, we intend to use an independent library for > numerical reasoning (to solve differential equations etc.), so the > aforementioned preprocessing step is not inconvenient. We intend to > use ASP solvers for logic reasoning, which is performed in tandem with > numerical reasoning. > > > We probably need pretty basic support for real numbers, a) real > numbers are not bounded (in the sense that they can be arbitrarily > huge and of any sign), but they will be finitely many, b) no reasoning > over dense real numbers (they are finite), c) real numbers may > participate in nonlinear equations, but these equations are not > constraints in the sense of CSP. I'm not sure I follow; would it be possible to give us an example of the whole process? Cheers,  Martin 
From: Martin <mjb@cs...>  20121013 21:51:10

On Fri, 20121012 at 13:15 0500, Linas Vepstas wrote: > I don't know the answer to the questions below, but want to add a > related question: is there any work to add support for anything > resembling "satisfiability modulo theories" concepts from model > theory? I notice that there are many SMT solvers out there, and they > seem to overlap ASP at least partly in function. Not really (as far as I know). There is some work by Ilkka and Tomi at Aalto which translates AnsProlog into difference logic and that could possibly be extended. Alternatively, I have got a design (but, alas, at the moment, not enought time to work on them  all help welcome) for creating a generic interface so that SMT solvers can be used as ASP solvers and the additional theories are 'accessible' in ASP. > I found ASP much easier to use though. SMTLIB is designed as a machine input / interface system and prioritises easy of implementation over human readbaility / usability. Cheers,  Martin 
From: Forrest Sheng Bao <forrest.bao@gm...>  20121012 22:56:06

I would recommend two works, AC(C) and EZCSP. AC(C): Integrating Answer Set Programming and Constraint Logic Programming by Veena S. Mellarkod, Michael Gelfond and Yuanlin Zhang http://isaim2008.unl.edu/PAPERS/SS1AI+*Logic*/MGelfondss1.pdf It also has a journal version here: http://www.springerlink.com/index/9G77J73424952M11.pdf EZCSP: http://marcy.cjb.net/ezcsp/index.html I normally do not use it because it requires Sicstus Prolog. AC(C) supports integrating constraint logic programming (in theory, over any domain) into ASP. Current AC(C) solver integrates CLP(R), i.e., constraint logic programming over real numbers, into ASP. I have used AC(C) for temporal planning and it went well on some problems where continuous variables have large domains (in these cases, discretizing them increases search space greatly). A brief overview of this application of AC(C) is at https://sites.google.com/site/pddl2acc/ You can find simplified and modified version of AC(C) in my slides/posters. Please feel free to let me know whether you might be interested in using AC(C). Since it is made by someones in my lab before, I would be happy to see people using it. As some others suggested, SMT is also a good candidate. Cheers, Forrest On Fri, Oct 12, 2012 at 12:38 PM, Ankesh <ankesh@...> wrote: > Hello, > > I am a PhD student at Rensselaer Polytechnic Institute, Troy, NY. I am > looking to implement a modelbuilder for continuouschange event calculus > (CCEC) [1]. Answer set solvers have been deployed to build models for > discretetime event calculus [2]. We would like to use solvers to implement > a reasoner for CCEC. > > As I understand, claspd doesn't support real numbers. Is there an ASP > solver that supports real numbers and real arithmetic, which we could use? > > (As far as we know, DLVcomplex supports real numbers and is available for > free, but the code is not open.) > > Thanks and regards, > > Ankesh > > P.S. We expect that given any CCEC description, finiteness of the domain > can be guaranteed with some preprocessing (ignoring some nuances related > to floatingpoint arithmetic, for now). > > [1] Reasoning about Discontinuities in the Event Calculus (1996). > http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.9979 > > [2] http://reasoning.eas.asu.edu/ecasp/ > 
From: Peter Schueller <schueller.p@gm...>  20121012 21:47:09

There is the HEX formalism with the dlvhex solver (opposed to DLV, the dlvhex solver is an Open Source software which uses as solver backends either its own builtin grounder and model builder (also open source), dlv (closed source), or gringo for grounding and clasp for solving (open source)). http://www.kr.tuwien.ac.at/research/systems/dlvhex/ HEX supports disjunctive logic programs and external predicates which can make use of any C++ computation that provides boolean values to the model builder. In that way HEX is ASP modulo external computations which can contain whole theories. (e.g., description logics) For the clasp backend, dlvhex uses a special interface of clasp which also allows custom theories to be integrated into the clasp model building process, so clasp contains support for "external theories" as already indicated by Martin in his reference to clingcon. Best, Peter On Fri, Oct 12, 2012 at 7:15 PM, Linas Vepstas <linasvepstas@...> wrote: > I don't know the answer to the questions below, but want to add a related > question: is there any work to add support for anything resembling > "satisfiability modulo theories" concepts from model theory? I notice that > there are many SMT solvers out there, and they seem to overlap ASP at least > partly in function. I found ASP much easier to use though. > >  linas > > On 12 October 2012 12:38, Ankesh <ankesh@...> wrote: >> >> Hello, >> >> I am a PhD student at Rensselaer Polytechnic Institute, Troy, NY. I am >> looking to implement a modelbuilder for continuouschange event calculus >> (CCEC) [1]. Answer set solvers have been deployed to build models for >> discretetime event calculus [2]. We would like to use solvers to implement >> a reasoner for CCEC. >> >> As I understand, claspd doesn't support real numbers. Is there an ASP >> solver that supports real numbers and real arithmetic, which we could use? >> >> (As far as we know, DLVcomplex supports real numbers and is available for >> free, but the code is not open.) >> >> Thanks and regards, >> >> Ankesh >> >> P.S. We expect that given any CCEC description, finiteness of the domain >> can be guaranteed with some preprocessing (ignoring some nuances related to >> floatingpoint arithmetic, for now). >> >> [1] Reasoning about Discontinuities in the Event Calculus (1996). >> http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.9979 >> >> [2] http://reasoning.eas.asu.edu/ecasp/ >> >> >> >>  >> Don't let slow site performance ruin your business. Deploy New Relic APM >> Deploy New Relic app performance management and know exactly >> what is happening inside your Ruby, Python, PHP, Java, and .NET app >> Try New Relic at no cost today and get our sweet Data Nerd shirt too! >> http://p.sf.net/sfu/newrelicdev2dev >> _______________________________________________ >> Potasscousers mailing list >> Potasscousers@... >> https://lists.sourceforge.net/lists/listinfo/potasscousers >> > > >  > Don't let slow site performance ruin your business. Deploy New Relic APM > Deploy New Relic app performance management and know exactly > what is happening inside your Ruby, Python, PHP, Java, and .NET app > Try New Relic at no cost today and get our sweet Data Nerd shirt too! > http://p.sf.net/sfu/newrelicdev2dev > _______________________________________________ > Potasscousers mailing list > Potasscousers@... > https://lists.sourceforge.net/lists/listinfo/potasscousers > 
From: Ankesh <ankesh@gm...>  20121012 19:32:58

Kind of real number support we need: 1) Any input disjunctive logic program (i.e. CCEC description) is expected to have a finite number of real numbers as constants. 2) The input real numbers combine according to the arithmetic equations (involving operators +, , *, /) defined in the CCEC description. Real numbers generated using those equations are finitely many (ignoring the nuances relating floatingpoint arithmetic), and can be predetermined. 3) With a little bit of preprocessing, we can restrict the domain of real numbers, in the input disjunctive logic program, to finitely many constants. As such, we intend to use an independent library for numerical reasoning (to solve differential equations etc.), so the aforementioned preprocessing step is not inconvenient. We intend to use ASP solvers for logic reasoning, which is performed in tandem with numerical reasoning. We probably need pretty basic support for real numbers, a) real numbers are not bounded (in the sense that they can be arbitrarily huge and of any sign), but they will be *finitely many*, b) no reasoning over dense real numbers (they are finite), c) real numbers may participate in nonlinear equations, but these equations are not constraints in the sense of CSP. Thanks, Ankesh On Fri, Oct 12, 2012 at 2:30 PM, Martin <mjb@...> wrote: > On Fri, 20121012 at 13:38 0400, Ankesh wrote: > > Hello, > > > > I am a PhD student at Rensselaer Polytechnic Institute, Troy, NY. I am > > looking to implement a modelbuilder for continuouschange event > > calculus (CCEC) [1]. Answer set solvers have been deployed to build > > models for discretetime event calculus [2]. We would like to use > > solvers to implement a reasoner for CCEC. > > > > As I understand, claspd doesn't support real numbers. Is there an ASP > > solver that supports real numbers and real arithmetic, which we could > > use? > > I do not know of one that has native support. It depends a little on > what kind of real number support you need. Linear? Nonlinear? > Bounded? Do you need the numbers to be dense (i.e. you can always pick > a point between two others)? It may be sufficient to use integrations > with constraint solvers like clingcon. Alternatively you might want to > look at some of the integrations with SMT solvers as there are some > solvers that support theories of (linear or nonlinear) reals. > > > (As far as we know, DLVcomplex supports real numbers and is available > > for free, but the code is not open.) > > As far as I know, none of the DLV varients have code available. I > believe this to be a deliberate policy decision by the group. > > > P.S. We expect that given any CCEC description, finiteness of the > > domain can be guaranteed with some preprocessing (ignoring some > > nuances related to floatingpoint arithmetic, for now). > > > > [1] Reasoning about Discontinuities in the Event Calculus > > (1996). http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.9979 > > > > [2] http://reasoning.eas.asu.edu/ecasp/ > > You might also be interested in: > > http://www.doc.ic.ac.uk/~mtc06/report.pdf > > HTH > > Cheers, >  Martin > > > 
From: Martin <mjb@cs...>  20121012 18:31:00

On Fri, 20121012 at 13:38 0400, Ankesh wrote: > Hello, > > I am a PhD student at Rensselaer Polytechnic Institute, Troy, NY. I am > looking to implement a modelbuilder for continuouschange event > calculus (CCEC) [1]. Answer set solvers have been deployed to build > models for discretetime event calculus [2]. We would like to use > solvers to implement a reasoner for CCEC. > > As I understand, claspd doesn't support real numbers. Is there an ASP > solver that supports real numbers and real arithmetic, which we could > use? I do not know of one that has native support. It depends a little on what kind of real number support you need. Linear? Nonlinear? Bounded? Do you need the numbers to be dense (i.e. you can always pick a point between two others)? It may be sufficient to use integrations with constraint solvers like clingcon. Alternatively you might want to look at some of the integrations with SMT solvers as there are some solvers that support theories of (linear or nonlinear) reals. > (As far as we know, DLVcomplex supports real numbers and is available > for free, but the code is not open.) As far as I know, none of the DLV varients have code available. I believe this to be a deliberate policy decision by the group. > P.S. We expect that given any CCEC description, finiteness of the > domain can be guaranteed with some preprocessing (ignoring some > nuances related to floatingpoint arithmetic, for now). > > [1] Reasoning about Discontinuities in the Event Calculus > (1996). http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.9979 > > [2] http://reasoning.eas.asu.edu/ecasp/ You might also be interested in: http://www.doc.ic.ac.uk/~mtc06/report.pdf HTH Cheers,  Martin 
From: Linas Vepstas <linasvepstas@gm...>  20121012 18:15:29

I don't know the answer to the questions below, but want to add a related question: is there any work to add support for anything resembling "satisfiability modulo theories" concepts from model theory? I notice that there are many SMT solvers out there, and they seem to overlap ASP at least partly in function. I found ASP much easier to use though.  linas On 12 October 2012 12:38, Ankesh <ankesh@...> wrote: > Hello, > > I am a PhD student at Rensselaer Polytechnic Institute, Troy, NY. I am > looking to implement a modelbuilder for continuouschange event calculus > (CCEC) [1]. Answer set solvers have been deployed to build models for > discretetime event calculus [2]. We would like to use solvers to implement > a reasoner for CCEC. > > As I understand, claspd doesn't support real numbers. Is there an ASP > solver that supports real numbers and real arithmetic, which we could use? > > (As far as we know, DLVcomplex supports real numbers and is available for > free, but the code is not open.) > > Thanks and regards, > > Ankesh > > P.S. We expect that given any CCEC description, finiteness of the domain > can be guaranteed with some preprocessing (ignoring some nuances related > to floatingpoint arithmetic, for now). > > [1] Reasoning about Discontinuities in the Event Calculus (1996). > http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.9979 > > [2] http://reasoning.eas.asu.edu/ecasp/ > > >  > Don't let slow site performance ruin your business. Deploy New Relic APM > Deploy New Relic app performance management and know exactly > what is happening inside your Ruby, Python, PHP, Java, and .NET app > Try New Relic at no cost today and get our sweet Data Nerd shirt too! > http://p.sf.net/sfu/newrelicdev2dev > _______________________________________________ > Potasscousers mailing list > Potasscousers@... > https://lists.sourceforge.net/lists/listinfo/potasscousers > > 
From: Ankesh <ankesh@gm...>  20121012 17:38:33

Hello, I am a PhD student at Rensselaer Polytechnic Institute, Troy, NY. I am looking to implement a modelbuilder for continuouschange event calculus (CCEC) [1]. Answer set solvers have been deployed to build models for discretetime event calculus [2]. We would like to use solvers to implement a reasoner for CCEC. As I understand, claspd doesn't support real numbers. Is there an ASP solver that supports real numbers and real arithmetic, which we could use? (As far as we know, DLVcomplex supports real numbers and is available for free, but the code is not open.) Thanks and regards, Ankesh P.S. We expect that given any CCEC description, finiteness of the domain can be guaranteed with some preprocessing (ignoring some nuances related to floatingpoint arithmetic, for now). [1] Reasoning about Discontinuities in the Event Calculus (1996). http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.9979 [2] http://reasoning.eas.asu.edu/ecasp/ 
From: Santiago Videla <santiago.videla@ir...>  20121002 16:23:53

Hi Roland, Thanks for this confirmation. Indeed, I found a bug in my code when reading/parsing the (suboptimal) answer sets. That's why I was computing a different score from the minimization statement. Regards, On Oct 1, 2012, at 7:10 PM, Roland Kaminski wrote: > On Monday 01 October 2012 18:29:15 Santiago Videla wrote: >> Hi, >> >> I'm running a minimization problem with gringo 3.0.3 and clasp 2.1. I'm >> interested not only in finding all the optimal models but all the >> suboptimal models for a given tolerance. For doing this I'm using: >> optall = OPTSCORE + OPTSCORE * TOLERANCE >> >> The problem is that when I do this, I'm finding that the score assigned to >> suboptimal models is different from what correspond to the minimization >> statement. I mean, for some models the score is exactly the sum of the >> weighed atoms in the minimization statement, but for others is something >> else. Is this possible? Is there any internal optimization made by gringo >> or clasp that could change the score? Is it possible to switch off this >> optimization? > gringo rewrites optimization statements in certain situations: >  negative weights >  maximize statements > > If you do not use any of the above, then the optimizations statement should > arrive at clasp "as encoded". The rewriting is necessary because the lparse > does not support the two features. > > In your case you might be able to not use optimization statements at all. Just > use weight constraints to achieve what you want: > > % #minimize [ a=1, b=2, c=1, d=4, e=1, f=4 ]. > % wich becomes > #const estimate = OPTSCORE + OPTSCORE * TOLERANCE. > : estimate+1 [ a=1, b=2, c=1, d=4, e=1, f=4 ]. > > Regards, Roland >  > Got visibility? > Most devs has no idea what their production app looks like. > Find out how fast your code is with AppDynamics Lite. > http://ad.doubleclick.net/clk;262219671;13503038;y? > http://info.appdynamics.com/FreeJavaPerformanceDownload.html_______________________________________________ > Potasscousers mailing list > Potasscousers@... > https://lists.sourceforge.net/lists/listinfo/potasscousers 
From: Roland Kaminski <kaminski@cs...>  20121001 17:30:11

On Monday 01 October 2012 18:29:15 Santiago Videla wrote: > Hi, > > I'm running a minimization problem with gringo 3.0.3 and clasp 2.1. I'm > interested not only in finding all the optimal models but all the > suboptimal models for a given tolerance. For doing this I'm using: > optall = OPTSCORE + OPTSCORE * TOLERANCE > > The problem is that when I do this, I'm finding that the score assigned to > suboptimal models is different from what correspond to the minimization > statement. I mean, for some models the score is exactly the sum of the > weighed atoms in the minimization statement, but for others is something > else. Is this possible? Is there any internal optimization made by gringo > or clasp that could change the score? Is it possible to switch off this > optimization? gringo rewrites optimization statements in certain situations:  negative weights  maximize statements If you do not use any of the above, then the optimizations statement should arrive at clasp "as encoded". The rewriting is necessary because the lparse does not support the two features. In your case you might be able to not use optimization statements at all. Just use weight constraints to achieve what you want: % #minimize [ a=1, b=2, c=1, d=4, e=1, f=4 ]. % wich becomes #const estimate = OPTSCORE + OPTSCORE * TOLERANCE. : estimate+1 [ a=1, b=2, c=1, d=4, e=1, f=4 ]. Regards, Roland 
From: Santiago Videla <santiago.videla@ir...>  20121001 16:29:34

Hi, I'm running a minimization problem with gringo 3.0.3 and clasp 2.1. I'm interested not only in finding all the optimal models but all the suboptimal models for a given tolerance. For doing this I'm using: optall = OPTSCORE + OPTSCORE * TOLERANCE The problem is that when I do this, I'm finding that the score assigned to suboptimal models is different from what correspond to the minimization statement. I mean, for some models the score is exactly the sum of the weighed atoms in the minimization statement, but for others is something else. Is this possible? Is there any internal optimization made by gringo or clasp that could change the score? Is it possible to switch off this optimization? Regards, 