daajoe / wcs2asp-1

The Encoding script

The encoding.py script is used to encode programs into ASP. This script can be executed with different options some of which are required. In general, the script has four parameters:

• -f - followed by the file we want to encode. This file is a .lp file, that contains the logic program that we want to encode. We elaborate later how this files looks [required].
• -o - followed by the observation we want to explain. This parameter is used for the abduction case, where we need to define the observation that we want to explain [optional, but required for the abduction case].
• -s - followed by the semantics we want to do the encoding. We have her e two options, either LS for Łukasiewicz and Gottwald semantics or K for Kleene semantics. The default value is LS [optional].
• -p - followed with Y for Yes or N for No. This parameter is used to differentiate whether we want to compute the supported or all minimal models for the L and S semantics. This parameter has no effect in the abduction case, or when we use the encoding for the K semantics [optional].
• -r - followed by the literal that we want to reason. This parameter is used for the abduction case as well. It is the parameter that tells the literal that we want to reason [optional, but required for the abduction case].

To add a negative literal, we write the classical negation with a - sign. For example if we want to add the negative atom of the atom l, we do with the following option -r="-l".

The .lp file that is used with the -f parameter indicates the logic program that we want to compute the knowledge minimal models of its weak completion. Every rule of the form a ← l_1 ⋀ l_2 ⋀ ... ⋀ l_n is written as a :- l_1, l_2, ...., l_n., every fact a ← T is written as a., and every assumption a ← ⊥ is written as -a.. Every rule ends with a . and lines starting with % are comments and will be ignored by the encoding.

After we run the script, it will generate a .lp file, which then will be executed with respect to the type of problem that we want to solve. The problem that we want to solve is encoded with respect to the mode, from which an additional file is added when we run the solver. We have three modes:

• mode: 1 it refers to the encoding file clingoEncodingMinimalModels.lp (if we want to compute all the minimal models under L, S, or K semantics). This mode is executed when we do not use the -o parameter and the -p parameter is N.
• mode: 2 it refers to the encoding file clingoEncodingSupportedModels.lp (if we want to compute the supported models under L and S semantics). This mode is executed when we want to compute the knowledge minimal models that can be computed with respect to the least fixed point operator Φ.
• mode: 3 it refers to the encoding file clingoForAbductionEncodingMinimalModels.lp (if we want to compute the minimal explanation for the abduction case). This mode is executed for the abduction case.

Remark: the parameters should be used with the corresponding clingo....lp file. For example if we use the clingoForAbductionEncodingMinimalModels.lp file, we need to use the -o and -r parameters as well. These tells which observation we want to explain, and for which literal we want to reason.