Parallel construction of dialectical trees for defeasible logic programming
Defeasible Logic Programming (DLP) [2] is an extension of logic programming based on the defeasible argumentation formalism developed in [12, 11, 6]. In DLP a query q succeeds when a justification for q is found. The defeasible argumentation formalism obtains justifications through a dialectical ana...
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Objeto de conferencia |
| Lenguaje: | Inglés |
| Publicado: |
1999
|
| Materias: | |
| Acceso en línea: | http://sedici.unlp.edu.ar/handle/10915/22245 |
| Aporte de: |
| Sumario: | Defeasible Logic Programming (DLP) [2] is an extension of logic programming based on the defeasible argumentation formalism developed in [12, 11, 6]. In DLP a query q succeeds when a justification for q is found. The defeasible argumentation formalism obtains justifications through a dialectical analysis, where arguments and counterarguments are considered For implementing this dialectical analysis, a dialectical tree is used [6].
In [4], implicitly exploitable parallelism for DLP was studied. Since DLP is an extension of Logic Prograrnming, the four kinds of parallelism already studied for traditional Logic Prograrnming can be exploited [7]: oR-parallelism, independent and dependent AND-parallelism, and also unification parallelism. Besides, there are new sources of parallelism that can be implicitly exploited in a defeasible argumentation system: (1) several arguments for a conclusion q can be constructed in parallel, (2) once an argument A for q is found, defeaters for A can be searched in parallel, and (3) the dialectical analysis between arguments and defeaters, can also be done in parallel. In this work we propose an implementation for exploiting this last source of parallelism. |
|---|