Defeasible-argumentation-based multi-agent planning
[EN] This paper presents a planning system that uses defeasible argumentation to reason about context information during the construction of a plan. The system is designed to operate in cooperative multi-agent environments where agents are endowed with planning and argumentation capabilities. Planni...
| Autores: | , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2017 |
| País: | España |
| Institución: | Universitat Politècnica de València (UPV) |
| Repositorio: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia |
| Idioma: | inglés |
| OAI Identifier: | oai:riunet.upv.es:10251/99167 |
| Acceso en línea: | https://riunet.upv.es/handle/10251/99167 |
| Access Level: | acceso abierto |
| Palabra clave: | Defeasible argumentation Multi-agent planning Multi-agent systems Cooperation LENGUAJES Y SISTEMAS INFORMATICOS |
| Sumario: | [EN] This paper presents a planning system that uses defeasible argumentation to reason about context information during the construction of a plan. The system is designed to operate in cooperative multi-agent environments where agents are endowed with planning and argumentation capabilities. Planning allows agents to contribute with actions to the construction of the plan, and argumentation is the mechanism that agents use to defend or attack the planning choices according to their beliefs. We present the formalization of the model and we provide a novel specification of the qualification problem. The multi-agent planning system, which is designed to be domain-independent, is evaluated with two planning tasks from the problem suites of the International Planning Competition. We compare our system with a non-argumentative planning framework and with a different approach of planning and argumentation. The results will show that our system obtains less costly and more robust solution plans. |
|---|