Abstract
The majority of BDI agent implementations perform means-end reasoning using statically defined plan libraries, thus limiting their flexibility at runtime. One of the reasons for this limiting choice in the implementation is the fact that deciding whether an agent is able to accomplish its goal or not is proven to be a very complex problem. On the other hand, advances in the field of propositional planning have enabled a larger class of problems to be tractable. Therefore, this
dissertation aims to describe the relationship between BDI agents and propositional planning. In order to accomplish such goal, we describe a mapping between the mental states used in the BDI model and propositional planning formalisms. The mapping thus defined is implemented in an extension to the X-BDI agent model that includes Graphplan as its propositional planning
algorithm. Finally, the implemented prototype is used to model and execute a number of case studies.
dissertation aims to describe the relationship between BDI agents and propositional planning. In order to accomplish such goal, we describe a mapping between the mental states used in the BDI model and propositional planning formalisms. The mapping thus defined is implemented in an extension to the X-BDI agent model that includes Graphplan as its propositional planning
algorithm. Finally, the implemented prototype is used to model and execute a number of case studies.
| Translated title of the contribution | Propositional Planning in BDI Agents |
|---|---|
| Original language | Portuguese |
| Type | Thesis |
| Publisher | Pontifícia Universidade Católica do Rio Grande do Sul |
| Publication status | Published - 2005 |
| Externally published | Yes |