g-BDI: A Graded Intensional Agent Model for Practical Reasoning Ana Casali 1 , Llu´ ıs Godo 2 , and Carles Sierra 2 1 Dept. of Computer Science, Universidad Nacional de Rosario (UNR) Centro Intl. Franco-Argentino de Ciencias de la Informaci´on y de Sistemas (CIFASIS) Av Pellegrini 250, 2000 Rosario, Argentine acasali@fceia.unr.edu.ar 2 Institut d‘Investigaci´o en Intel·lig` encia Artificial (IIIA) - CSIC Campus Universitat Aut`onoma de Barcelona s/n 08193 Bellaterra (Catalunya), Spain {godo,sierra}@iiia.csic.es Abstract. In intentional agents, actions are derived from the mental attitudes and their relationships. In particular, preferences (positive de- sires) and restrictions (negative desires) are important proactive atti- tudes which guide agents to intentions and eventually to actions. In this paper we overview recent developments about a multi-context based agent architecture g-BDI to represent and reasoning about gradual no- tions of desires and intentions, including sound and complete logical for- malizations. We also show that the framework is expressive enough to describe how desires, together with other information, can lead agents to intentions and finally to actions. As a case-study, we will also describe the design and implementation of recommender system on tourism as well as the results of some experiments concerning the flexibility and performance of the g-BDI model. 1 Introduction In the recent past, an increasing number of theories and architectures have been proposed to provide multiagent systems a formal support for their reasoning and decision making models, among them the so-called BDI architectures [16,9,15]. We consider that making the BDI architecture more flexible will allow for de- signing and developing intensional agents potentially capable to have a better performance in uncertain and dynamic environments. Along this research line we have developed a general model for graded BDI agents (g-BDI agents for short), specifying an architecture able to deal with the environment uncertainty (via graded beliefs) and with graded mental proactive attitudes (via desires and intentions). In the g-BDI model, belief degrees represent the extent to which the agent believes formulas hold true. Degrees of positive or negative desires allow the agent to set different levels of preference or rejection respectively. In- tention degrees also give a preference measure but, in this case, modelling the V. Torra, Y. Narukawa, and M. Inuiguchi (Eds.): MDAI 2009, LNAI 5861, pp. 5–20, 2009. c  Springer-Verlag Berlin Heidelberg 2009