Loading...
Publication
Integração de Emoção e Raciocínio em Agentes Inteligentes
Authors
Advisor(s)
Abstract(s)
This thesis addresses the influence of emotional phenomena on reasoning and decision-making processes of intelligent agents, in order to make possible the use of agents with this kind of cognitive abilities in concrete environments where real-time behavior is needed. The following are specific contributions of this thesis: an emotion model to support the modeling of emotional phenomena in intelligent agents, where the dynamics underlying those phenomena are emphasized; and an agent model able to support the modeling of the relation between emotional and cognitive phenomena based on the dynamic and continuous nature of that relation. These models support the implementation of mechanisms for regulation and adaptation of the reasoning and decision-making processes, which are able to take advantage of the relation between emotional and cognitive phenomena to focus the cognitive activity. These mechanisms enable an agent to control the computational resources used in cognitive activity and the time expended in that activity. They also enable an agent to take advantage of past experiences to anticipate future situations, through the formation of autobiographic emotional memories and through the exploration of those memories by prospective reasoning. A generic agent architecture is proposed that supports the implementation of agents of different types and levels of complexity integrating emotional and cognitive aspects. As an experimental support, three prototypes are presented, aimed at different application contexts: changing environment dynamism context; sensory and time restrictions context; social reasoning context.
Description
Keywords
Intelligent agents artificial intelligence multi-agent systems cognitive models emotion models adaptation learning resource-bounded reasoning