A modelagem do aprendiz em ambientes de aprendizagem por computador baseados em atividades de resolução de problemas.

Abstract

The aim of this dissertation is to specify the learner modeling process for tutor agents of the MATHEMA architect ure. MATHEMA architecture is a framework for the development of computer learning environments based on problem solving activities. The specification is done following two lines of action. First, the establishment of general directions for learner modeling in MATHEMA agents. Such directions summarize what informations about the learners can be modeled: how to represent, acquire and maintain these informations in learner model: and mainly, what decisions can be made by MATHEMA agents having a student model. Second, the definition of a particular Learner Modeling System to be used in MATHEMA agents. This Learner Modeling- System is specified by means of VDM-SL formal specification language. Its main features are: (1) the learner model representation done by an overlay of MATHEMA agents Tutor System Curriculum: (2) the acquisition of informations done by a model tracing process based on problem resolution; (3) the student model maintenance done by a maintenance process on fuzzy overlay models. Finally, it is shown how the Learner Model System represented, acquired and maintained informations can effectively be used by MATHEMA agents Tutor System during selection of problems to be passed to learners.