Uma ontologia de tarefa de investigação para uso de agentes inteligentes

Abstract

Currently, many computer applications have benefited from the use of intelligent agents. An intelligent agent can be understood from the perspective of a software component that performs certain tasks. Thus, there are different types of intelligent agents and choosing the right type depends on the environment and the tasks that it performs. In turn, knowledge of tasks can be represented through ontologies. The ontology of tasks describes a generic task related vocabularies that are independent of the domain of expertise. For example, a task of investigation can occur in different areas and can be performed both by human agents as by software agents. In this context, this dissertation aims to create a task ontology covering structural and behavioral aspects of investigative environment used by intelligent agents. For that purpose was served, was made a bibliographical survey of publications involving the major issues involved, namely: ontologies, research tasks and intelligent agents. To evaluate the proposed ontology was developed a prototype applied to the police domain, where intelligent agents perform collections and analyses of data sources with the intention of generating an information interlacement that helps the human’s investigators in their decision-making. As a result, this work presents generic investigation task ontology for use of intelligent agents and a prototype applied to the field of policing aimed at assisting the development of more complex applications related to the police investigation.