ONTOCADE: UM AMBIENTE CASE BASEADO EM ONTOLOGIAS PARA ANÁLISE E PROJETO NA ENGENHARIA DE DOMÍNIO MULTIAGENTE

Abstract

This work proposes ONTOCADE (Ontology-based Environment for Computer- Aided Domain Engineering), an ontology-based CASE environment for analysis and design in Multi-Agent Domain Engineering. Success in Multi-Agent Domain Engineering, the process of creating reusable software abstractions for the development of a set of multi-agent software applications in a domain or problem-solving area, depends on the availability of appropriate methodologies and tools providing adequate support for ontologies, communications, mobility, and autonomy, among others central topics of agent-based systems. ONTOCADE supports the application of the phases of MADEM, a methodology for Multi-Agent Domain Engineering. According to the rules of integration for composing an environment, its tools use a shared data model and incorporate the knowledge of methods for coordinating and guiding the execution of the phases of the software development lifecycle. For that, ONTOCADE is based on ontologies, structures particularly appropriate for representing knowledge and high-level software abstractions, because they present clear and unambiguous terminology. ONTOCADE uses a generic ontology, ONTOMADEM, for guiding the user in modeling tasks. To define ONTOMADEM, it was used the MADEM knowledge to generate a semantic network with the representation of the concepts of the methodology. For its design, the ONTOMADEM was generated by mapping the semantic network to an ontology represented by a hierarchy of classes. To function as a runtime environment, it was chosen Protégé, an extensible platform for creating knowledge-based applications. ONTOMADEM is an extension of the meta-ontology of Protégé. In this way, ONTOCADE executes as a plugin, that is, an aggregated component, which takes advantages of all the functionalities provided by Protégé and offers specifics functions, such as domain and architectural modeling.