Uma arquitetura multiagente para sistemas de supervisão e controle de processos

• Ano de defesa: 2013
• Autor(a) principal: Queiroz, Jonas Felipe Pereira de [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Computação; Sistemas multiagentes; Agentes inteligentes (Software); Ontologias (Recuperação da informação); Multiagent systems
• Link de acesso: http://hdl.handle.net/11449/110382

The technological advances occurred in last years have contributed to increase the use of sensors and controllers in industries. In these environments, the major challenge is the development of complex computer systems to integrate heterogeneous components, to manage and analyze the large amount of produced data, in order to automate the operators’ activities to supervise and control involved processes. In this context, multi-agent approach offers a promising and innovative way to understand, develop, manage and maintain complex systems. The objective of this work is to specify a reference architecture using the multi-agent approach. The architecture was proposed to be used as a reference model for the development of computational systems for processes monitoring and control. The architecture has been specified through an analysis of the aspects and requirements found in a set of related works that use the multi-agent approach in the development of applications in the field of supervision and control. Based on the proposed architecture was developed a system prototype to supervise oil well drilling operations. In this prototype, an ontology was defined to describe knowledge of oil well drilling domain, in order to represent operations, abnormalities and agents. This prototype was tested using a real data set, with the tests was possible to validate some aspects of the proposed architecture