A JADE-based tool to assess the network dynamics of mechatronic multiagent systems

• Autor(a) principal: Moura, Catarina Viegas de Freitas
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/14109
• topic: Inference tool; Multiagent systems; Self-organization; Agent's interactions; Agent's networks; Agent's networks visualization

The increase in requirements and diversified demand in the fields of production has been triggering constant changes in manufacturing areas. This scenario has prevented traditional systems to meet the needed requirements to support these changes. New paradigms have been emerging in the fields of manufacturing in order to support the evolution of manufacturing. These paradigms intend to supply modern production systems with a better responsiveness, flexibility, robustness and reconfiguration. However, systems with these characteristics imply the need to predict their behavior in various production scenarios in order to avoid failures. In this sense, it is necessary to study and understand the behavior of these systems. Methods that allow the study of distributed systems are still very scarce. This thesis proposes a multiagent architecture that supports a tool to infer the agents’ network in any multiagent system. The developed tool, through the networks that infers, allows to study the interactions’ dynamics, often hidden in this type of systems, that contributes to the emergence of desired and undesired features. The proposed tool infers multiagent networks based on the interactions that occur between agents in real time and codifies the information resultant of those interactions in a compatible format with specialized tools for network analysis. To prove and test the effectiveness and applicability of the developed tool, an instance of an Evolvable Production System was simulated and analyzed by the tool. The usage of tool in the study and observation of the system allowed to test the performance of the tool in a highly dynamic environment and to validate the system’s operation.