K2: uma arquitetura para a adaptação de agentes de software ao contexto

• Autor(a) principal: Lemke, Ana Paula
• Data de Publicação: 2011
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional PUCRS
• Texto Completo: http://hdl.handle.net/10923/1573
• topic: INFORMÁTICA; SISTEMAS MULTIAGENTES; ARQUITETURA DE COMPUTADOR

Agent technology is increasingly seen as an attractive approach to develop applications for pervasive environments [Gun08b]. However, many existing agent platforms support only the development of agents able to deal with a limited set of situations (the agents need to be reprogrammed when faced with unexpected situations). Software development for complex environments such as pervasive environments is difficult since the developers cannot predict every possible execution context the application will have at design time. So, it is necessary to develop software agents capable of learning and adapting to meet the conditions of a new environment, or a new context. Considering this issue, this thesis proposes an architecture to create context adaptative agents – the K2 architecture. The idea is to provide an architecture that supports the modification of structural elements of a software agent, adapting its behavior and structure according to perceived changes in the context. One of the characteristics of the proposed architecture is the separation between adaptative and non-adaptative (default) agent behaviors, which will be accomplished by using elements named adaptors, whose deployment is supported by aspect-oriented programming. Three examples illustrate the architecture feasibility and applicability.