Dynamic composition of service oriented multi-agent system in self-organized environments

The increasing relevance of complex systems in dynamic environments has received special attention from researchers during the last decade. Due to the need of a flexible and quick response to the clients' requirements, such systems become an important challenge. In this paper, self-organizing mechanisms capable to compose services in an automatic, flexible and decentralized manner are presented, mostly in which their adaptive behavior is concerned. Due to the distributed approach, we also investigate the adaptation regarding the structure of each entity. We thus propose an innovative self-learning mechanism that allows the distributed entities to learn structural relations allowing the system's evolution. This hypothesis were explored and validated by implementing a multi-agent system, in accordance with trust mechanisms to improve the interaction of agents. The achieved results show the correct agent's states in which the agents must evolve and self-organize, improving the system benefits band increasing the organization performance.