چكيده به لاتين
Holonic structures are hierarchical constructions of holons that have been suggested and developed for the solution of complex problems, and satisfying objectives such as limiting interaction domains, decreasing uncertainty, or the formation of high-level goals in multi-agent systems; in order to lever- age their flexibility in dynamic environments.
Although holonic multi-agent systems have been deployed widely in the modeling and solution of complex problems, most of their key concepts, like the formation of holons and their dynamic orga- nization, make use of very basic and domain dependent models, due to the novelty of the research in this field. In this dissertation, an attempt is made to utilize social concepts to propose a novel method for the organization of holonic structures.
The outcomes of this research are contributions in both static and dynamic holonic multi-agent systems. In the static approach, a goal hierarchy of the system is assumed and the concepts from organizational theory have been used to construct a holonic structure inside a deployed multi-agent system, based on their roles and skill. The efficiency of the suggested method was assessed in object tracking application of wireless sensor network; and according to the empirical results, it has been successful in managing the energy level of the sensors.
In the second part of the research, a comprehensive distributed and dynamic method is proposed. This approach is based on many social concepts such as social diversity, social exchange theory, social norms, and social psychology. Beside, it has been empowered by means of learning methods to manage the dynamic environments. Various parts of the proposed method have been evaluated in application environments of function optimization and task assignments. The results prove the high capability of the contribution in large scaled distributed problems.
