A Generic Agent Organisation Framework for Autonomic Systems

Autonomic computing is being advocated as a tool for managing large, complex computing systems. Specifically, self-organisation provides a suitable approach for developing such autonomic systems by incorporating self-management and adaptation properties into large-scale distributed systems. To aid in this development, this paper details a generic problem-solving agent organisation framework that can act as a modelling and simulation platform for autonomic systems. Our framework describes a set of service-providing agents accomplishing tasks through social interactions in dynamically changing organisations. We particularly focus on the organisational structure as it can be used as the basis for the design, development and evaluation of generic algorithms for self-organisation and other approaches towards autonomic systems.