Decentralised Structural Adaptation in Agent Organisations

Abstract

Abstract. Autonomic computing is being advocated as a tool for maintaining and managing large and complex computing systems. Self-organising multi-agent systems provide a suitable paradigm for developing such autonomic systems. Towards this goal, we demonstrate a robust, decentralised approach for structural adaptation in explicitly modelled problem solving agent organisations. Our method is based on self-organisation principles and enables the agents to modify the organisational structure to achieve a better allocation of tasks across the organisation in a simulated task-solving environment. The agents forge and dissolve relations with other agents using their history of interactions as guidance. We empirically show that the efficiency of organisations using our approach is close to that of organisations having an omniscient central allocator and considerably better than static organisations or those changing the structure randomly.

Decentralized approaches for self-adaptation in agent organizations

“…However, we mainly build on the ideas presented by Jin and Levitt [1996], Ferber and Gutknecht [1998] and Hannoun et al [2000]. A preliminary version of the consolidated model was presented by Kota et al [2008]. However, here we extend it by making the organisation open and dynamic (see Section 3.2).…”

Capability-based task allocation in emergency-response environments: A coalition formation approach

“…The organizational structure defines the informational, controlling, and communicational patterns and features of the task environment [8][9][10]. In this research, following the TOMACS, the initial structure of the organization is formed once the system begins to work and reorganization occurs during the system operation, along with the occurrence of reorganization triggers [11].…”

Organisational adaptation of multi-agent systems in a peer-to-peer scenario

“…It also includes information about which action (Act) it is performing on each piece, such as serving, receiving or no action. Equation 34 contains these denitions note that in our current P2P implementation, there is only possession/activity information about a single piece.…”

“…In addition to previous organisation-centred approaches, there are several works [42,40,17,34] that focus on the emergence and/or adaptation of norms from an agent-centred perspective (ACMAS). However, these approaches use methods that depend on participants' implementation or that assume cooperation among agents or that are driven by individual goals that may not be aligned with the social welfare.…”

“…This service is particularly useful when environmental or agent population changes distract the organisation from its goals. As a matter of fact, organisation adaptation has been studied for more than a decade now [15,26,28,25] and it still constitutes a major research topic [34,44], since it can help to obtain the system's expected performance under changing circumstances and/or in dynamic environments. Runtime adaptation is also in line with the computational organisational theory, which claims that the best organisation designs are domain and context dependent [13].…”