Home // International Journal On Advances in Intelligent Systems, volume 8, numbers 3 and 4, 2015 // View article

Structural Adaptation for Self-Organizing Multi-Agent Systems: Engineering and Evaluation

Authors:
Thomas Preisler
Wolfgang Renz
Tim Dethlefs

Keywords: Self-Organizing Systems; Multi-Agent Systems; Structural Adaptation; Decentralized Coordination; Engineering Approach

Abstract:
Over one decade of research in engineering of selforganization (SO) has established SO as a decentralized way to develop self-adaptive systems. However, such SO systems may show unwanted behavior under certain conditions, e.g., a decrease of performance and/or starvation, even when apparently well engineering. A promising concept to overcome such dynamical in-efficiencies in SO systems is to realize the dynamic exchange or reconfiguration of the coordination processes responsible for the self-organizing behavior in terms of a structural adaptation. In this paper, we propose an architecture and engineering approach to support the self-adaptive, structural exchange (or reconfiguration) of self-organizing coordination processes based on distributed Multi-Agent technology. Here, a sensor in each agent detects any decrease of specified SO performance indicators, which initiates a distributed adaptation process that allows the exchange (or reconfiguration) of the self-organizing coordination processes, enabling the system to adapt to changing conditions automatically. The proposed engineering approach is explained exemplary and evaluated using a case study in which a set of autonomous robots is needed to be coordinated in order to achieve their collaborative goal of mining commodities. Based on two different coordination strategies, it will be shown how the autonomous entities benefit from self-organizing coordination processes that support the emergent formation of local mining groups among the robots, and how the concept of structural adaptation helps to overcome a failure that negatively influences these coordination processes

Pages: 413 to 425

Copyright: Copyright (c) to authors, 2015. Used with permission.

Publication date: December 30, 2015

Published in: journal

ISSN: 1942-2679