Engineering Self-Organising Systems Nature-Inspired Approaches to So
As information handling systems get more and more complex, it becomes increasingly difficult to manage them using traditional approaches based on centralized and pre-defined control mechanisms. Over recent years, there has been a significant increase in t
- PDF / 4,391,379 Bytes
- 309 Pages / 430 x 660 pts Page_size
- 31 Downloads / 151 Views
Abstract. A self-organising system functions without central control, and through contextual local interactions. Components achieve a simple task individually, but a complex collective behaviour emerges from their mutual interactions. Such a system modifies its structure and functionality to adapt to changes to requirements and to the environment based on previous experience. Nature provides examples of self-organisation, such as ants food foraging, molecules formation, or antibodies detection. Similarly, current software applications are driven by social interactions (negotiations, transactions), based on autonomous entities or agents, and run in highly dynamic environments. The issue of engineering applications, based on the principles of selforganisation to achieve robustness and adaptability, is gaining increasing interest in the software research community. The aim of this paper is to survey natural and artificial complex systems exhibiting emergent behaviour, and to outline the mechanisms enabling such behaviours. Keywords: Self-organisation, self-organising application, emergence, collective behaviour, multi-agent systems.
1
Introduction
The study of self-organising systems is a field that has been explored at least since 1953 with the work done by Grass´e [25], who studied the behaviour of insect societies. Many systems in nature demonstrate self-organisation, such as planets, cells, organisms and societies. All these systems exhibit recurrent properties inherent to self-organisation. The simplest form of self-organisation can be achieved by the arrangement of parts of a system in such a way as to be non-random. Considerable research has already been undertaken to study such systems. Self-organising systems are often encountered in many scientific areas including biology, chemistry, geology, sociology, and information technology. A large number of software self-organising systems are designed based on natural mechanisms of self-organisation. Furthermore, recent research has been oriented towards introducing self-organisation mechanisms specifically for software applications, as well as entire development techniques supporting selforganisation [19]. This trend originates from the fact that current software applications need to cope with requirements and constraints stemming from the G. Di Marzo Serugendo et al. (Eds.): AAMAS 2003 Ws ESOA, LNAI 2977, pp. 1–19, 2004. c Springer-Verlag Berlin Heidelberg 2004
2
Giovanna Di Marzo Serugendo et al.
increased dynamism, sophisticated resource control, autonomy and decentralisation inherent in contemporary business and social environments. The majority of these characteristics and constraints are the same as those which can be observed in natural systems exhibiting self-organisation. This survey firstly defines self-organisation, and presents examples of selforganising systems taken from natural life. Subsequently, it describes the different mechanisms enabling social, natural and artificial organisations to achieve a coherent global behaviour. Finally, it reviews seve
Data Loading...
