|InterJournal Complex Systems, 1861
|Manuscript Number: |
Submission Date: 2006
|Dual phase evolution – a mechanism for self-organization in complex systems|
One of the successes of complexity theory has been to identify patterns (e.g. networks) and processes (e.g. feedback) that underlie large classes of phenomena from widely different fields (Green 2000, Green & Bransden, 2006). In recent studies (e.g.. Green et al, 2000, Green & Sadedin 2005), we have shown that processes governing species evolution in a landscape are similar in nature to a wide range of critical phenomena, which we term dual phase evolution. Landscapes exhibit two phases - a connected phase, in which selection predominates; and a disconnected phase, in which variation predominates. Disturbances (such as fires, cometary impacts) flip a landscape from connected to disconnected phases, leading both to the isolation of populations necessary for speciation to occur and to the explosive spread of new species. Based on such findings, we argue that evolution within landscapes exemplifies a family of mechanisms that differs from other widely known phenomena, such as self-organized criticality. In essence, our research suggests that underlying self-organization and emergence in many complex systems is a mechanism that incorporates: 1. State spaces that possess dual phases, with variation (exploration) dominant in one phase and selection (exploitation) dominant in the other. 2. Changes between these two phases are mediated by perturbation and criticality. 3. Order accumulates via repeated phase changes; 4. Perturbations often exploit chaos as a source of novelty; 5. Novel order crystallizes via mechanisms that include synchronization, positive feedback and encapsulation. We show that that many processes exhibit parts of this general mechanism. Many optimization algorithms (e.g. simulated annealing) exploit phase changes in the connectivity of the solution landscape to mediate between global search (exploration) and local search (exploitation). We show how the process may work in several systems, including social networks (Bransden and Green 2005) and food webs (Green & Sadedin 2005). Relevant references from our research Bransden, T.G. and Green, D.G. (2005). Getting along with your neighbours - emergent cooperation in networks of adaptive agents. In Ohuchi, A., Suzuki, K., Gen, M. and Green, D.G. (eds). Asia-Pacific Symposium on Intelligent and Evolutionary Systems (IES2005). Future University-Hakodate, Japan. Green, D.G. (2000). Self-Organization in complex systems. In Complex Systems. T.J. Bossomaier & D.G. Green (eds) Cambridge University Press, pp. 7-41. Green, D.G., Newth. D. and Kirley, M. (2000). Connectivity and catastrophe - towards a general theory of evolution. In M.A. Bedau et al. (eds.) Artificial Life VII: Proceedings of the Seventh International Conference. pp 153-161. MIT Press. Green, D.G. and Bransden, T.G. (2006). Complexity theory. In McGraw-Hill Encyclopedia of Science and Technology. McGraw-Hill, New York. pp. 507-511. Green, D.G. and Sadedin, S. (2005). Interactions matter - Complexity in landscapes and ecosystems. Ecological Complexity 2, 117-130. Green, D.G., Klomp, N.I., Rimmington, G.R. & Sadedin, S. (2005). Complexity in Landscape Ecology. Springer, Amsterdam.
|Submit referee report/comment|