Nonlinear Sciences – Adaptation and Self-Organizing Systems
Scientific paper
2004-03-28
Nonlinear Sciences
Adaptation and Self-Organizing Systems
PhD. thesis (169 pages, 1 tex file, 89 ps and eps files, for better resolution images of Figures 1.1-1.4 contact the author)
Scientific paper
I analyse a model of an evolving network represented as a directed graph; each node corresponds to one molecular species and the links to catalytic interactions between species. Over short timescales the graph remains fixed while relative populations of the molecular species change according to a set of coupled differential equations. Over long timescales the system is subject to periodic perturbations, each of which adds one new node to the graph, with random links to other nodes, and removes one node with the least relative population. Starting from a sparse random graph, a small autocatalytic set (ACS) inevitably forms and then grows by accreting nodes until it spans the entire graph. The resultant fully autocatalytic graph, whose probability of forming by pure chance is very small, nevertheless forms in this model in an average time that grows only logarithmically with the size of the system. ACSs can also get destroyed, often accompanied by the sudden extinction of a large number of species. I show that the largest of the extinction events in this model are caused by one of three mechanisms, each of which produces a specific discontinuous change in the graph's topology. The model is analytically tractable: two theorems are proved which determine the set of nodes with least relative population in the attractor, for any given graph. This in turn can be used to analytically demonstrate the inevitability of the formation and growth of ACSs and calculate the associated timescales. Finally, I show that the formation and growth of ACSs is robust to the relaxation of many of the idealizations made to enhance the analytical tractability of the model.
No associations
LandOfFree
Formation and Destruction of Autocatalytic Sets in an Evolving Network Model does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Formation and Destruction of Autocatalytic Sets in an Evolving Network Model, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Formation and Destruction of Autocatalytic Sets in an Evolving Network Model will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-381405