Adaptation Noise And Self Organizing Systems

A cellular game is a dynamical system in which cells, placed in some discrete structure, are regarded as playing a game with their immediate neighbors. Individual strategies may be either deterministic or stochastic. Strategy success is measured according to some universal and unchanging criterion. Successful strategies persist and spread; unsuccessful ones disappear. In this thesis, two cellular game models are formally defined, and are compared to cellular automata. Computer simulations of these models are presented. Conditions providing maximal average cell success, on one and two-dimensional lattices, are examined. It is shown that these conditions are not necessarily stable; and an example of such instability is analyzed. It is also shown that Nash equilibrium strategies are not necessarily stable. Finally, a particular kind of zero-depth, two-strategy cellular game is discussed; such a game is called a simple cellular game. It is shown that if a simple cellular game is left/right symmetric, and if there are initially only finitely many cells using one strategy, the zone in which this strategy occurs has probability 0 of expanding arbitrarily far in one direction only. With probability 1, it will either expand in both directions or disappear. Computer simulations of such games are presented.

This paper shows how to determine all the unidimensional two-state cellular automaton rules of a given number of inputs which conserve the number of active sites. These rules have to satisfy a necessary and sufficient condition. If the active sites are viewed as cells occupied by identical particles, these cellular automaton rules represent evolution operators of systems of identical interacting particles whose total number is conserved. Some of these rules, which allow motion in both directions, mimic ensembles of one-dimensional pseudo-random walkers. Numerical evidence indicates that the corresponding stochastic processes might be non-Gaussian.

The relevance of chaos to evolution is discussed in the context of the origin and maintenance of diversity and complexity. Evolution to the edge of chaos is demonstrated in an imitation game. As an origin of diversity, dynamic clustering of identical chaotic elements, globally coupled each to other, is briefly reviewed. The clustering is extended to nonlinear dynamics on hypercubic lattices, which enables us to construct a self-organizing genetic algorithm. A mechanism of maintenance of diversity, ``homeochaos", is given in an ecological system with interaction among many species. Homeochaos provides a dynamic stability sustained by high-dimensional weak chaos. A novel mechanism of cell differentiation is presented, based on dynamic clustering. Here, a new concept -- ``open chaos" -- is proposed for the instability in a dynamical system with growing degrees of freedom. It is suggested that studies based on interacting chaotic elements can replace both top-down and bottom-up approaches.

In addition to the emergent complexity of patterns that appears when many agents come in interaction, it is also useful to characterize the dynamical processes that lead to their self-organization. A set of ergodic invariants is identified for this purpose, which is computed in several examples, namely a Bernoulli network with either global or nearest-neighbor coupling, a generalized Bak-Sneppen model and a continuous minority model.

This work approaches human chromosome mapping by developing algorithms for ordering markers associated with radiation hybrid data. Motivated by recent work of Boehnke et al. [1], we formulate the ordering problem by developing stochastic spin models to search for minimum-break marker configurations. As a particular application, the methods developed are applied to 14 human chromosome-21 markers tested by Cox et al. [2]. The methods generate configurations consistent with the best found by others. Additionally, we find that the set of low-lying configurations is described by a Markov-like ordering probability distribution. The distribution displays cluster correlations reflecting closely linked loci.

Given a (finite) string of zeros and ones, we report a way to determine if the number of ones is less than, greater than, or equal to a prescribed number by applying two sets of cellular automaton rules in succession. Thus, we solve the general density classification problem using cellular automaton.

Large distributed multiagent systems are characterized by vast numbers of agents trying to gain access to limited resources in an unpredictable environment. Agents in these system continuously switch strategies in order to opportunistically find improvements in their utilities. We have analyzed the fluctuations around equilibrium that arise from strategy switching and discovered the existence of a new phenomenon. It consists of the appearance of sudden bursts of activity that punctuate the fixed point, and is due to an effective random walk consistent with overall stability. This clustered volatility is followed by relaxation to the fixed point but with different strategy mixes from the previous one. This phenomenon is quite general for systems in which agents explore strategies in search of local improvements.

Clustering and correlation effects are frequently observed in chaotic systems in situations where, because of the positivity of the Lyapunov exponents, no dimension reduction is to be expected. In this paper, using a globally coupled network of Bernoulli units, one finds a general mechanism by which strong correlations and slow structures are obtained at the synchronization edge. A structure index is defined, which diverges at the transition points. Some conclusions are drawn concerning the construction of an ergodic theory of self-organization.

A survey of the patterns of synonymous codon preferences in the HIV env gene reveals a relation between the codon bias and the mutability requirements in different regions in the protein. At hypervariable regions in gp120 , one finds a greater proportion of codons that tend to mutate non-synonymously, but to a target that is similar in hydrophobicity and volume. We argue that this strategy results from a compromise between the selective pressure placed on the virus by the induced immune response, which favours amino acid substitutions in the complementarity determining regions, and the negative selection against missense mutations that violate structural constraints of the env protein.

This paper will consider the coevolution of species which are symbiotic in their interaction. In particular, we shall analyse the interaction of squirrels and oak trees, and develop a mathematical framework for determining the coevolutionary equilibrium for consumption and production patterns.