arXiv subjects

H. Waelbroeck

Publications and source records attributed to H. Waelbroeck.

17 recordsLinked to original sources

arXiv2000

Schemata Evolution and Building Blocks

In the light of a recently derived evolution equation for genetic algorithms we consider the schema theorem and the building block hypothesis. We derive a schema theorem based on the concept of effective fitness showing that schemata of higher than average effective fitness receive an exponentially increasing number of trials over time. The equation makes manifest the content of the building block hypothesis showing how fit schemata are constructed from fit sub-schemata. However, we show that generically there is no preference for short, low-order schemata. In the case where schema reconstruction is favored over schema destruction large schemata tend to be favored. As a corollary of the evolution equation we prove Geiringer's theorem.

nlin.AO↗

arXiv2000

Schemata as Building Blocks: Does Size Matter?

We analyze the schema theorem and the building block hypothesis using a recently derived, exact schemata evolution equation. We derive a new schema theorem based on the concept of effective fitness showing that schemata of higher than average effective fitness receive an exponentially increasing number of trials over time. The building block hypothesis is a natural consequence in that the equation shows how fit schemata are constructed from fit sub-schemata. However, we show that generically there is no preference for short, low-order schemata. In the case where schema reconstruction is favoured over schema destruction large schemata tend to be favoured. As a corollary of the evolution equation we prove Geiringer's theorem. We give supporting numerical evidence for our claims in both non-epsitatic and epistatic landscapes.

nlin.AO↗

arXiv1998

Discrete Chaos

We propose a theory of deterministic chaos for discrete systems, based on their representations in binary state spaces $ Ω$, homeomorphic to the space of symbolic dynamics. This formalism is applied to neural networks and cellular automata; it is found that such systems cannot be viewed as chaotic when one uses the Hamming distance as the metric for the space. On the other hand, neural networks with memory can in principle provide examples of discrete chaos; numerical simulations show that the orbits on the attractor present topological transitivity and a dimensional phase space reduction. We compute this by extending the methodology of Grassberger and Procaccia to $ Ω$. As an example, we consider an asymmetric neural network model with memory which has an attractor of dimension $D_a = 2$ for $N = 49$.

chao-dyn↗

arXiv1998

Causal Set Dynamics: A Toy Model

We construct a quantum measure on the power set of non-cyclic oriented graphs of N points, drawing inspiration from 1-dimensional directed percolation. Quantum interference patterns lead to properties which do not appear to have any analogue in classical percolation. Most notably, instead of the single phase transition of classical percolation, the quantum model displays two distinct crossover points. Between these two points, spacetime questions such as "does the network percolate" have no definite or probabilistic answer.

gr-qc↗

arXiv1997

Emergence of Algorithmic Languages in Genetic Systems

In genetic systems there is a non-trivial interface between the sequence of symbols which constitutes the chromosome, or ``genotype'', and the products which this sequence encodes --- the ``phenotype''. This interface can be thought of as a ``computer''. In this case the chromosome is viewed as an algorithm and the phenotype as the result of the computation. In general only a small fraction of all possible sequences of symbols makes any sense for a given computer. The difficulty of finding meaningful algorithms by random mutation is known as the brittleness problem. In this paper we show that mutation and crossover favour the emergence of an algorithmic language which facilitates the production of meaningful sequences following random mutations of the genotype. We base our conclusions on an analysis of the population dynamics of a variant of Kitano's neurogenetic model wherein the chromosome encodes the rules for cellular division and the phenotype is a 16-cell organism interpreted as a connectivity matrix for a feedforward neural network. We show that an algorithmic language emerges, describe this language in extenso, and show how it helps to solve the brittleness problem.

adap-org↗

arXiv1997

Self-Adaptation in Evolving Systems

A theoretical and experimental analysis is made of the effects of self-adaptation in a simple evolving system. Specifically, we consider the effects of coding the mutation and crossover probabilities of a genetic algorithm evolving in certain model fitness landscapes. The resultant genotype-phenotype mapping is degenerate, there being no direct selective advantage for one probability versus another. We show that the action of mutation and crossover breaks this degeneracy leading to an induced symmetry breaking among the genotypic synonyms. We demonstrate that this induced symmetry breaking allows the system to self-adapt in a time dependent environment.

adap-org↗

arXiv1997

Codon Bias and Mutability in HIV Sequences

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.

adap-org↗

arXiv1997

Symmetry Breaking and Adaptation: Evidence from a Toy Model of a Virus

We argue that the phenomenon of symmetry breaking in genetics can enhance the adaptability of a species to changes in the environment. In the case of a virus, the claim is that the codon bias in the neutralization epitope improves the virus' ability to generate mutants that evade the induced immune response. We support our claim with a simple ``toy model'' of a viral epitope evolving in competition with the immune system. The effective selective advantage of a higher mutability leads to a dominance of codons that favour non-synonymous mutations. The results in this paper suggest the possibility of emergence of an algorithmic language in more complicated systems.

adap-org↗

arXiv1996

Analysis of the Effective Degrees of Freedom in Genetic Algorithms

An evolution equation for a population of strings evolving under the genetic operators: selection, mutation and crossover is derived. The corresponding equation describing the evolution of schematas is found by performing an exact coarse graining of this equation. In particular exact expressions for schemata reconstruction are derived which allows for a critical appraisal of the ``building-block hypothesis'' of genetic algorithms. A further coarse-graining is made by considering the contribution of all length-l schematas to the evolution of population observables such as fitness growth. As a test function for investigating the emergence of structure in the evolution the increase per generation of the in-schemata fitness averaged over all schematas of length l, $Δ_l$, is introduced. In finding solutions of the evolution equations we concentrate more on the effects of crossover, in particular we consider crossover in the context of Kauffman Nk models with k=0,2. For k=0, with a random initial population, in the first step of evolution the contribution from schemata reconstruction is equal to that of schemata destruction leading to a scale invariant situation where the contribution to fitness of schematas of size l is independent of l. This balance is broken in the next step of evolution leading to a situation where schematas that are either much larger or much smaller than half the string size dominate over those with $l \approx N/2$. The balance between block destruction and reconstruction is also broken in a k>0 landscape. It is conjectured that the effective degrees of freedom for such landscapes are landscape connective trees that break down into effectively fit smaller blocks, and not the blocks themselves. Numerical simulations confirm this ``connective tree hypothesis'' by showing that correlations drop off with connective distance and not with intrachromosomal distance.

adap-org↗

arXiv1996

Symmetry Breaking and Adaptation: The Genetic Code of Retroviral Env Proteins

Although several synonymous codons can encode the same aminoacid, this symmetry is generally broken in natural genetic systems. In this article, we show that the symmetry breaking can result from selective pressures due to the violation of the synonym symmetry by mutation and recombination. We conjecture that this enhances the probability to produce mutants that are well-adapted to the current environment. Evidence is found in the codon frequencies of the HIV {\it env} protein: the codons most likely to mutate and lead to new viruses resistant to the current immunological attack, are found with a greater frequency than their less mutable synonyms.

adap-org↗

arXiv1996

A Hamiltonian Lattice Theory for Homogeneous Curved Spacetimes in 2+1 Dimensions

We propose an exact Hamiltonian lattice theory for (2+1)-dimensional spacetimes with homogeneous curvature. By gauging away the lattice we find a generalization of the ``polygon representation'' of (2+1)-dimensional gravity. We compute the holonomies of the Lorentz connection ${\bf A}_i = ω_i^a {\bf L}_a + e_i^a {\bf K}_a$ and find that the cycle conditions are satisfied only in the limit $Λ\to 0$. This implies that, unlike in (2+1)-dimensional Einstein gravity, the connection ${\bf A}$ is not flat. If one modifies the theory by taking the cycle conditions as constraints, then one finds that the constraints algebra is first-class only if the Poisson bracket structure is deformed. This suggests that a finite theory of quantum gravity would require either a modified action including higher-order curvature terms, or a deformation of the commutator structure of the metric observables.

gr-qc↗

arXiv1995

Quantization of 2+1 Gravity on the Torus

We use the polygon representation of 2+1--dimensional gravity to explicitly carry out the canonical quantization of a universe with the topology of a torus. The mapping-class-invariant wave function for a quantum ''big bounce'', is reminiscent of the interference patterns of linear gratings. We consider the ``problem of time'' of quantum gravity: for one choice of internal time the universe recovers a semiclassical interpretation after the bounce, with a wave packet centered at a single geometry; for another choice of internal time, the quantum solutions involve interference between macroscopically distinct universes.

gr-qc↗

arXiv1995

Chaotic Dynamics of Binary Systems

We propose a theory of chaos for discrete systems, based on their representation in a space of ``binary histories'', $ {\cal B^{\infty}} $. We show that $ {\cal B^{\infty}} $ is a metrizable Cantor set which embeds the attractor $Λ$, itself also a Cantor set.

chao-dyn↗

arXiv1994

Deterministic Chaos in Tropical Atmospheric Dynamics

We examine an 11-year data set from the tropical weather station of Tlaxcala, Mexico. We find that mutual information drops quickly with the delay, to a positive value which relaxes to zero with a time scale of 20 days. We also examine the mutual dependence of the observables and conclude that the data set gives the equivalent of 8 variables per day, known to a precision of $2\%$. We determine the effective dimension of the attractor to be $D_{eff} \approx 11.7$ at the scale $3.5\% < R/R_{max} < 8\%$. We find evidence that the effective dimension increases as $R/R_{max} \to 0$, supporting a conjecture by Lorenz that the climate system may consist of a large number of weakly coupled subsystems, some of which have low-dimensional attractors. We perform a local reconstruction of the dynamics in phase space; the short-term predictability is modest and agrees with theoretical estimates. Useful skill in predictions of 10-day rainfall accumulation anomalies reflects the persistence of weather patterns, which follow the 20-day decay rate of the mutual information.

comp-gas↗

arXiv1994

Recognition of Temporal Sequences of Patterns

Using an asymmetric associative network with synchronous updating, it is possible to recall a sequence of patterns. To obtain a stable sequence generation with a large storage capacity, we introduce a threshold that eliminates the contribution of weakly correlated patterns. For this system we find a set of evolution equations for the overlaps of the states with the patterns to be recognized. We solve these equations in the limit of the stationary cycle, and obtain the critical value of the capacity as a function of the threshold and temperature. Finally, a numerical simulation is made, confirming the theoretical results.

comp-gas↗

arXiv1994

Homotopy Invariants and Time Evolution in (2+1)-Dimensional Gravity

We establish the relation between the ISO(2,1) homotopy invariants and the polygon representation of (2+1)-dimensional gravity. The polygon closure conditions, together with the SO(2,1) cycle conditions, are equivalent to the ISO(2,1) cycle conditions for the representa- tions of the fundamental group in ISO(2,1). Also, the symplectic structure on the space of invariants is closely related to that of the polygon representation. We choose one of the polygon variables as internal time and compute the Hamiltonian, then perform the Hamilton-Jacobi transformation explicitly. We make contact with other authors' results for g = 1 and g = 2 (N = 0).

gr-qc↗

arXiv1994

Canonical Quantization of (2+1)-Dimensional Gravity

We consider the quantum dynamics of both open and closed two- dimensional universes with ``wormholes'' and particles. The wave function is given as a sum of freely propagating amplitudes, emitted from a network of mapping class images of the initial state. Interference between these amplitudes gives non-trivial scattering effects, formally analogous to the optical diffraction by a multidimensional grating; the ``bright lines'' correspond to the most probable geometries.

gr-qc↗