Publications

1996

• L'ensemble de Mandelbrot nuageux
  
  • Colonna Jean-François
  , 1996. Cloudy Mandelbrot set (L'ensemble de Mandelbrot nuageux)
• Montagnes et nuages legers
  
  • Colonna Jean-François
  , 1996. Mountains and light clouds (Montagnes et nuages legers)
• Montagnes et nuages legers
  
  • Colonna Jean-François
  , 1996. Mountains and light clouds (Montagnes et nuages legers)
• Analysis of Sound Signals with High Resolution Matching Pursuit
  
  • Gribonval Rémi
  • Bacry Emmanuel
  • Mallat Stéphane
  • Depalle Philippe
  • Rodet Xavier
  , 1996, pp.125-128. Sound recordings include transients and sustained parts. Their analysis with a basis expansion is not rich enough to represent efficiently all such components. Pursuit algorithms choose the decomposition vectors depending upon the signal properties. The dictionary among which these vectors are selected is much larger than a basis. Matching pursuit is fast to compute, but can provide coarse representations. Basis pursuit gives a better representation but is very expensive in terms of calculation time. This paper develops a high resolution matching pursuit: it is a fast, high time-resolution, time-frequency analysis algorithm, that makes it likely to be used far musical applications (10.1109/TFSA.1996.546702)
  DOI : 10.1109/TFSA.1996.546702
• Montagnes et de nuages légers (vue aérienne)
  
  • Colonna Jean-François
  , 1996. Mountains and light clouds (bird's-eye view) (Montagnes et de nuages légers (vue aérienne))
• Vue artistique d'un ensemble de Julia dans le corps des quaternions
  
  • Colonna Jean-François
  , 1996. Artistic view of a quaternionic Julia set (Vue artistique d'un ensemble de Julia dans le corps des quaternions)
• Intégration de 128 coupes d'un ensemble de Julia dans le corps des quaternions le long de son quatrième axe
  
  • Colonna Jean-François
  , 1996. Integration of 128 tridimensional cross-sections of a quaternionic Julia set computed along its fourth axis (Intégration de 128 coupes d'un ensemble de Julia dans le corps des quaternions le long de son quatrième axe)
• Du Soleil à Pluton (échelles non linéaires)
  
  • Colonna Jean-François
  , 1996. From the Sun to Pluto (non linear scales) (Du Soleil à Pluton (échelles non linéaires))
• Evolutionary Algorithms for Constrained Parameter Optimization Problems
  
  • Michalewicz Zbigniew
  • Schoenauer Marc
  Evolutionary Computation, Massachusetts Institute of Technology Press (MIT Press), 1996, 4 (1), pp.1-32. Evolutionary computation techniques have received a lot of attention regarding their potential as optimization techniques for complex numerical functions. However, they have not produced a significant breakthrough in the area of nonlinear programming due to the fact that they have not addressed the issue of constraints in a systematic way. Only recently several methods have been proposed for handling nonlinear constraints by evolutionary algorithms for numerical optimization problems; however, these methods have several drawbacks and the experimental results on many test cases have been disappointing. In this paper we (1) discuss difficulties connected with solving the general nonlinear programming problem, (2) survey several approaches which have emerged in the evolutionary computation community, and (3) provide a set of eleven interesting test cases, which may serve as a handy reference for future methods.
• Zoom sur une représentation bidimensionnelle de l'ensemble de Mandelbrot, avec visualisation des arguments
  
  • Colonna Jean-François
  , 1996. Bidimensional zoom in on the Mandelbrot set with display of the arguments (Zoom sur une représentation bidimensionnelle de l'ensemble de Mandelbrot, avec visualisation des arguments)
• Marche aléatoire isotrope de 64 particules sur un réseau cubique tridimensionnel
  
  • Colonna Jean-François
  , 1996. Isotropic random walk of 64 particles on a tridimensional square lattice (Marche aleatoire isotrope de 64 particules sur un reseau cubique tridimensionnel)
• Vue artistique du système solaire
  
  • Colonna Jean-François
  , 1996. Artistic view of the Solar System (Vue artistique du systeme solaire)
• Integration du probleme des N-corps (N=10) montrant le veritable systeme solaire
  
  • Colonna Jean-François
  , 1996. N-body problem integration (N=10) displaying the actual Solar System (Integration du probleme des N-corps (N=10) montrant le veritable systeme solaire)
• Integration du probleme des N-corps (N=10) montrant le veritable systeme solaire, notre Terre etant choisie comme origine des coordonnees
  
  • Colonna Jean-François
  , 1996. N-body problem integration (N=10) displaying the actual Solar System, our Earth being at the origin of the coordinates (Integration du probleme des N-corps (N=10) montrant le veritable systeme solaire, notre Terre etant choisie comme origine des coordonnees)
• Genetic Algorithms for Air Traffic Control System
  
  • Delahaye Daniel
  • Alliot Jean-Marc
  • Durand Nicolas
  • Schoenauer Marc
  , 1996. no abstract
• An induction-based control for genetic algorithms
  
  • Sebag Michèle
  • Schoenauer Marc
  • Ravisé Caroline
  , 1996, 1063, pp.100-119. This paper presents a Machine Learning approach to control genetic algorithms. From examples gathered through spying evolution or experimenting on populations, induction extracts a rule-based characterization of which evolutionary events are good or bad for evolution. Such rule base allows for further generations to escape most disruptive or unproductive changes, according to a civilized rather than Darwinian evolution scheme. An evolutionary event is described as mutating a chromosome (at given bit—string positions) or crossing over two chromosomes (with given crossing points), and labeled by comparing the fitness of the offspring with that of its parents. Knowledge induced from such events allows to predict the effects of further operators, thereby filtering further undesirable events. Experiments on some artificial problems are discussed. (10.1007/3-540-59286-5_85)
  DOI : 10.1007/3-540-59286-5_85
• Topological Optimum Design using Genetic Algorithms
  
  • Kane Couro
  • Schoenauer Marc
  Control and Cybernetics, Polish Academy of Sciences, 1996, 25 (5), pp.1059-1088. Structural topology optimization is addressed through Genetic Algorithms: A set of designs is evolved following the Darwinian survival-of-fittest principle. The goal is to optimize the weight of the structure under displacement constraints. This approach demonstrates high flexibility, and breaks many limits of standard optimization algorithms, in spite of the heavy requirements in term of computational effort: Alternate optimal solutions to the same problem can be found; Structures can be optimized with respect to multiple loadings; The prescribed loadings can be applied on the unknown boundary of the solution, rather than on the fixed boundary of the design domain; Different materials as well as different mechanical models can be used, as witnessed by the first results of Topological Optimum Design ever obtained in the large displacements model. But these results could not have been obtained without careful specific handling of the specific aspects of topological genetic optimization: First, specific genetic operators (crossover, mutation) were introduced; Second, special attention was paid to the design of the objective function; The nonlinear geometrical effects of the large displacement model lead to non viable solutions, unless some constraints are imposed on the stress field.
• Evolutionary identification of rheological models
  
  • Sebag Michèle
  • Schoenauer Marc
  • Maitournam Habibou
  , 1996, pp.465-470.
• Contrôle d'un algorithme génétique
  
  • Sebag Michèle
  • Schoenauer Marc
  Revue des Sciences et Technologies de l'Information - Série RIA : Revue d'Intelligence Artificielle, Lavoisier, 1996, 10, pp.389-428. Après avoir rappelé le principe des algorithmes génétiques (AG), cet article s'inté­resse à l'efficacité de ces méthodes d'optimisation. Les liens entre la représentation choisie et les opérateurs d'évolution manipulant cette représentation sont discutés. Deux problèmes artificiels illustrant des limites attendues ou inattendues des AG sont présentés. Enfin, le nombre important de paramètres et de choix structurels intervenant dans un algorithme génétique pose le problème du contrôle d'un AG. Cet article propose une méthode nouvelle pour le contrôle d'un AG, fondée sur l'apprentissage symbolique.
• Mechanical Inclusions Identification by Evolutionary Computation
  
  • Schoenauer Marc
  • Kallel Leila
  • Jouve François
  Revue Européenne des Éléments Finis, HERMÈS / LAVOISIER, 1996, 5 (5-6), pp.619-648. The problem of the identification of mechanical inclusion is theoretically ill-posed, and to-date numerical algorithms have demonstrated to be inaccurate and unstable. On the other hand, Evolutionary Algorithms provide a general approach to inverse problem solving. However, great care must be taken during the implementation: The choice of the representation, which determines the search space, is critical. Three representations are presented and discussed. Whereas the straightforward mesh-dependent representation suffers strong limitations, both mesh-independent representation provide outstanding results on simple instances of the identification problem, including experimental robustness in presence of noise.
• Wavelet based fractal analysis of DNA sequences
  
  • Arneodo A.
  • Daubentoncarafa Y.
  • Bacry Emmanuel
  • Graves P. V.
  • Muzy J. F.
  • Thermes C.
  Physica D: Nonlinear Phenomena, Elsevier, 1996, 96 (1-4), pp.291-320. no abstract (10.1016/0167-2789(96)00029-2)
  DOI : 10.1016/0167-2789(96)00029-2
• Evolutionary identification of macro-mechanical models
  
  • Schoenauer Marc
  • Sebag Michèle
  • Jouve François
  • Lamy Bertrand
  • Maitournam Habibou
  , 1996, pp.467-488. This chapter illustrates the potential of genetic programming (GP) in the field of macro­ mechanical modeling, addressing the problem of identification of a mechanical model for a material. Two kinds of models are considered. One-dimensional dynamic models are rep­resented via symbolic formulations termed rheological models, which are directly evolved by GP. Three-dimensional static models of hyperelastic materials are expressed in terms of strain energy functions. A model is rated based on the distance between the behavior predicted by the model, and the actual behavior of the material given by a set of me­ chanical experiments. The choice of GP is motivated by strong arguments, relying on the tree-structure of rheological models in the first case, and on the need for first and second order derivatives in the second case. Key issues are the exploration of viable individuals only, and the use of Gaussian mutations to optimize numerical constants.