A new algorithm for training feedforward multilayer neural networks is proposed. It is based on recursive least squares procedures and U-D factorization, which is a well-known technique in filter theory. It will be shown that due to the U-D factorization method, our algorithm requires fewer computations than the classical RLS applied to feedforward multilayer neural network training.
Evolutionary Algorithms, also known as Genetic Algorithms in a former terminology, are probabilistic algorithms for optimization, which mimic operators from natural selection and genetics. The paper analyses the convergence of the heuristic associated to a special type of Genetic Algorithm, namely the Steady State Genetic Algorithm (SSGA), considered as a discrete-time dynamical system non-generational model. Inspired by the Markov chain results in finite Evolutionary Algorithms, conditions are...
Simple Temporal Networks (STN) allow conjunctions of minimum and maximum distance constraints between pairs of temporal positions to be represented. This paper introduces an extension of STN called Time–dependent STN (TSTN), which covers temporal constraints for which the minimum and maximum distances required between two temporal positions x and y are not necessarily constant but may depend on the assignments of x and y. Such constraints are useful to model problems in which the duration of an...
A number of methodological papers published during the last years testify that a need for a thorough revision of the research methodology is felt by the operations research community – see, for example, [Barr et al., J. Heuristics1 (1995) 9–32; Eiben and Jelasity, Proceedings of the 2002 Congress on Evolutionary Computation (CEC'2002) 582–587; Hooker, J. Heuristics1 (1995) 33–42; Rardin and Uzsoy, J. Heuristics7 (2001) 261–304]. In particular, the performance evaluation of nondeterministic methods,...
L'apprentissage automatique à partir d'exemples consiste généralement à caractériser un ensemble d'objets dénotant un concept. Nous avons développé deux méthodes d'apprentissage symbolique, LEGAL et LEGAL-E, qui s'appuient sur le même modèle d'apprentissage, et utilisent une technique de généralisation descendante, basée sur la logique des propositions et sur la structure de treillis de Galois, pour produire un ensemble de descriptions structurées et ordonnées. Elles diffèrent dans leur approche...
Most of the techniques used in text classification are determined by the occurrences of the words (terms) appearing in the documents, combined with the user feedback over the documents retrieved. However, in our model, the most relevant terms will be selected from a previous fuzzy classification given by the genetic algorithm guided by the user feedback, but using techniques from Machine Learning. A feature selection process is carried out through a Genetic Algorithm in order to find the most discriminatory...