IBP-Laforia
1996/32:
Rapport de Recherche Laforia /
Laforia research reports
37 pages - Mars/March 1997 -
French document.
PostScript : Ko /Kb
Titre / Title: - S.A.G.A.C.E. : Solution Algorithmique Génétique pour l'Anticipation de Comportements Evolutifs
Abstract : The intelligence of an individual is commonly evaluated according to some particular abilities : Learning, Adaptation, Anticipation. An artificial intelligent system ought to be judged on the same criteria. Most of the 'intelligent' realizations on computers are concerned with Learning (problem solving, games, etc.) or Adaptation (Adaptive systems, robots, etc.). Only a few of these systems deal with Anticipation and when they do, they rarely take into account its whole complexity (problematic, influence of Anticipation on its object, countermeasures, etc.).
This report is intended to give an in-depth look at the different aspects of Anticipation and to present the S.A.G.A.C.E. method (Genetic Algorithmic Solution for the Anticipation of evolutive Behaviors). This method that we have developped is able to predict and to anticipate, with good accuracy, evolutive (indeed adaptive) observed behaviors.
Keywords : Anticipation / Modelling / Learning / Genetic algorithms / Incomplete information games / S.A.G.A.C.E.
Publications internes Laforia 1996 / Laforia research reports 1996