Abstract
In the late 1970s, Feigenbaum showed how to apply the ideas of the renormalization group to the problem of transition to chaos by period doubling. Starting with Feigenbaum's proof of universality, Jean-Pierre Eckmann's seminar attempted to explain its multiple ramifications : generalization to other universality classes, and also generalization to an arbitrary number of dimensions, which provided a link to experimental results. The seminar ended with an analysis of the Hamiltonian case, for which the universality constant is almost double that of the original one-dimensional case.
