| Contents |
| We describe an algorithm which is able to approximate invariant measures of dynamical systems up to small errors in the Wasserstein distance, and its practical implementation. The use of Wasserstein distance, allows to replace some difficult a priori estimations on the regularity of the invariant measure and exploit as much as possible some a posteriori estimates which are made by the computer. We will show a variation of the algorithm for computing the invariant measure of general homeomorphisms of the circle with irrational rotation number, where we can rigorously compute the invariant measure up to small errors. Such maps form a family of examples which are not mixing but just ergodic, and for which we are able to compute the invariant measure. |
|