Weak mixing and sparse equidistribution

Max Auer
2024-08-28
Abstract:The celebrated Birkhoff Ergodic Theorem asserts that, for an ergodic map, orbits of almost every point equidistributes when sampled at integer times. This result was generalized by Bourgain to many natural sparse subsets of the integers. On the other hand, the behaviour of orbits of \textbf{all} points in a dynamical system is much less understood, especially for sparse subsets of the integers. We generalize a method introduced by A. Venkatesh to tackle this problem in two directions, general $\mathbb{R}^d$ actions instead of flows, and weak mixing, rather than mixing, actions. Along the way, we also establish some basic properties of weak mixing and show weak mixing for the time 1-map of a weak mixing flow. The main examples are equidistribution along $\{(n^{1+\epsilon_i})_{i=1}^d\}$, which is a direct generalisation of Venkatesh's example, and equidistribution along random subsets of $\mathbb{Z}^d$ for horosperical action.
Dynamical Systems
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the equidistribution behavior of the orbits of all points on sparse subsets in dynamical systems. Specifically, the paper focuses on how to generalize the method proposed by A. Venkatesh to deal with problems in two directions: one is to generalize from flows to more general \( \mathbb{R}^d \) actions; the other is to generalize from mixing actions to weakly mixing actions. ### Background and Problem Description of the Paper 1. **Birkhoff Ergodic Theorem**: - The Birkhoff ergodic theorem states that if \( \phi \) is a measure - preserving ergodic mapping, then for almost every point \( x \), its orbit is almost everywhere equidistributed under integer - time sampling. - The mathematical expression is: \[ \frac{1}{N} \sum_{n = 0}^{N - 1}f(\phi^n(x))\to\int_X f\,d\mu\quad\text{when}\quad N\to\infty \] 2. **Ergodicity of Sparse Subsets**: - A natural question is whether the above results can be generalized to different sequences, such as sparse subsets. - Specifically, for a given sequence \( (b_n)_{n\geq0}\subset\mathbb{N} \), consider the following expression: \[ \frac{1}{N} \sum_{n = 0}^{N - 1}f(\phi^{b_n}(x)) \] - Does this expression converge? This is a core problem of the paper. ### Main Contributions of the Paper 1. **Generalization of the Method**: - The author generalizes A. Venkatesh's method and applies it to more general \( \mathbb{R}^d \) actions and weakly mixing actions. - Through this method, the author is able to study the equidistribution properties on sparse subsets. 2. **The Role of Weak Mixing**: - Weak mixing is a weaker concept than mixing, but it is more common in many practical problems. - The author proves that the time - 1 mapping under weak mixing actions is also weakly mixing, and establishes some basic properties of weak mixing. 3. **Specific Examples**: - The paper gives two main examples: - The equidistribution along \( \{(n_1+\epsilon_i)^d\}_{i = 1} \), which is a direct generalization of Venkatesh's result. - The equidistribution along a random subset of \( \mathbb{Z}^d \), especially for the horospherical action. ### Conclusion The paper solves the problem of equidistribution behavior on sparse subsets in dynamical systems by generalizing existing methods. These results are not only of great theoretical significance, but also provide new tools and perspectives for further research.