English

Approximately Jumping Towards the Origin

Probability 2024-12-06 v1 Dynamical Systems

Abstract

Given an initial point x0Rdx_0 \in \mathbb{R}^d and a sequence of vectors v1,v2,v_1, v_2, \dots in Rd\mathbb{R}^d, we define a greedy sequence by setting xn=xn1±vnx_{n} = x_{n-1} \pm v_n where the sign is chosen so as to minimize xn\|x_n\|. We prove that if the vectors viv_i are chosen uniformly at random from Sd1\mathbb{S}^{d-1} then elements of the sequence are, on average, approximately at distance xnπd/8\|x_n\| \sim \sqrt{\pi d/8} from the origin. We show that the sequence (xn)n=1(\|x_n\|)_{n=1}^{\infty} has an invariant measure πd\pi_d depending only on dd and we determine its mean and study its decay for all dd. We also investigate a completely deterministic example in d=2d=2 where the vnv_n are derived from the van der Corput sequence. Several additional examples are considered.

Cite

@article{arxiv.2412.04284,
  title  = {Approximately Jumping Towards the Origin},
  author = {Alex Albors and François Clément and Shosuke Kiami and Braeden Sodt and Ding Yifan and Tony Zeng},
  journal= {arXiv preprint arXiv:2412.04284},
  year   = {2024}
}
R2 v1 2026-06-28T20:24:24.821Z