Theorem 5.5 -- the lazy walk on the torus mixes in order
ProvedMarkovMixing.torus_mixingmarkov-chainsmixing-timesprobability
The -dimensional discrete torus is the graph whose vertices are the -tuples of residues mod , two vertices being adjacent when they agree in all coordinates but one and differ by there. The lazy random walk on it stays put with probability and otherwise moves to a uniformly chosen neighbour; its stationary distribution is uniform. For a tolerance , the mixing time is the first time at which , where is the total variation distance.
The theorem (Theorem 5.5 of Levin–Peres–Wilmer) asserts: for every dimension there is a constant , depending only on , such that for every side length and every tolerance ,
The walk on the torus mixes in order steps, uniformly in the side length — proved in the book by a coordinatewise coupling.
Preamble
import Definitions.Def_mm_coupling import Mathlib.Analysis.SpecialFunctions.Log.Base
Formal statement
namespace MarkovMixing
/-- **Theorem 5.5** (LPW): the lazy random walk on the `d`-dimensional torus
`ℤ_n^d` satisfies `t_mix(ε) ≤ c(d) n² log₂(ε⁻¹)` for a constant `c(d)`
depending only on the dimension `d`. -/
theorem torus_mixing (d : ℕ) (hd : 0 < d) :
∃ c : ℝ, 0 < c ∧ ∀ (n : ℕ) [NeZero n], 2 ≤ n → ∀ ε : ℝ, 0 < ε → ε ≤ 1 / 2 →
(mixingTime (lazy (graphWalk (torusGraph d n)))
(uniformDist (Fin d → ZMod n)) ε : ℝ) ≤
c * n ^ 2 * Real.logb 2 ε⁻¹ := by
sorry
end MarkovMixing
Source
D. A. Levin, Y. Peres, E. L. Wilmer, Markov Chains and Mixing Times, AMS 2009, https://documents.epfl.ch/groups/i/ip/ipg/www/2013-2014/Random_Walks/markovmixing.pdf, Section 5.3.2, Theorem 5.5, pp. 65-66