1: \begin{abstract}

2: \noindent

3: 	%

4: We analyse the convergence to equilibrium of the \textit{Bernoulli--Laplace urn model}:

5: \begin{itemize}[noitemsep, topsep=\smallskipamount]

6: 	\item 

7: 	initially,

8: 	one urn contains $k$ red balls and a second $n-k$ blue balls;

9: 	

10: 	\item 

11: 	in each step,

12: 	a pair of balls is chosen uniform and their locations are switched.

13: \end{itemize}

14: Cutoff is known to occur at $\tfrac12 n \log \min\set{k, \sqrt n}$ with window order $n$ whenever $1 \ll k \le \tfrac12 n$.

15: %	the process is far from equilibrium after $\tfrac12 n (\log \min\set{k, \sqrt n} - C)$ swaps ($C$ large),

16: %	but close after $\tfrac12 n (\log \min\set{k, \sqrt n} + C)$.

17: We refine this by determining the \textit{limit profile}:

18: a function $\Phi$

19: such that

20: \begin{center-small}

21: \(\displaystyle

22: 	d_\TV\rbb{ \tfrac12 n \log \min\set{k, \sqrt n} + \theta n }

23: \to

24: 	\Phi(\theta)

25: \Qas

26: 	n \to \infty

27: \Qforall

28: 	\theta \in \mbr.

29: \)

30: \end{center-small}\noindent

31: Our main technical contribution, of independent interest, approximates a rescaled chain by a diffusion on $\mbr$ when $k \gg \sqrt n$, and uses its explicit law as a Gaussian process.

32: 	%

33: \end{abstract}

34: