Mala

Metropolis-adjusted Langevin algorithm

Mala算法/Langevin Monte Carlo (LMC)v by wikipedia

Algorithm
1.Initialization：set $X_0=x_0$ (and recursively define an app $X_k$ to the true solution $X(k\tau )$)
2. (proposal dis)the transition probability density

Note that $X_{k+1}$ is normally distributed with mean $X_k+\tau \nabla \log \pi (X_k)$ and covariance equal to $2\tau$ times the $d\times d$ identity matrix.
3. This proposal is accepted or rejected according to the Metropolis–Hastings algorithm:

注意：

1. Compared to naive Metropolis–Hastings, MALA has the advantage that it usually proposes moves into regions of higher $\pi$ probability, which are then more likely to be accepted.
2. On the other hand, when $\pi$ is strongly anisotropic (i.e. it varies much more quickly in some directions than others), it is necessary to take $0<\tau \ll 1$ in order to properly capture the Langevin dynamics;
   the use of a positive-definite preconditioning matrix $A\in {\mathbb{R}}^{d\times d}$ can help to alleviate this problem, by generating proposals according to
   ${\tilde{X}}_{k+1}:=X_k+\tau A\nabla \log \pi (X_k)+\sqrt{2\tau A}{\xi }_k$

操作guide:
In practical applications, the optimal acceptance rate for this algorithm is $0.574$.