Processing math: 0%
Accelerating Gaussian Diffusions
Hwang, Chii-Ruey ; Hwang-Ma, Shu-Yin ; Sheu, Shuenn-Jyi
Ann. Appl. Probab., Tome 3 (1993) no. 4, p. 897-913 / Harvested from Project Euclid
Let \pi(x) be a given probability density proportional to \exp(-U(x)) in a high-dimensional Euclidean space \mathbb{R}^m. The diffusion dX(t) = -\nabla U(X(t))dt + \sqrt 2 dW(t) is often used to sample from \pi. Instead of -\nabla U(x), we consider diffusions with smooth drift b(x) and having equilibrium \pi(x). First we study some general properties and then concentrate on the Gaussian case, namely, -\nabla U(x) = Dx with a strictly negative-definite real matrix D and b(x) = Bx with a stability matrix B; that is, the real parts of the eigenvalues of B are strictly negative. Using the rate of convergence of the covariance of X(t) [or together with EX(t)] as the criterion, we prove that, among all such b(x), the drift Dx is the worst choice and that improvement can be made if and only if the eigenvalues of D are not identical. In fact, the convergence rate of the covariance is \exp(2\lambda_M(B)t), where \lambda_M(B) is the maximum of the real parts of the eigenvalues of B and the infimum of \lambda_M(B) over all such B is 1/m \operatorname{tr} D. If, for example, a "circulant" drift \bigg(\frac{\partial U}{\partial x_m} - \frac{\partial U}{\partial x_2},\frac{\partial U}{\partial x_1} - \frac{\partial U}{\partial x_3}, \cdots, \frac{\partial U}{\partial x_{m-1}} - \frac{\partial U}{\partial x_1}\bigg) is added to Dx, then for essentially all D, the diffusion with this modified drift has a better convergence rate.
Publié le : 1993-08-14
Classification:  Diffusions,  convergence rate,  stochastic relaxation,  Monte Carlo method,  stability matrix,  Ornstein-Uhlenbeck process,  reversible process,  image analysis,  covariance matrix,  60J60,  62E25,  65C05,  82B31,  68U10
@article{1177005371,
     author = {Hwang, Chii-Ruey and Hwang-Ma, Shu-Yin and Sheu, Shuenn-Jyi},
     title = {Accelerating Gaussian Diffusions},
     journal = {Ann. Appl. Probab.},
     volume = {3},
     number = {4},
     year = {1993},
     pages = { 897-913},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1177005371}
}
Hwang, Chii-Ruey; Hwang-Ma, Shu-Yin; Sheu, Shuenn-Jyi. Accelerating Gaussian Diffusions. Ann. Appl. Probab., Tome 3 (1993) no. 4, pp.  897-913. http://gdmltest.u-ga.fr/item/1177005371/