Accelerating diffusions
Hwang, Chii-Ruey ; Hwang-Ma, Shu-Yin ; Sheu, Shuenn-Jyi
Ann. Appl. Probab., Tome 15 (2005) no. 1A, p. 1433-1444 / Harvested from Project Euclid
Let U be a given function defined on ℝd and π(x) be a density function proportional to exp−U(x). The following diffusion X(t) is often used to sample from π(x), ¶ \[dX(t)=-\nabla U(X(t))\,dt+\sqrt{2}\,dW(t),\qquad X(0)=x_{0}.\] ¶ To accelerate the convergence, a family of diffusions with π(x) as their common equilibrium is considered, ¶ \[dX(t)=\bigl(-\nabla U(X(t))+C(X(t))\bigr)\,dt+\sqrt{2}\,dW(t),\qquad X(0)=x_{0}.\] ¶ Let LC be the corresponding infinitesimal generator. The spectral gap of LC in L2(π) (λ(C)), and the convergence exponent of X(t) to π in variational norm (ρ(C)), are used to describe the convergence rate, where ¶ λ(C)=Sup {real part of μ:μ is in the spectrum of LC, μ is not zero}, ¶ \[\rho(C)=\operatorname {Inf}\biggl\{\rho\dvtx \int \vert p(t,x,y)-\pi(y)\vert \,dy\le g(x)e^{\rho t}\biggr\}.\] ¶ Roughly speaking, LC is a perturbation of the self-adjoint L0 by an antisymmetric operator C⋅∇, where C is weighted divergence free. We prove that λ(C)≤λ(0) and equality holds only in some rare situations. Furthermore, ρ(C)≤λ(C) and equality holds for C=0. In other words, adding an extra drift, C(x), accelerates convergence. Related problems are also discussed.
Publié le : 2005-05-14
Classification:  Diffusion,  convergence rate,  acceleration,  spectral gap,  spectrum,  variational norm,  ergodicity,  MCMC,  Monte Carlo Markov process,  60J60,  47D07,  65B99,  35P05
@article{1115137980,
     author = {Hwang, Chii-Ruey and Hwang-Ma, Shu-Yin and Sheu, Shuenn-Jyi},
     title = {Accelerating diffusions},
     journal = {Ann. Appl. Probab.},
     volume = {15},
     number = {1A},
     year = {2005},
     pages = { 1433-1444},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1115137980}
}
Hwang, Chii-Ruey; Hwang-Ma, Shu-Yin; Sheu, Shuenn-Jyi. Accelerating diffusions. Ann. Appl. Probab., Tome 15 (2005) no. 1A, pp.  1433-1444. http://gdmltest.u-ga.fr/item/1115137980/