an:01750589
Zbl 1022.81042
Faris, William G.
Ornstein-Uhlenbeck and renormalization semigroups
EN
Mosc. Math. J. 1, No. 3, 389-405 (2001).
00080461
2001
j
81T17 82B28 47D06 60J60
renormalization group; Ornstein-Uhlenbeck semigroup; measures on Hilbert space
An abstract characterization of the Ornstein-Uhlenbeck semigroup corresponding to the renormalization group realized as a combination of a scaling with a Gaussian convolution is given. Precisely, let \(H= L^2(\mathbb{R}^n, d^nx)\) be the Hilbert space, \(n>2\) and \(D= x\cdot\nabla+\nu/2\), \(-A= D-1\). Let \(\exp(-tL)\) be the Uhlenbeck-Ornstein semigroup corresponding to the renormalization group, with flow \(\exp(-tA)\) generated by \(-A\). Let \(\nu\) be the invariant Gaussian measure with covariance \(G= C+aS\), \(C= (-\Delta)^{-1}(1- \exp(\alpha\Delta))\), \(S= (-\Delta)^{-1}\), \(a>0\), where \(\alpha\) determines the cutoff. Then it is proved \(G= RS\), where \(R\), \(S\) commute and \(R\) is selfadjoint and invertible. Furthermore, the Hilbert space adjoint semigroup \(\exp(-tL^*)\) is the Ornstein-Uhlenbeck semigroup with flow \(G\exp(-tA^*)G^{-1}= R\exp(tA)R^{-1}\) generated by \(-GA^* G^{-1}= RAR^{-1}\) (Theorem 7.1).
To prove the theorem, first measures on Hilbert space is reviewed following \textit{N. N. Vakhaniya}, \textit{V. I. Tarieladze} and \textit{S. A. Chobanyan} [Probability distributions on Banach spaces, Dordrecht (1987; Zbl 0698.60003)], in Sect. 2. Ornstein-Uhlenbeck semigroup (Mehler semigroup) is reviewed in Sect. 3 following \textit{V. I. Bogachev}, \textit{M. R??ckner} and \textit{B. Schmuland}, [Generalized Mehler semigroups and applications, Probab. Theory Related Fields 105, 193-225 (1996; Zbl 0849.60066)]. Sect. 4 and 5 are devoted to the study of the generator of the Ornstein-Uhlenbeck semigroup and the adjoint semigroup. Then abstract renormalization group is treated in Sect. 6. After these preparations, Theorem 7.1 is proved. The RG example suggests the program of looking for other pairs consisting of a diffusion given by \(Q\) and a linear vector field. One such example taking the covariance \(Q\) to be an integral operator was given by \textit{G. Da Prato} and \textit{J. Zabczyk} [Stochastic equations in infinite dimensions, Encyclopedia of Mathematics and its Applications, 44. Cambridge (1992; Zbl 0761.60052)]. This example and its generalization are discussed in Sect. 8 and 9.
Akira Asada (Takarazuka)
Zbl 0698.60003; Zbl 0849.60066; Zbl 0761.60052