Download Fourier integral operators, quantization, and spectra of by Weinstein A. PDF

By Weinstein A.

Similar mathematics books

Extra resources for Fourier integral operators, quantization, and spectra of Riemannian manifolds (CNRS colloquium)

Example text

Justify the preceding assertion. – Optional and predictable projections can be deﬁned for every nonnegative measurable process X as the limit in n of the corresponding projections of the processes X ∧ n. In this case, projections may take value +∞. 7] are still valid. – Let X, Y and Z be bounded measurable stochastic processes. Moreover, assume that Y is optional and Z is predictable. Prove that Π(OX) = ΠX, O(XY ) = Y (OX), Π(XZ) = Z(ΠX). 8] The following states that the difference between the progressively measurable and the optional σ-algebras is rather small.

If T ⊆ A, then T is a predictable stopping time. 6, T is a stopping time. 5. 6. They are given as exercises. – Let X be a right-continuous predictable stochastic process with values in [0, ∞], whose all trajectories are nondecreasing, a ∈ [0, ∞]. Prove that T := inf {t : Xt a} is a predictable stopping time. We should warn the reader against a possible mistake in a situation which is quite typical in stochastic calculus (though it does not appear in this book). For example, let a = (at ) be a nonnegative progressively measurable (or predictable) stochastic t a ds, where the integral is taken pathwise.

4, we get EXT {T <∞} = Eξ {T = EMT − s} {T = EM∞ s} {T = EZT s} = EE(M∞ |FT − ) {T s} {T <∞} for every predictable stopping time T . Hence, Z = ΠX. 14 at the beginning of the proof, and we should take Y := M [0,s] at the end. – Let X be a bounded measurable process. Show that there exist a sequence {Tn } of stopping times and an evanescent set N such that {OX = ΠX} ⊆ Tn ∪ N. 3 on monotone classes. The following theorem establishes a connection between projections and conditional expectations. – Let X be a bounded measurable process.