A global view of Brownian penalisations (Q1029294): Difference between revisions

With the help of Feynman-Kac type penalisation results for Wiener measure \(W\) have been obtained by \textit{B. Roynette} and \textit{M. Yor} [Penalising Brownian paths. Lecture Notes in Mathematics 1969. Berlin: Springer (2009; Zbl 1190.60002)]. The aim of this monograph is to deepen the understanding of \(W\). It consists of four chapters. Chapter 1 ``Existence and main properties of \(W\)''. For the 1-dimensional Brownian motion \((X_t, t \geq 0)\) the construction of \(W\) is given again, so that the monograph may be read independently of previous papers. Among the main properties of \(W\) presented here, let us cite: i) the close links between \(W\) and probabilities obtained by penalising Wiener measure by certain functionals; ii) the existence of integral representation formulae for the measure \(W\); iii) many examples of martingales which converge to 0 as \(t \to \infty\); iv) a general penalisation theorem for Wiener measure; v) the existence of invariant measures for several Markov processes taking values in function spaces. Chapter 2 ``Existence and main properties of \(W^{(2)}\)''. The results relative to the 1-dimensional Brownian motion are extended to 2-dimensional Brownian motion (for which values in \(\mathbb{C}\) instead of \(\mathbb{R}^2\) are considered and complex notation is used). The role of the measure \(W\) is played by a positive and \(\sigma\)-finite measure \(W^{(2)}\). The properties of \(W^{(2)}\) are analogous to those of \(W\). For the \(\mathbb{C}\)-valued Brownian motion \((X_t, t \geq 0)\), the winding process \[ (\theta_t, t \geq 0)=\left(\theta_0+\mathrm{Im }\int_0^t \frac{dX_s}{X_s}, t \geq 0\right) \] is studied under \(W^{(2)}\). A Spitzer type limit theorem is obtained about the asymptotic behavior in distribution for \(\theta_t\) as \(t \to \infty\). Chapter 3 ``The analogue of the measure \(W\) for a class of linear diffusions''. It is devoted to the transcription of several of the preceding results to a certain class of linear diffusions. Fundamental examples of such diffusions are the Bessel processes with dimension \(d = 2(1 - \alpha)\) for \(0 < d < 2\). Chapter 4 ``An analogue of the measure \(W\) for discrete Markov chains''. For Markov chains taking values in a countable set the analogue of the preceding results is obtained. Here some measure is defined which plays the role of the measure \(W\) in the precedings chapters. Many examples are considered, in particular, random walks on trees.