Séminaire de probabilités. V. Université de Strasbourg (Q2546296)

As with previous volumes in this series the book under review is a large and varied collection of articles from the University of Strasbourg probability seminar. These include original contributions, new proofs of old results, and commentaries upon or alternative expositions of old or recent work (some unpublished) -- all from the wide field of probability, topological measure theory and potentials, and always presented with the elegance and precision that we have come to expect from \textit{P. A. Meyer} and his school. We turn to a brief notice of each of the articles. \textit{Ph. Artzner} [Fonctions caractéristiques et mesures planes invariantes par rotations (1--16)] begins with an interesting discussion of measures in the plane invariant under rotations centred at the origin, their associated measures on \(\mathbb R^+\) and characteristic functions. (It is surprising that no reference is made to the paper of \textit{J. F. C. Kingman}, Acta Math. 109, 11--53 (1963; Zbl 0121.12803)], the case \(n=2\) is surely related]. This is followed by an exposition by \textit{Patrice Assouad} [Démonstration de la ``Conjecture de Chung'' par Carleson (17--20)] of L. Carleson's (unpublished) solution of a conjecture of K. L. Chung which plays a part in some recent work on processes with independent increments. Next \textit{Jean Bretagnolle} [Résultats de Kesten sur les processus à accroissements indépendants (21--36)] discusses some aspects of the work of \textit{H. Kesten} [Mem. Am. Math. Soc. 93, 129 p. (1969; Zbl 0186.50202)] and some unpublished extensions (to separable l.c.a. groups) due to S. C. Port and C. Stone. \textit{R. Cairoli} [Décomposition de processus a indices doubles (37--57)] obtains (and applies) decomposition theorems for double indexed processes analogous to ones of P. A. Meyer. And then we come to an extremely elegant introduction by \textit{P. Cartier} [Introduction à l'étude des mouvements browniens à plusieurs paramètres (58--75)] to P. Levy's theory of Brownian motion with several parameters. This includes a discussion of the general theory of Gaussian random functions culminating in a construction of the object of interest -- a Gaussian random function \(X(t)\) indexed by a real Hilbert space \(T\) with (i) \(X(0)=0\), (ii) \(E[(X(t) - X(t'))^2] = \Vert t - t'\Vert\) and (iii) \(C_X(t,t') = \tfrac12 \{\Vert t\Vert + \Vert t'\Vert - \Vert t - t'\Vert\}\). A simple proof by \textit{K. L. Chung} [A simple proof of Doob's convergence theorem (76)] of a convergence theorem in potential theory of Doob precedes a series of six papers by \textit{C. Dellacherie} [Quelques commentaires sur les prolongements de capacités (77--81); Une démonstration du théorème de séparation des ensembles analytiques (82--85); Correction à ``ensembles aléatoires. II'' (86); Les théorèmes de Mazurkiewicz-Sierpinski et de Lusin (87--102); Espaces paves et rabotages (103--126)]. These papers are excellent expositions of certain topics in the theory of capacities and associated topology and measure theory. They include a. derivation of the separation theorem for analytic sets using the fundamental result concerning capacitability, thus reversing the method Sion used to establish the latter result; a detailed discussion of the previously not very accessible theorems of Mazurkiewicz-Sierpinski and Lusin used by the author in recent work; and a systematic study (in fairly abstract form) of paved sets and ``rabotages globals''. In the next two papers \textit{C. Dellacherie} and \textit{C. Doléans-Dade} [Un contre-exemple au problème des laplaciens approches (127--137)] and the second author alone [Une martingale uniformément intégrable mais non localement de carré intégrable (138--140)] give examples of processes with certain specific properties, the latter example being of a uniformly integrable martingale \((X_t)\) such that \(E[X_T^2] =+\infty\) for all stopping times \(T\not\equiv 0\). Then Mme. Doléans-Dade [Intégrales stochastiques par rapport à une famille de probabilités (141--146)] poses and solves a problem involving stochastic integrals with respect to a family \((P^\mu)_{\mu\in M}\) of pr. measures. A detailed discussion of some previously announced work on regenerative sets by \textit{B. Maisonneuve} [Ensembles régénératifs, temps locaux et subordinateurs (147--169)] brings us to P. A. Meyer's extensive contribution to this book. In a sequence of ten articles \textit{P. A. Meyer} covers an extraordinary range of topics -- [Sur un article de Dubins (170--176)] the exposition of a method of Dubins for realising a discrete time martingale as a stopped Brownian motion, [Processus de Poisson ponctuels, d'après K. Ito (177--190)] a general theory of Poisson point processes (and applications) after Ito, [Démonstration simplifiée d'un théorème de Knight (191--195)] a simplified proof of a theorem of Knight characterising \(n\)-dimensional Brownian motion, [Représentation intégrale des fonctions excessives résultats de Mokobodzki (196--208)] the exposition of results of Mokobodzki on the integral representation of excessive functions, [Le retournement du temps, d'après Chung et Walsh (213--236)] time reversal in Markov processes after Chung and Walsh and a number of other contributions and discussions [Un théorème sur la répartition des temps locaux (209--210); Deux petits résultats de théorie du potentiel (211--212); Travaux de H. Rost en théorie du balayage (237--250); Solutions de l'équation de Poisson dans le cas récurrent (251--269); Balayage pour les processus de Markov continus à droite, d'après Shih Chung Tuo (270--274)]. \textit{John B. Walsh} [Two footnotes to a theorem of Ray (283--289)] gives some remarks on two theorems of Ray followed by an interesting discussion [Some topologies connected with Lebesgue measure (290--310)] (with applications to potential theory and Markov chains) of some topologies of \(\mathbb R\) which are finer than the usual and have properties tantamount to ignoring sets of Lebesgue measure zero. \textit{Takesi Watanabe} [On balayées of excessive measures and functions with respect to resolvents (311--341)] extends the theory of Deny of balayage with respect to excessive measures and functions of a single kernel, replacing the single kernel by a resolvent. Three papers by \textit{Michel Weil} [Décomposition d'un temps terminal (342--346); Quasi-processus et énergie (347--361); Conditionnement par rapport au passé strict (362--372)], two relating to stopping times and one about quasi processes and energy conclude this excellent collection of articles. The only papers we have not so far noted are one by \textit{Daniel Revuz} [Remarque sur les potentiels de mesure (275--277)] consisting of remarks on potentials of measures and \textit{J. de Sam Lazaro} and \textit{P. A. Meyer} [Une remarque sur le flot du movement brownien (278--282)] on aspects of Brownian motion.