No label defined (Q4952337)

The present survey paper deals with the relationship between Hausdorff dimension of certain exceptional subsets of the Brownian path and critical exponents for probabilities of events related to Brownian motion. The author first introduces a general framework and then discusses several examples, including cone points, cut points, frontier points and pioneer points. As an example, cut points are those points \(B(t)\) on a planar Brownian path, which separate the path in two disjoint parts \(B[0,t)\) and \(B(t,1]\). The Hausdorff dimension of the set of cut points is shown to be \(2-\xi_2(1,1)\), where \(\xi_2(1,1)\) is an intersection exponent defined as follows: The probability that the paths of two independent Brownian motions started at distinct points on the unit sphere does not intersect before the motions leave the ball of radius \(n\) is logarithmically asymptotically equal to \(-n\xi_2(1,1)\). At the time when the survey was written only conjectures for the exact value of intersection exponents were known. Meanwhile however, Lawler, Werner and Schramm, in a series of deep and exciting papers, have been able to prove several of these conjectures, including \(\xi_2(1,1)=5/4\). These new developments make the present paper even more relevant. In the last two sections results for Brownian motion are carried over to discrete random walks. Altogether, this is a very recommendable survey paper on a subject of great current interest. For the entire collection see [Zbl 0933.00038].