The number of two-dimensional maxima (Q4330058)

Let \(n\) points be placed uniformly at random in a subset \(A\) of the plane. A point is said to be maximal in the configuration if no other point is larger in both coordinates. It is shown that for large \(n\) and for many sets \(A\), the number \(W\) of maximal points is approximately normally distributed. The argument, which uses Stein's method, is applicable in higher dimensions, although the geometry becomes more complicated there. In the case where \(A\) is convex, the results complement those of \textit{L. Devroye} [Comput. Math. Appl. 25, No. 5, 19-31 (1993; Zbl 0769.60044)]. In the case that \(A\) is the unit square, \(W\) has the same distribution as the number of record values in a sequence of \(n\) independent observations from the same continuous distribution, first analyzed by A.~Rényi. In the case that \(A\) is a triangle and \(n\) is a Poisson variable, \(W\) is asymptotically equivalent to the number of germs in a Johnson-Mehl process; see for example \textit{S. N. Chiu }and \textit{M. P. Quine} [Ann. Appl. Probab. 7, No. 3, 802-814 (1997; Zbl 0888.60016)].