Computing half-plane and strip discrepancy of planar point sets

We present efficient algorithms for two problems concerning the discrepancy of a set S of n points in the unit square in the plane. First, we describe an algorithm for maintaining the half-plane discrepancy of S under insertions and deletions of points. The algorithm runs in O(nlogn) worst-case time per update, and it requires only relatively simple data structures. Second, we give an algorithm that computes the strip discrepancy of S in O(n2a(n)logn) time, where a(n) is the extremely slowly growing functional inverse of Ackermann's function.