Chan’s Algorithm • It is Jarvis’s march applied to big blobs of points Chan’s algorithm • Each blob is the convex hull of m points • What is m ? • m = min(exp(2, exp(2,t)), n) Chan’s algorithm • Partition P into r disjoint sets of size m • Compute convex hull of each set • Set p0 = (-∞, 0) and let p1 be the bottommost point • Do an m – step Jarvis march (note this) • Stop if the march returns to p1 ,else increase m and repeat Chan’s algorithm • Analysis – Time-complexity in O(n log h) » The time-complexity of each iteration is in O(n log m) where m = exp(2, exp(2,t)), for t =1, 2, … » Number of iterations is bounded above by log log h » Thus the complexity is O(n exp(2, log log h + 1)). This simplifies to O(n log h)