Parallel algorithms for height balancing binary trees

Abstract

The authors present a simple parallel algorithm to height-balance a binary tree. The algorithm accepts any arbitrary binary tree as its input and yields an optimally shaped binary tree. For any arbitrary binary tree of n nodes the algorithm has a time complexity of O(lgn) and utilizes O(n) processors on a EREW PRAM model. The algorithm uses Euler tours and list ranking, which form the building blocks for many parallel algorithms.<>