Parallel Algorithms for Maximal Independent Set and Maximal Matching

Abstract

We present a derandomization process which leads to efficient parallel algorithms for the maximal independent set and maximal matching problems. The derandomization of the general pairs PROFIT/COST problem depends on the derandomization of the bit pairs PROFIT/COST problem which follows Luby’s approach of using an O(n) sized sample space. This simplifies the approach presented in [16].