Unravelling Greedy Algorithms

Abstract

In my previous Functional Pearls article (Bird, 1992), I proved a theorem giving conditions under which an optimization problem could be implemented by a greedy algorithm. A greedy algorithm is one that picks a ‘best’ element at each stage. Here, we return to this theorem and extend it in various ways. We then use the theory to solve an intriguing problem about unravelling sequences into a smallest number of ascending subsequences.