Worst-case analysis of generalized heapsort algorithm revisited

Abstract

In this paper we present a generalized heapsort algorithm and its worst-case complexity analysis. The weighted sum of the number of comparisons and movements has been defined as a measure of the complexity. Theoretical analysis suggests that, by this criterion, 3-heap should be optimal in the worst case, contradicting the findings of A. Paulik (Paulik, A., 1990, Worst-case analysis of a generalized heapsort algorithm. Information Processing Letters, 36, 159–65.). Experimental data supporting the analysis are also presented.