Runtime analysis of evolutionary algorithms for the depth restricted (1,2)-minimum spanning tree problem

Abstract

The Minimum Spanning Tree problem is a well-known combinatorial optimization problem, which has attracted much attention from the researchers in the field of evolutionary computing. Within the paper, a constrained version of the problem named Depth Restricted (1-2)-Minimum Spanning Tree problem is considered in the context of evolutionary algorithms, which had been shown to be NP-hard. We separately investigate the expected time (i.e., the expected number of fitness evaluations) of the (1+1) EA, the Multi-Objective Evolutionary Algorithm and its two variants adapted to the constrained version, to obtain an approximate solution with ratio 2 or 3/2 with respect to several different fitness functions. In addition, we observe a close connection between the constrained version and the Set Cover problem, and present a simple evolutionary algorithm for the 3-Set Cover problem. Based on the approximate solution returned by our evolutionary algorithm for the 3-Set Cover problem, an approximate solution with ratio better than 3/2 for the constrained version can be constructed.