Optimal Communication Spanning Tree Test Instances

Abstract

Searching the literature for standard test problems for the optimal communication spanning tree (OCST) problem reveals that many researchers use their private test problems which are mostly not published. As a result, the comparison of different search algorithms or representations is a difficult and time consuming task. It is not possible to check quickly if a new search method is better than the existing ones. Furthermore, applicants hesitate to use new and efficient search methods or representations if they can not be tested on a variety of different test problems and solve these problems well and reliably. Therefore, the building up of a collection of test instances for the OCST problem is necessary.