A New 0–1 ILP Approach for the Bounded Diameter Minimum Spanning Tree Problem

The bounded diameter minimum spanning tree (BDMST) problem is NP-hard and appears e.g. in communication network design when considering certain quality of service requirements. Prior exact approaches for the BDMST problem mostly rely on network flow-based mixed integer linear programming and Miller-Tucker-Zemlin-based formulations. This article presents a new, compact 0–1 integer linear programming model, which is further strengthened by dynamically adding violated connection and cycle elimination constraints within a branch-and-cut environment. The proposed approach is empirically compared to two recently published formulations. It turns out to work well in particular on dense instances with tight diameter bounds.