Heuristics for the strong generalized minimum label spanning tree problem

In this work we introduce and study the strong generalized minimum label spanning tree (GMLST), a novel optimization problem defined on edge-labeled graphs. Given a label set associated to each edge of the input graph, the aim is to look for the spanning tree using the minimum number of labels. Differently from the previously introduced GMLST problem, including a given edge in the solution means that all its labels are used. We present a mathematical formulation, as well as three heuristic approaches to solve the problem. Computational results compare the performances of the proposed algorithms.