A Memetic Algorithm for the Optimum Communication Spanning Tree Problem

For the NP-hard Optimum Communication Spanning Tree (OCST) problem a cost minimizing spanning tree has to be found, where the cost depends on the communication volume between each pair of nodes routed over the tree. We present a memetic algorithm (MA) for this problem and focus our discussion on the evaluation of recombination operators for the OCST. The proposed algorithm outperforms evolutionary algorithms (EA) for known benchmark instances and outperforms state-of-the-art solvers for non-Euclidean instances.

• Type: Conference
• Year: 2007
• Where: HM
• Authors: Thomas Fischer, Peter Merz
• Keywords: Artificial Intelligence, Communication Spanning Tree, HM 2007, Outperforms Evolutionary Algorithms, Spanning Tree