Maximum Leaf Spanning Tree Problem Benchmarks
Description
This database contains experimental problems designed to study the Maximum Leaf Spanning Tree Problem (MLSTP). More specifically, the dataset can be used to evaluate the performance of algorithms developed to solve MLSTP. We generated a set of large-scale instances. We also collected existing benchmarks {Lucena, A., Maculan, N. & Simonetti, L. Reformulations and solution algorithms for the maximum leaf spanning tree problem. Comput Manag Sci 7, 289–311 (2010). https://doi.org/10.1007/s10287-009-0116-5, GENDRON, B., LUCENA, A., DA CUNHA, A. S. & SIMONETTI, L. (2014), "Benders Decomposition, Branch-and-Cut, and Hybrid Algorithms for the Minimum Connected Dominating Set Problem", INFORMS Journal on Computing, 26, 645-657, doi: https://doi.org/10.1287/ijoc.2013.0589.} which were included in the dataset. These instances are used in a study entitled “A New Formulation and Algorithm for Maximum Leaf Spanning Tree Problem with an Application in the Forest Fire Detection” which will be appeared in ----. DOI reference: http://dx.doi.org/10.17632/w98s4tvfn8.1
Files
Steps to reproduce
This database contains experimental problems designed to study the Maximum Leaf Spanning Tree Problem (MLSTP). The dataset is organized in three separate folders, including Existing_Instances which covers all previously reported instances in literature {Lucena, A., Maculan, N. & Simonetti, L., Reformulations and solution algorithms for the maximum leaf spanning tree problem. Comput Manag Sci 7, 289–311 (2010). https://doi.org/10.1007/s10287-009-0116-5, GENDRON, B., LUCENA, A., DA CUNHA, A. S. & SIMONETTI, L. (2014), "Benders Decomposition, Branch-and-Cut, and Hybrid Algorithms for the Minimum Connected Dominating Set Problem", INFORMS Journal on Computing, 26, 645-657, DOI: https://doi.org/10.1287/ijoc.2013.0589.}, New_Large_Scale_Instances having 54 new large-scale instances generated randomly, and a real-world instance related to Arasbaran forest in Northwestern Iran. We have generated new random large-scale instances based on two key parameters, namely, the number of graph nodes and graph density. The number of nodes of graphs ranges between 300 to 2000 and graph density varies from 0.05 to 0.7. The data for Arasbaran forest is the pre-determined locations of sensors for early fire detection. The edges of the graph are defined based on sensors' connectivity distance. More specifically, sensors are supposed to be fully connected with neighboring sensors.