Instances for “Rapid Influence Maximization on Social Networks: The Positive Influence Dominating Set Problem"

Published: 29 September 2021| Version 1 | DOI: 10.17632/ywfgkk5pky.1
Contributors:
,
Rui Zhang

Description

We provide the instances used in the paper “Rapid Influence Maximization on Social Networks: The Positive Influence Dominating Set Problem”, by S. Raghavan and Rui Zhang. This repository contains the 100 instances used in the paper. All the instances used in the paper are provided in a compressed archive. The accompanying data is contained in the following file: • PIDS_Instances.zip Description: There is one main folder, which contains 100 instances based on 10 real-world graphs. For graphs Gnutella, Anybeat, Advogato, Escorts, Hamster, Ning, and Delicious, the setting is as follows: For each instance file, there are m + 2 lines. The first m lines provide the edges in the graphs. Nodes are labeled from 0 to n where n is the largest number in the first m lines. The (m + 1)th line contains the weight (b) for each node. The (m + 2)th line contains the threshold value (g) for each node. For graphs Flixster, Youtube, and Lastfm, the setting is as follows: Each real-world graph “G” is described by the file named “G_Graph.txt” which contains the edges in the graph. Nodes are labeled from 0 to n, where n is the largest number in the file. Each line provides the two end nodes of an edge. The 10 instances associated with each graph “G” are provided in the 10 files named “G_i.txt” for i in {0, 1, · · · , 9}. In each file, there are two lines. The first line contains the weight (b) for each node. The second line contains the threshold value (g) for each node. The excel file “PIDS_Results.xlsx” reports, for each instance, the upper and lower bounds obtained in the paper.

Institutions

University of Maryland at College Park, University of Colorado Boulder

Categories

Integer Programming, Social Networks, Model Formulation, Peer Influence, Facet, Branch and Cut