Heterogeneous Parallel Critical Nodes Detection

The Critical Node Detection Problem (CNDP) is a popular NP-complete graph-theoretical optimization problem with many real-world applications such as influencers in a social network, spreading of pandemics, project critical tasks as well as military strategy planning among many others. Past efforts to solve CNDP problems used heuristics solution with some vendor specific parallel GPU attempts that do not scale well on a heterogeneous system with a combination of CPU and GPU cores. In this paper, we propose and experimentally evaluate a novel heterogeneous algorithm for CNDP that significantly outperforms the earlier best known GPU specific parallel implementation by 2X in the context of real-world Twitter data.