Reliable Cluster on Uncertain Multigraph

In real-world heterogeneous network, computing devices are generally equipped with multiple network devices. Also, the connection between devices can be changed dynamically, thus, the connectivity between devices needs to be modeled deliberately, after probabilistic models. From these two factors, this paper proposes to use uncertain multigraph to model heterogeneous network. Furthermore, this paper considers the reliable cluster problem on this uncertain multigraph. Finding reliable cluster can identify set of network devices that unlikely to be disconnected and should be assigned as core network devices. An efficient clustering algorithm is presented to address the problem and the evaluation results show that the proposed algorithm can find reliable clusters faster than base-line algorithms.