DASH: A novel method for dynamically selecting key nodes to spread information rapidly under the graph burning model

Abstract

Seldom research considers both the “speed” of information spreading and also the dynamic way of selecting seeds, i.e., choosing node step by step instead of one-time selecting. Graph burning model is used here to meet these requirements of information spreading, whose objective is to find a shortest sequence of nodes and activate them sequentially so that the network is eventually activated completely. This problem has been proven to be NP-Hard. However, few heuristic or approximate algorithms exist, even for trees. Thus, in this paper we propose a novel tree burning algorithm called DASH, where the node with both a strong burning capacity and large spacing to the existing activated nodes is selected as a new seed sequentially. This DASH algorithm is implemented and compared with existing algorithms on 20 spanning subtrees of real networks and 15 randomly generated trees, and the experimental results demonstrate that DASH performs better.