the algorithm behaves similar to the degree-based heuristics that favors vertices withhigh degree. However, after a certain number of vertices have been selected, VirAdswill make the selection based on the information within d-hop neighbor around theconsidered vertices, which is different from degree-based heuristic that considersonly one-hop neighborhoodship.Given those measures, VirAds selects in each step the vertex u with the highest e f f ectiveness which is defined as n(e)u + n(a)u. After that, the algorithm needs toupdate the measures for all the remaining vertices.It is introduced in [13] that the cost-effective, massive and fast propagation problem (CFM) can be easily shown to be NP-hard by a reduction from the set coverproblem. It is also proved that there is unlikely an approximation algorithm withfactor less than O(logn). However, if we assume the network is power-law, theiralgorithm is an approximation algorithm for this problem with a constant factor.
đang được dịch, vui lòng đợi..
