|
Key nodes measure nature of the problem can be attributed to the node importance measure and scheduling problems, the problem of an effective solution to help guide financial, power, supply chain, the Internet, and many other applications to carry out. The current research mainly for the right static social networks, the dynamic weighting network research is relatively small, clearly further research is necessary. Considering the reality single factor measurement node importance limitations, gives the equivalent point concept of the right, on the one hand, the point right as a basic factor, which indeed partially reflects the node importance, on the other hand, Point right on the other nodes diffusion effect, the greater the distance, the smaller the impact, taking into account the global network node point right on the central node affected, given based on the equivalent point right node importance ranking method, which further improves Measurement accuracy of the results. Taking into account the reality weighted network community structure characteristics are given first grouped computing ideas, classic community grouping algorithm can right chart has a good effect, in a weighted graph but with the expected results there are some gaps, To solve this problem, are given based on the incremental distance matrix grouping concept model, taking into account the two communities after the merger channel matrix changes, given the evaluation of the packet the quality of distance increment index for the direct calculation of the path matrix time cost big enough, using a dynamically updated way to shorten the calculation time, so that the packet measurement method can be applied to a larger scale networks. Considering real weighting network changes with time dynamic characteristics, for the path in which the matrix gives the dynamic update method for the joining node, delete node, while the right changes, etc. is given a different update method, greatly reducing the computational overhead, and gives a computational node on the shortest path to a new method, so that it can better meet the actual dynamic network applications. Finally, combining these three aspects are given based on incremental distance matrix groups seeking node of the importance of dynamic algorithm (a dynamic algorithm to calculate the importance of nodes based on the distance-increment matrix grouping, IDD), which mainly includes data packet initialization, data after joining the grouping selected from the matrix updates and Node Importance calculate four aspects, combined with C-DBLP data and verified by experiments that the algorithm is feasible and effective.
|