Research On The Sequence Of The Nodes Importance In Complex Network

Many complicated relationship in real life can be modeled by complex networks.We can discover the characteristics and regulations of the existing networks through theresearch on complex networks. One of the most important tasks for complex networksis to evaluate the importance of nodes. As different nodes have different characteristics,it is of high practical values to evaluate the performance of different nodes. The maincontributions of this thesis are summarized as follows:1. A peeling off sorting algorithm based on elimination technique is proposed. Wefirst peel off the least important nodes until all of the nodes are peeled off, then wereverse the order and get the final order of all the nodes. The simulation results showthat the proposed algorithm can address the limitations of the existing sortingalgorithms, e.g., degree sorting and closeness sorting algorithms. The results also showthat our algorithm is more effective than the compared ones.2. We analyzed the limitations of the closeness sorting algorithm, and then basedon this analysis, an improved closeness sorting algorithm is proposed. In this method,we design a weight function for closeness sorting. The results indicate that withdifferent weight functions, different nodes importance sequences are obtained. Finally, astable weight function and a stable closeness sorting algorithm are proposed. The mainidea is that the final sequence of nodes is unchanged even if we enlarge the differencesamong nodes. The experimental results show that the closeness sorting algorithm withweight function is effective.