Research On Community Structure Division Problems Of Weighted Complex Network Based On Partheno-genetic Algorithm

At present,many scientists are focusing on the research for complex network With the deepening of the complex network research,it was found that that the non-weighted network could only give a qualitative description of the existence or nonexistence for the interaction between a set of given nodes.While in practical network analysis procedures,it is essentially important for the relationships between the vertices or the differences of the interaction strength, for example, the broadband flux for the Internet network、the cooperation number in the scientists’cooperation network are both main factors influencing the equality of the network. But the weighted network can objectively reflect a richer information in order to satisfy the people’s needs. community structure is one of the important characteristics of complex network, it is naturally especially important for the study of the community structure of weighted networkThe most algorithms on community division of complex network are advised for the non-weighted ones, only a few algorithms are constructed for the weighted ones. In this article,we divided an algorithm on the community division of weighted complex network based on the Partheno-Genetic Algorithm. In this paper,firstly,we introduced the concept of weighted complex network metrics, generalized the module function widely used in the non-weighter networks to the weighted ones,and proposed a new fitness function QwF to evaluate the community division of weighted complex network. Secondly,we gave an introducition to the algorithms on community division of both the non-weighted networks and the weighted ones. Finally,under the theoretical basis of the Partheno-Genetic Algorithm,we extended the coding processing used in non-weighted complex network to the community division of the weighted ones, constructed a new method for initialization combining with the characteristics of the weighted complex networks,and introduced three kinds of optimization operators,which advanced the efficiency of the algorithm.