A novel weighted evolving network model based on the clique overlapping growth was proposed.The model shows different network characteristics under two different selection mechanisms that are preferential selection and random selection.On the basis of mean-field theory,this model under the two different selection mechanisms was analyzed.The analytic equations of distributions of the number of cliques that a vertex joins and the vertex strength of the model were given.It is proved that both distributions follow the scale-free power-law distribution in preferential selection mechanism and the exponential distribution in random selection mechanism,respectively.The analytic expressions of exponents of corresponding distributions were obtained.The agreement between the simulations and analytical results indicates the validity of the theoretical analysis.Finally,three real transport bus networks(BTNs) of Beijing,Shanghai and Hangzhou in China were studied.By analyzing their network properties,it is discovered that these real BTNs belong to a kind of weighted evolving network model with clique overlapping growth and random selection mechanism that was proposed in this context.
A novel scale-flee network model based on clique （complete subgraph of random size） growth and preferential attachment was proposed. The simulations of this model were carried out. And the necessity of two evolving mechanisms of the model was verified. According to the mean-field theory, the degree distribution of this model was analyzed and computed. The degree distribution function of vertices of the generating network P（d） is 2m^2m1^-3（d-m1 ＋ 1）^-3, where m and m1 denote the number of the new adding edges and the vertex number of the cliques respectively, d is the degree of the vertex, while one of cliques P（k） is 2m^2Ek^-3, where k is the degree of the clique. The simulated and analytical results show that both the degree distributions of vertices and cliques follow the scale-flee power-law distribution. The scale-free property of this model disappears in the absence of any one of the evolving mechanisms. Moreover, the randomicity of this model increases with the increment of the vertex number of the cliques.
