Modeling of Evolving Knowledge Network with Node Deletion

A new type of knowledge evolving model which comprises knowledge-based local world preferential attachment and degree-based node preferential deletion is studied. A series of numerical simulations of the degree distribution of the knowledge networks are presented. The results indicate that the degree distribution of the knowledge networks at first has an approximate state of scale-invariant, and then it turns into an exponential distribution. Moreover, the truncation in p(k) increases as the adding probability pa decreases. In addition, the probability of the numbers having relatively more connections is greater when the adding probability is larger. Finally, we compare the effect of the size of local world on degree distribution. The results of simulations show that the probability of the numbers having relatively larger degree will increases as the size of local world increases which indicate that the connectivity is more and tighter as M increases.