Research On Computer Behavior Based On Complexity Theory

Guided by the thought of system integrity, complex behavior of computer network are thoroughly study based on complexity theory.(1)Based on chaos theory complex behavior property of computer networks traffic with different time-scale are deeply researched, the results indicate the universality of the chaos and scale-free behavior in real computer networks.(2)The network behavior evolvement model that correctly describes function mechanism of general real networks is presented, and formal description of the evolvement rules is implemented, the fluctuations formula of input traffic is also deduced.(3) Based on the mean field model of network traffic, the dynamical properties of network traffic system is analyzed, and the condition of chaotic dynamical behavior is achieved under mean field hypothesis.(4) The phase transition phenomena and complexity of global behavior based on network structure property are investigated, and theoretically explain that complexity of network global behavior is maximal near the phase transition traffic point, and the heterogeneity of topology structure has enhancement effect on complicated degrees of network global behavior. (5)On the complex network model, the approximate representation of the local and global phase transition traffic are proposed respectively based on node stabilization and system stabilization condition, the relation of local and global phase transition traffic is clarified, analytical results about variation law of the performance index of system and the degrees departure of network global behavior from stable state under different input traffic load are put forward.(6)Two prediction models of chaos network traffic, namely wavelet weighted local-region model, and mutual information and principal componential wavelet weighted local-region model are put forward.