Use of random topics as practical control signals in a social network model

In this paper, we study practical strategies for controlling the behaviour of a synthetic social network modelling the dynamic diffusion of knowledge. The problem of controlling the evolution of complex networks has been extensively studied in recent years and remarkable theoretical results have been achieved. However, still largely unexplored is the analysis of realistic control strategies for complex networks and the special case of social networks. Our model of knowledge diffusion in a social network is used for simulating and evaluating possible control strategies of social network behaviour. Our approach is to exploit the controlled injection of random topics into some driver nodes for influencing the overall dynamics. This way, it is possible to modify some key control parameters in a deterministic way with realistic inputs, considering the strong practical constraints of social networks with respect to control measures. Control parameters considered are: The injection interval of random topics, the rate of driver nodes with respect to the network size, and the selection criteria of driver nodes. Finally, we discuss possible applications and the challenges that social networks pose to the issue of network control.