Title:A novel pinning observability strategy for Boolean networks

Abstract:Observability, as is well-known, is of biological and engineering significance for the study of complex networks with a mass of state variables (SVs), while sensors are commonly impossible or very difficult to be inflicted on all SVs. Taking an unobservable system into account, it is significantly crucial to design an operably effective control strategy under which the system achieves observability. In this study, a novel design strategy of pinning controller is developed for an unobservable Boolean network by taking advantage of the original network structure (NS) with respect to (w.r.t.) $n$ SVs rather than the traditionary state transition diagram w.r.t. $2^n$ vertices based on the semi-tensor product of matrices. The application of the original NS information dramatically reduces the time computationally complexity from $O(2^{4n})$ to $O((n-m)2^n)$ with $m$ being the output number. Moreover, the designed pinning controller by our approach can easily pick the pinning nodes and its feedback form is more concise. With regard to simulation, a biological network, called T-LGL survival network, with $18$ SVs successfully becomes to be observable under the designed pinning controller.