Detecting faulty edges of complex dynamical networks based on compressive sensing

Abstract

As an important technology to improve network reliability, fault diagnosis has gained wide attention in complex dynamical networks. However, few studies focused on detecting the structure of broken edges when faults occur. In this paper, due to the natural sparsity of complex dynamical networks, a completely data-driven method based on compressive sensing is established to detect the structure of faulty edges from limited measurements. The least absolute shrinkage and selection operator algorithm is applied to solve the reconstruction problem. In addition, the method is also applicable to multilayer networks. The faulty edges in both the intralayer network and the interlayer network can be fully identified. Compared with other methods, the main advantages of the proposed method lie in two aspects. First, the structure of faulty edges can be obtained directly with limited measurements. Second, the proposed method is less time consuming and more efficient due to less data processing. Numerical simulations involving single-layer, multilayer and real-world complex dynamical networks are given to demonstrate the accuracy of detecting the structure of faulty edges from the proposed method.