Wireless Sensor Network (WSN) as one of the representatives of the Internet of Things technology has also received much attention. To accurately diagnose fault sensor nodes, a fault diagnosis method based on fireworks algorithm optimization convolutional neural network algorithm is proposed. The weights and biases of the convolutional neural networks are optimized by using the self-regulating mechanism of global and local searching ability of fireworks algorithm. So the problem of convolution neural network in extreme judgment and limited convergence speed is solved, to effectively realize the fault diagnosis of the WSN. Simulation experiments show that this algorithm has higher fault diagnosis accuracy than other classic WSN fault diagnosis algorithms. INDEX TERMS convolution neural network, fault diagnosis, fireworks algorithm, MM* model, wireless sensor network.
Matching composition network is a family of interconnection networks, including the BC network and the hyper Petersen network, etc. In this paper, we prove that the local diagnosability of an MCN can be obtained by adding one to that of the component with some structural restrictions. Additionally, we obtain a sufficient condition for verifying that an MCN with fault-free edges (faulty edges, respectively) has strong local diagnostic properties. By applying these results, the BC network BC n and the hyper Petersen network HP n are n l -diagnosable at each node belonging to them, for n ⩾ 3. These interconnection networks with fault-free edges (faulty edges, respectively) possess the strong local diagnosability property. KEYWORDSfault diagnosis, local diagnosability, matching composition network, PMC model INTRODUCTIONWith the scale of the high-performance parallel computer systems growing larger, the system architecture has become more and more complex.However, the complex architecture of these multiprocessor systems has a negative impact on the reliability of such systems. Due to the influence of many factors such as production process, system integration, and system operating environment, the more processors and communication links, the higher the probability of failure. In some cases, some of the processors fail, which may cause the entire system to lose its ability to work.Consequently, promptly locating and replacing the faulty processors are absolutely essential for guaranteeing the reliability of a multiprocessor system.During the development of system diagnosis, researchers proposed several diagnostic models. The PMC model is undoubtedly one of the most widely used models. 1 There are two different diagnostic strategies depending on whether all faulty nodes in the system are identified one or more times under the PMC model. One is one-step diagnostic strategy and the other is sequential diagnostic strategy. 1 These two diagnostic strategies are both based on the same assumption that all neighbor processors of any one of the processors in the system may fail simultaneously. Therefore, the number of faulty processors that these diagnostic strategies can identify must be limited by the minimum number of neighbors for any one of the processors in the system. To overcome this restriction and increase the system's fault diagnosis capabilities, the researchers proposed different diagnostic strategies and made many attempts. The t/k-diagnosability 2 approach increases the diagnostic capabilities of the system by allowing at most k processors to be diagnosed incorrectly. Obviously, the shortcoming of the t/k-diagnostic strategy is that it may diagnose a fault-free node as a faulty node. Araki et al 3 proposed the (t, k)-diagnosis by introducing a new parameter k, which extended the application of sequential diagnostic strategies and revealed the correlation between one-step diagnosis and generalized sequential diagnosis. However, the choice of the parameter k is a challenge that affects diagnostic efficiency. Provi...
Aiming at the characteristics of MM* model fault diagnosis, a fireworks algorithm based on a dual population strategy is designed. The dual population of the algorithm is operated independently in parallel, and cooperative operator and optimal operator are cross-executed in the iterative process. The cooperative operator enables two populations to exchange effective information, avoiding the premature maturity of the algorithm. The optimal operator helps to strengthen the global search power of the algorithm and improve the convergence rate of the algorithm. At the same time, the constraint equation is designed, a new fitness function is proposed, and the mutation operator and selection strategy are optimized. The experimental comparison shows that the algorithm improves the efficiency and accuracy of system-level fault diagnosis and has good practicability. Finally, the correctness of the algorithm is proved by theory, and the time complexity of the algorithm is analyzed. INDEX TERMS System-level fault diagnosis, MM* model, dual population, fireworks algorithm, t-diagnosable systems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.