Aiming at the problem that conventional tracking algorithms are difficult to deal with abrupt motion efficiently, an optimization algorithm called hybrid Teaching-learning-based optimization with Adaptive Grasshopper Optimization Algorithm (TLGOA) is proposed in this paper. Firstly, the non-linear strategy based on tangent function is used to replace the linear mechanism in the standard Grasshopper Optimization Algorithm (GOA). The improved adaptive GOA (AGOA) can avoid the local trapping problem and enhance the global optimization ability, which can handle the problem of abrupt motion. Secondly, considering that Teaching-learning-based optimization (TLBO) has obviously local exploitation operator and fast convergence, a hybrid TLGOA tracker is designed by combining the advantages of both AGOA and TLBO. The approach can enable better tracking accuracy and efficiency. Finally, extensive experimental results show that the proposed algorithm has obvious advantages over other algorithms, and also prove that TLGOA tracker is very competitive compared to other state-of-the-art trackers, especially for abrupt motion tracking.

Higher accuracy in cluster failure prediction can ensure the long-term stable operation of cluster systems and effectively alleviate energy losses caused by system failures. Previous works have mostly employed BP neural networks (BPNNs) to predict system faults, but this approach suffers from reduced prediction accuracy due to the inappropriate initialization of weights and thresholds. To address these issues, this paper proposes an improved arithmetic optimization algorithm (AOA) to optimize the initial weights and thresholds in BPNNs. Specifically, we first introduced an improved AOA via multi-subpopulation and comprehensive learning strategies, called MCLAOA. This approach employed multi-subpopulations to effectively alleviate the poor global exploration performance caused by a single elite, and the comprehensive learning strategy enhanced the exploitation performance via information exchange among individuals. More importantly, a nonlinear strategy with a tangent function was designed to ensure a smooth balance and transition between exploration and exploitation. Secondly, the proposed MCLAOA was utilized to optimize the initial weights and thresholds of BPNNs in cluster fault prediction, which could enhance the accuracy of fault prediction models. Finally, the experimental results for 23 benchmark functions, CEC2020 benchmark problems, and two engineering examples demonstrated that the proposed MCLAOA outperformed other swarm intelligence algorithms. For the 23 benchmark functions, it improved the optimal solutions in 16 functions compared to the basic AOA. The proposed fault prediction model achieved comparable performance to other swarm-intelligence-based BPNN models. Compared to basic BPNNs and AOA-BPNNs, the MCLAOA-BPNN showed improvements of 2.0538 and 0.8762 in terms of mean absolute percentage error, respectively.