Recently, Ou Shiqin, a faculty member from the College of Mathematics and Statistics at Guizhou University (GZU), has made notable academic progress. Ou is part of two key research teams: the team led by Professor Wang Jinrong at the Guizhou Base of the National Tianyuan Mathematics Southwest Center, and the Guizhou Data-Driven Modeling, Learning and Optimization Innovative Talent Team. As the first author, Ou has published a series of academic papers on neural networks in top-tier journals, including IEEE Transactions on Fuzzy Systems (a leading journal in the field of computer science). These published works provide new theoretical frameworks and methodological foundations for research in related fields, marking a significant contribution to the discipline.
Focused on fuzzy neural networks, the research introduced a state-dependent switching mechanism to construct a state-dependent switching fuzzy neural network model. It further explored issues related to multistability and fixed-time multistability under state-dependent switching of system parameters, with the aim of providing a theoretical basis for the application of neural networks in associative memory. The results show that the state-dependent switching mechanism increases the equilibrium states of the system. Compared with traditional non-switching neural network models, this enables greater storage capacity when applied to associative memory. Additionally, considering the almost periodic nature of system parameters, the study demonstrated that a meta-switching fuzzy neural network equipped with Gaussian activation functions possesses stable almost periodic solutions. Below are presented the phase diagrams of a 2-element switching neural network (with Gaussian activation functions) under two conditions: time-varying almost periodic parameters and non-time-varying parameters.
In addition, based on the multistability of the system, a multi-controller combined control strategy was proposed to achieve fixed-time convergence of multiple equilibrium states. Criteria for determining fixed-time almost periodicity, periodicity, and multistability were established. Related findings have been submitted and accepted by IEEE Transactions on Fuzzy Systems (2025, DOI: 10.1109/TFUZZ.2025.3585694). Based on neural network systems with multiple equilibrium states, the multistability synchronization problem of the corresponding drive-response systems was investigated, with relevant results published in Neural Networks, 180(2024), 106713.
Editor: Zhang Chan
Chief Editor: Li Xufeng
Senior Editor: Ding Long
Translator: Wang Xiaomin
