Stochastic-Sampling-Based Event-Triggered Control for Switching Reaction–Diffusion Neural Networks
研究了切换反应扩散神经网络在随机采样下的多异步时空采样数据控制问题，引入驻留概率更精确描述随机动态，并设计基于随机采样的事件触发方案优化传输频率，通过数值仿真验证了有效性。
This article addresses the issue of multiasynchronous time-space sampled-data control (SDC) for switching reaction-diffusion neural networks (SRDNNs) under stochastic sampling. Unlike the well-known transition probabilities, sojourn probabilities (SPs) are introduced to more precisely represent the stochastic dynamics of SRDNNs. Instead of using a fixed sampling interval, a stochastic variable is employed to describe the aperiodic nature of the sampling period, leading to a stochastic-sampling-based event-triggered scheme to optimize the transmission frequency. To enhance flexibility, a novel time-space SDC strategy is developed that conducts sampling simultaneously in both temporal and spatial dimensions while employing switching gains. Finally, the efficacy and superiority of the proposed control strategy are confirmed through a numerical simulation.