事件触发下马尔可夫跳跃神经网络的随机同步

Stochastic synchronization of Markovian jump neural networks via event-triggered control

中文关键词  神经网络  随机同步  事件触发控制  马尔可夫跳跃参数

英文关键词  neural networks  stochastic synchronization  event-triggered control  Markovian jump parameters

中文摘要

      本文研究了事件触发机制下带有随机噪声的马尔可夫跳跃神经网络的随机同步问题. 为了更有效地降低 数据传输量和节约网络资源, 本文采用了一种事件触发控制. 当传输误差和状态误差满足触发条件时, 数据才能够 被传输, 使得主从系统可以在有限的资源和带宽下实现同步. 通过构建新的Lyapunov泛函, 以及使用广义Dynkin公 式和不等式分析方法, 得到误差系统的稳定性条件, 并能够进一步保证主系统和从系统的随机同步. 最后利用 MATLAB进行仿真实验, 结果表明与采样数据控制相比, 事件触发控制能够有效减少数据传输次数, 同时该数值例 子验证了所得结果的可行性以及有效性.

英文摘要

      In this paper, the stochastic synchronization problem for Markovian jump neural networks via an eventtriggered mechanism was studied. In order to reduce the amount of data transmission more effectively and save network resources, an event-triggered mechanism was adopted in this paper. Data can be transmitted when the transmission error and the state error satisfy the event-triggered condition so that the master-slave systems could synchronize with limited resources and bandwidth. The stability condition of the error system was obtained by establishing new Lyapunov functional, utilizing the generalized Dynkin formula, and using the inequality analysis method, which further ensured the stochastic synchronization of the master system and the slave system. In the end, the simulation experiment was carried out through MATLAB. The results showed that compared with the sampled data control, the event-triggered control can effectively reduce the number of data transmissions, and the numerical example verified the feasibility and effectiveness of the results.