ON THE ASYMPTOTIC BEHAVIOR OF HOPFIELD NEURAL NETWORK WITH PERIODIC INPUTS

ＩｎｔｒｏｄｕｃｔｉｏｎＳｉｎｃｅｔｈｅｓｐｅｃｉａｌｓｕｐｅｒｉｏｒｉｔｙｏｆａｒｔｉｆｉｃｉａｌｎｅｕｒａｌｎｅｔｗｏｒｋｓｔｅｃｈｎｏｌｏｇｙｉｎｖａｒｉｏｕｓｅｎｇｉｎｅｅｒｉｎｇｔｅｃｈｎｉｑｕｅｓｆｉｅｌｄｓ,ｓｕｃｈａｓｏｐｔｉｍｉｚａｔｉｏｎ ,ａｓｓｏｃｉａｔｉｖｅｍｅｍｏｒｉｅｓ,ｐａｔｔｅｒｎｒｅｃｏｇｎｉｔｉｏｎ ,ｓｉｇｎａｌｐｒｏｃｅｓｓｉｎｇａｎｄａｕｔｏｍａｔｉｃｃｏｎｔｒｏｌ,ｔｈｅｒｅｈａｓｂｅｅｎｉｎｃｒｅａｓｉｎｇｉｎｔｅｒｅｓｔｉｎｔｈｅｉｎｖｅｓｔｉｇａｔｉ…     

等离子体激励器的静特性与高速特性仿真研究

A parametric study of high-frequency plasma jet actuator was carried out, using the experiment- tally measured energy distribution law of arc discharge as an ideal heat source. The influence of the exit angle of the actuator on the flow field was explored. The jet flow field characteristics of the spark discharge actuator under supersonic flow (Ma0=2.0) were investigated. The results show that the energy density of heat flux increases and the jet front and forward shock wave moves faster with the decrease of discharge region, and the smaller the exit angle of the jet is, the stronger the momentum injection ability of the actuator along the flow direction is. The rule still applies under high-speed air flow conditions. Compared to the static condition, the momentum injection capability of the jet is stronger and the influence domain is larger under supersonic flow conditions.     

具有周期输入Hopfield型神经网络的全局渐近性质

在不假定非线性激励函数有界和可微的条件下,应用Mawhin的重合度理论及Liapunov函数法给出一类具有周期输入的Hopfield型神经网络存在周期解及其全局指数稳定的充分条件。     

非自治Liénard型系统的周期解

利用Mawhin的重合度理论及Brousk定理给出了一类非自治二阶微分系统周期解存在的一些充分条件,所获结果推广与扩展了有关文献的相应结果.     

非自治Lienard型系统的周期解

利用Mawhin的重合度理论及Brousk定理给出了一类非自治二阶微分系统周期解存在的一些充分条件，所获结果推广与扩展了有关文献的相应结果。