Dynamic analysis and implementation of an absolute memristor Hopfield neural network
The Hopfield neural network is a brain-like neural network that can exhibit rich dynamical behaviors,especially chaos.To explore these rich dynamical phenomena,a Hopfield neural network model was investigated with memristive synaptic weights,which has three equilibrium points.The dynamic behavior of the system was analyzed by dynamic maps,bifurcation diagrams,Lyapunov exponent spectrum,phase trajectory diagrams,and time series diagrams.Investigation was also performed to study the effect of memristive synaptic coupling strength and initial conditions on the system's dynamic behavior.The experimental results show that the constructed Hopfield neural network can not only exhibits period-doubling bifurcation,chaotic and periodic windows,and symmetric co-existing attractors under different memristive strength,but also produces various types of asymmetric coexistence attractors with the change of the initial state value(such as periodic and chaotic coexistence attractors,single-scroll and double-scroll coexistence attractors),and also discovers transient chaos phenomena.Finally,the digital circuit of the system was implemented by FPGA technology,verifying the correctness and realizability of the proposed system.
万方数据
维普
