A Multistable Memristor and Its Application in Fractional-Order Hopfield Neural Network

Nowadays multistability memristors constitute a relatively rare research topic. In this paper, a locally active memristor with multistability is constructed and the mechanism of multistable dynamics is also expounded. In addition, the locally active memristor is used to replace the self-connection synaptic weight in a fractional-order Hopfield neural network. There are not only infinite coexisting attractors but also asymmetric coexisting attractors in this memristive fractional-order neural network. Coexisting dynamic is investigated through dynamic behavior analysis methods including bifurcation diagram, Lyapunov exponent spectrum, phase diagrams and complexity analysis. Finally, the analog circuit is implemented by discrete components and the circuit is simulated in PSIM.