Multi-scroll attractor and its broken coexisting attractors in cyclic memristive neural network

Area of Science:

• Neuroscience
• Chaos Theory
• Nonlinear Dynamics

Background:

• Memristive neural networks offer novel computational paradigms.
• Generating multi-scroll chaotic attractors is crucial for complex system modeling.
• Existing multi-scroll systems often have complex structures.

Purpose of the Study:

• To propose a simple-structured memristive neural network.
• To investigate the generation of multi-scroll chaotic attractors.
• To analyze the dynamic behaviors and circuit feasibility.

Main Methods:

• Designing a three-neuron memristive neural network with self-connections.
• Utilizing phase portraits, bifurcation diagrams, and Lyapunov exponents for analysis.
• Implementing a circuit realization platform to validate the system.

Main Results:

• The proposed network generates multiple coexisting multi-scroll attractors.
• Amplitude control and breakup of attractors into centrosymmetric pairs were observed.
• Abundant dynamic behaviors were systematically studied.

Conclusions:

• The simple memristive neural network effectively generates controllable multi-scroll attractors.
• The system exhibits rich dynamics and practical feasibility through circuit implementation.
• This work contributes to the development of advanced memristive systems.