Stability analysis for uncertain switched neural networks with time-varying delay

Area of Science:

• Control theory
• Artificial intelligence
• Dynamical systems

Background:

• Switched neural networks are complex systems with applications in various fields.
• Time-varying delays introduce significant challenges in analyzing system stability.
• Existing methods often oversimplify by assuming all subsystems are stable, leading to conservative results.

Purpose of the Study:

• To investigate the exponential stability of uncertain switched neural networks with time-varying delays.
• To develop a novel approach that accounts for mode-dependent properties of subsystems.
• To derive less conservative stability conditions compared to existing methods.

Main Methods:

• Categorization of subsystems into stable and unstable based on mode-dependent properties.
• Utilization of Lyapunov-like function method.
• Application of the average dwell time technique.

Main Results:

• Derivation of delay-dependent sufficient conditions for exponential stability.
• Demonstration that distinguishing between stable and unstable subsystems yields less conservative criteria.
• Validation of the proposed approach through two numerical examples.

Conclusions:

• The proposed method effectively guarantees the exponential stability of uncertain switched neural networks with time-varying delays.
• The approach offers improved (less conservative) stability criteria by differentiating subsystem stability.
• Numerical examples confirm the validity and advantages of the developed technique.