L1 Stability Criteria of Neural Networks with Bounded or Unbounded Time-Varying Delays

This paper aims to analyze the L1 stability of neural networks (NNs) with bounded or unbounded time-varying delays. Firstly, a new method based on system solutions is proposed, from which the L
1
stability criteria can be constructed. The presented method does not need to establish any Lyapunov-Krasovskii functional (LKF), which can greatly reduce the amount of calculation. So, the obtained L
1
stability conditions are simpler, and can be checked by standard software tools. Additionally, the advantages of the proposed conditions on L1 stability are demonstrated via representative numerical examples. It is worth emphasizing that this article is the first time to analyze the L1 stability performance of neural networks systems with bounded or unbounded time-varying delays.