[N-1] Nonlinear Problems

3:15 PM - 3:30 PM JST(6:15 AM - 6:30 AM UTC)Candidate for Academic Incentive Award

[N-1-21] Analysis of Global Stable Periodic Orbits in Discrete-time Hysteresis Neural Networks

◎Ryota Toyama¹, Ryoga Nakamura¹, Toshimichi Saito¹ (1. Hosei Univ.)

Keywords:

Hysteresys Neural Networks,Periodic Orbit,Stability

This paper studies fundamental dynamics of a discrete-time hysteresis neural network. The network is characterized by hysteresis threshold parameters and connection parameters. Depending on the parameters, the network can generate a variety of binary periodic orbits and binary fixed points. For simplicity, we analyze 6-dimensional networks comprehensively and clarify generation of a global stable binary periodic orbit: all initial points fall into the orbits. The periodic orbits are applicable to control signals of switching circuits and systems.