Realization of limit cycles by neural networks with piecewise linear activation function

Abstract

A simple method of realizing a prescribed limit cycle by means of a continuous-time recurrent neural network with the piecewise-linear activation function is proposed. Since the proposed method is formulated as a set of linear inequalities, we can easily determine parameters of the network by using, for example, the simplex method.

Topics

3 Figures and Tables

Download Full PDF Version (Non-Commercial Use)