Lotka-Volterra Recurrent Neural Networks with Delays
- 312 Downloads
The Lotka-Volterra model of RNNs was first proposed in . Derived from conventional membrane dynamics of competing neurons, it has found successful applications in winner-take-all, winner-share-all and k-winner-take-all problems, see [15, 16, 62]. Due to the application potential of this class of neural networks, it is necessary and useful to study its general dynamic properties.
KeywordsEquilibrium Point Convergence Analysis Recurrent Neural Network Positive Equilibrium Exponential Convergence
Unable to display preview. Download preview PDF.