Premium
Finding all stable equilibria of cellular neural networks
Author(s) -
Fajfar Iztok,
Bratkovic Franc
Publication year - 1994
Publication title -
international journal of circuit theory and applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.364
H-Index - 52
eISSN - 1097-007X
pISSN - 0098-9886
DOI - 10.1002/cta.4490220205
Subject(s) - piecewise linear function , artificial neural network , simple (philosophy) , computer science , set (abstract data type) , sign (mathematics) , binary number , tree (set theory) , algorithm , task (project management) , mathematical optimization , mathematics , artificial intelligence , combinatorics , mathematical analysis , philosophy , geometry , arithmetic , management , epistemology , economics , programming language
Abstract An efficient algorithm is given for finding all stable equilibrium states of a cellular neural network. the method is based upon a new theorem concerning the network dynamics. the theorem implies the implementation of a simple sign test which, combined with a binary tree search, efficiently performs the task of searching the equilibria. Since the proposed algorithm uses information about the network architecture and dynamic behaviour, it exhibits two important advantages over other more general methods for searching the equilibria of resistive piecewise‐linear circuits. First, the regions which cannot possibly include stable steady states can be eliminated even before starting to perform the algorithm. Secondly, there is no need to solve any set of linear equations, which is usually the most tedious part of equilibrium‐searching methods.