ANALYSIS OF TIME-VARYING CELLULAR NEURAL NETWORKS FOR QUADRATIC GLOBAL OPTIMIZATION

The algorithm for quadratic global optimization performed by a cellula r neural network (CNN) with a slowly varying slope of the output chara cteristic (see References 1 and 2) is analysed. It is shown that the o nly CNN which finds the global minimum of a quadratic function for any values of the input parameters is the network composed by only two ce lls. If the dimension is higher than two, even the CNN described by th e simplest one-dimensional space-invariant template (A) over cap=[A(1) ,A(0),A(1)], fails to find the global minimum in a subset of the param eter space. Extensive simulations show that the CNN described by the a bove three-element template works correctly within several parameter r anges; however, if the parameters are chosen according to a random alg orithm, the error rate increases with the number of cells. (C) 1998 Jo hn Wiley & Sons, Ltd.