Bifurcations Due to Different Neutral Delays in a Fractional-Order Neutral-Type Neural Network
Chengdai Huang, Heng Liu, Tingwen Huang, Jinde Cao · IEEE Transactions on Emerging Topics in Computational Intelligence · 2023
This article lucubrates the bifurcations in respect of a fractional-order neutral-type neural network(FONTNN) with two nonidentical delays. To begin with, the characteristic equation of the linearized FONTNN is deliberated and the nonidentical delays-induced bifurcation conditions are attained. It elucidates that the devised FONTNN is capable of maintaining stability performance when choosing a lesser time delay. Afterwards, FONTNN illustrates exceptional stability performance by detailed comparisons with regard to integer-order neutral-type delayed NNs. The effects of the coefficients for self-inhibition or time delay on the bifurcation points are carefully undertaken. It manifests that the coefficients for self-inhibition or time delay can postpone(advance) the onset of bifurcation of the developed FONTNN. Furthermore, fractional-order retarded delayed neural network(FORDNN) exhibits better performance in comparison with FONTNN. Conversely, it hints that the acquired stability results are fabulously inaccurate once neglecting the reverberations of neutral delays for FONNs. The correctness of the developed results is lastly indicated via numerical experiments.