Synchronization Of A Class Of Fuzzy Neural Networks On Time Scales

The biggest advantages of artificial neural networks with massively parallel computing and distributed storage, nonlinear mapping, strong robustness and fault tolerance, nonlocality, self-adaptive, self-organizing and self-learning, non-convex and association learning. These properties make the neural networks has become a frontier topic relates to physiological science, cognitive, science, psychology, information-computing,microelectronics, optics, electronics and so on, which imply multi-disciplines mutually intercross, comprehensive.And with the control methods of chaos synchronization are developed rapidly, it also let the applied fields of chaos synchronized systems to extensive spread. There-fore, research the synchronous conditions of fuzzy neural networks are very important, which for the systems with time delays, reaction diffusion operators or impulsive effects to determine the stability and applications plays an important role.Synchronization of a class of fuzzy neural networks on time scales is investigated, in this paper.Chapter Ⅰ will briefly introduce the development and the study significance of the neural networks, and also roughly present the development and the investigate signifi-cance of the chaos synchronous phenomena.Chapter Ⅱ is summarily introduce the development and the study significance of differential and integral calculus on time scales and list some basic definitions or theo-rems, which is useful theory.The conditions of synchronization for a class of fuzzy neural networks with mixed delays on time scales be investigated in Chapter III. This chapter is arranged as follows: In part Ⅰ, based on differential and integral calculus on time scales, Lyapunov functional and inequality skills, we establish some sufficient conditions to guarantee the globally exponential synchronization of reaction-diffusion recurrent fuzzy neural networks with continuous and infinite distributed delays, while control gain matrix is diagonal matrix; in part Ⅱ, through configurating a new equivalent system and use the same methods, we obtain some sufficient conditions to ensure the globally exponential synchronization of fuzzy neural networks with mixed delays and impulsive effects, while control gain matrix is positive definite matrix.The results extend some references largely.