Font Size: a A A

Dynamics Of Periodic Delayed Planar Neural Networks

Posted on:2009-03-19Degree:MasterType:Thesis
Country:ChinaCandidate:X ChenFull Text:PDF
GTID:2178360275450675Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
Over the past decade,artificial neural network theory and its application have development of high-profile.Artificial neural network is a function of the human brain inspired by the development of non-biological information processing systems. It can been roughly divided into two groups:hardware or software,which is based on the realization of the artificial neural network,in order to complete some of the specific information processing functions for the purpose of such networks may wish to project known as artificial neural networks.The other is a project of artificial neural network proposed by the mathematical model to express the recurrent neural network.The main type of artificial neural network is to study the characteristics of its power to project artificial neural networks revealing the features and functionality, design and development project for the artificial neural network theory to provide assurance and support.The Hopfield neural network in which the unique features caused by the large number of scholars study on the neural network research in the theory and application of all aspects of the outcome of a breakthrough.Neural network theory applied research has infiltrated a large number of projects in the area,such as intelligent control,pattern recognition,adaptive filtering and signal processing,nonlinear optimization,sensing technology,the robot.It has shown significant aspects of the application.It is the outstanding feature of super-strong peacekeeping non-linear, self-organization,adaptive and self-learning ability,as well as non-local,unsteady and non-convexity and so on.Its application at home and abroad will become a hot issue one of the study.Based on these studies,the main features of this paper is two-dimensional networks of varying delay PDCNN existence of periodic solutions as well as periodic solutions for global stability of the index.In the first part of the neural network research,the background of domestic and foreign research and this paper research on the content and significance are introduced.The theory of coincidence degree and continuity of the theorem are applied to prove the PDCNN cycle in the verification of power.Therefore,in the second part of the paper,we focused on these two fundamental theories of knowledge and background applications.In the third part of the operation through a number of techniques and the use of inequality,has been the existence of periodic solutions.In the fourth part use of the Gronwall's inequality and the right upper Dini derivative,as well as some skill points, we have been requested.Use the same methods we have verified more than two-dimensional time-varying delay Cohen-Grossberg delayed existence of periodic solutions.
Keywords/Search Tags:PDCNN, Neural Networks, Cohen-Grossberg, Periodic solutions, Global Exponential Stability
PDF Full Text Request
Related items