| Stabılıty Analysıs Of Bı-Dırectıonal Assocıatıve Memory Neural Networks
This thesis studies the stability properties of a general class of bidirectional associative memory Neural Networks (BAM). In this thesis, a new sufficient condition for the existence, uniqueness and global asymptotic stability of BAM is presented without assuming the symmetry of interconnection matrices, and the boundedness and differentiability of the activation functions.
In Chapter 1, the main objective of this thesis is explained in details.
In Chapter 2, the concepts that are necessary to understand this thesis are explained by the support of examples.
In Chapter 3, the terms, mathematical expressions that are used in this thesis are explained, some theorems, rules and hypothesis that are thought to be useful to know, it is mentioned about the stability concept that builds the main part of this thesis and how the analysis of stability is made and the Stability Analysis of Bidirectional Associative Memory Neural Networks.
In Chapter 4, the stability analysis that was carried out in Chapter 3, is simulated by MATLAB, the values that are obtained are showed with figures and graphs. The results obtained can be tested easily by using MATLAB.
In Chapter 5, a general assessment is made about the subject of our thesis by carrying about the stability analysis that was made in Chapter 3 and the values that are showed in Chapter 4. The results obtained are compared with the other results in the literature.
ABO SAALEEK Mohammad ,
Danışman : Doç.Dr. A.Halim ZAİM
Ana Bilim Dalı : Bilgisayar Müh.
Mezuniyet Yılı : 2006
Tez Savunma Jürisi : Doç.Dr. A.Halim ZAİM (Danışman)
Prof.Dr.Ahmet SERTBAŞ
Doç.Dr.Sabri ARIK
Prof.Dr.Mahmut ÜN
Yrd.Doç.Dr.Oğuzhan ÖZTAŞ
Dostları ilə paylaş: | 
