Implementation of Hopfield Neural Network using Double Gate MOSFET

ISBN-13: 9783639166644 / Angielski / Miękka / 2009 / 100 str.

Amit Borundiya

Opis

Chaotic phenomenon has a good ability to search for an optimal solution of a problem with many local minima. The current work uses the chaos theory with focus on electrical circuits. Chua's circuit is a good example to demonstrate the type of non- linearity in the electrical circuit which causes chaos.The opamp based hysteresis neuron model is implemented and its stability is explained. Stability criterion of the network can be used as a cost function and when the network reaches its stable point, this corresponds to a solution. This is explained in regards to N-queens problem.The circuit topology for solving 8-queens problem is proposed that can be extended to N-queens.The hysteresis neuron model is implemented using the double gate MOSFET. Once we are able to build this analog circuit representation of the network, the solution can be obtained in a constant time irrespective of the network size.

O autorze

Amit Parasmal Borundiya completed his M S from Ohio University,
USA after completing B.Tech from National Institute of
Technology, Kurushetra,India. His research interests are in the
field of VLSI Design and Neural Network based applications.
Apart from academics, his interests are also in soccer and
hiking.