Influence of Parallelism on Computational speed and error prediction for Artificial Neural Network based Parameter Optimization in Nano devices

Nayak, Dayananda and Hamsaveni, N and Puttamadappa, C and Subir Kumar, S (2011) Influence of Parallelism on Computational speed and error prediction for Artificial Neural Network based Parameter Optimization in Nano devices. International Journal of Engineering Research and Industrial Application, 4 (1). pp. 277-292. ISSN 0974-1518

[img] PDF
2. IJERIA, Vol.4, No.I, Feb 2011.pdf - Published Version
Restricted to Registered users only

Download (298kB) | Request a copy
Official URL: http: //www.ascent-journals.com

Abstract

A soft computing tool, Artificial Neural Network (ANN), is used in the present work to get optimized system parameters of GaAs and (In, Ga)As quantum wells for desired device characteristics. A combination of device scaling with new device structure and/ or new material has stimulated active researches in the area of nanodevices. To extract the best performance from a device it is very essential to choose the parameters intelligently as the performance is controlled by them. Hence to predict the best parameters at a time it is better to think over some special computational methods. As an attempt towards such computing, adaptive intelligent controls by neural network are deployed here to predict the device parameters so as to get the best performance from the device. In the present work we have investigated in details the influence of parallelism on computational speed and error prediction for ANN based parameter optimization in nanodevices. Special attenuation is taken to investigate the effects parallelism on training time, rms errors for change in the number of data set and iteration numbers.

Item Type: Article
Subjects: Engineering > MIT Manipal > Electronics and Communication
Engineering > MIT Manipal > Instrumentation and Control
Depositing User: MIT Library
Date Deposited: 20 Oct 2016 10:11
Last Modified: 20 Oct 2016 10:11
URI: http://eprints.manipal.edu/id/eprint/147284

Actions (login required)

View Item View Item