Title :
Neural network technique for the speed-up of Monte-Carlo based semiconductor simulators
Author :
Matei, R. ; Dima, G. ; Profirescu, M.D.
Author_Institution :
R&D Centre, Univ. Politehnica of Bucharest, Romania
Abstract :
The paper presents a way for improving the simulation time of the Monte-Carlo based simulators using a neural network structure. A multi-layered feed-forward neural network trained with a quasi-Newton algorithm was used. As an example, the extraction of the bulk transport parameters of a III-V compound semiconductor is discussed
Keywords :
III-V semiconductors; Monte Carlo methods; Newton method; feedforward neural nets; III-V compound semiconductor; Monte Carlo simulation; bulk transport; multilayered feedforward neural network; quasi-Newton algorithm; Computational modeling; Computer networks; Electronic mail; Feedforward neural networks; Feedforward systems; III-V semiconductor materials; Multi-layer neural network; Neural networks; Parallel processing; Research and development;
Conference_Titel :
Semiconductor Conference, 2000. CAS 2000 Proceedings. International
Conference_Location :
Sinaia
Print_ISBN :
0-7803-5885-6
DOI :
10.1109/SMICND.2000.890254
