Stochastic Process Models in Device Physics
Abstract
The aim of this project was to undertake a statistical study of Monte Carlo methods in device physics using tools from the modern theory of probability and stochastic processes. The Monte Carlo approach has been used to obtain numerical estimates of parameters by simulating the physical process on a computer rather than attempting to solve the integrodifferential equations which arise. Thus this method estimates averages and can be applied to any quantity expressible as such but may require an enormous number of simulations before statistically reliable results are obtained, especially if the quantity of interest is subject of very low probability of occurrence. A probabilistic study was undertaken of the behaviour of individual electrons, complementing the Monte-Carlo approach. The methodology required the formulation of some novel stochastic processes and suggested new types of conditions for ergodicity of Markov processes.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 20, 1989
- Accession Number
- ADA212284
Entities
People
- Fred M. Hoppe
Organizations
- University of Michigan