"Filter for optimization of stochastic processes"
No Thumbnail Available
Authors
Meeting name
Sponsors
Date
Journal Title
Format
Thesis
Subject
Abstract
A search method for optimization of stochastic systems was proposed by Luecke (1970) in which the effect of noise was reduced by replication of experiments. In this work, a further reduction of the effect of noise is obtained with a filter based on object function determinations made at previous points in the course of the search. This filter was based on sequential regression analyses that fitted past data to a quadratic. The value of the object function at the new search point as predicted by the quadratic was combined with the measured value to give an unbiased, minimum variance estimate. Use of this estimator yielded a significant improvement in the response of the search, but a wandering, random phenomena was observed as the search approached the neighborhood of he optimum. This behavior indicated that noise reduction was insufficient in that region so a further modificatior was made to increase filtering as the search progressed. Substantial improvements were obtained and this modified search was far more efficient than the original procedure.
Table of Contents
PubMed ID
Degree
M.S.
Thesis Department
Rights
OpenAccess.
License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License.