Degree Type

Dissertation

Date of Award

1983

Degree Name

Doctor of Philosophy

Department

Statistics

Abstract

One of the major developments in the theory of visual search is the establishment of a performance model based on fitting the search time distribution. Such a distribution is examined, based on a paper by Morawski et al;A modification of the traditional traveling salesman problem is made to relate specifically to the development of optimal search strategies. The modification involves inserting "capture" probabilities at the "cities" to be visited, and adapts the traditional dynamic programming algorithms to this added stochastic feature. A countably infinite version of this stochastic modification is formulated. For this formulation, typical ingredients of infinite dynamic programs are explored; these include: the convergence of the optimal value function, Bellman's functional equation, and the construction of optimal (in this case only conditionally optimal) strategies;Visual search is a process involving certain deterministic, as well as random, components. This idea is incorporated into a second search model for which the expected value, variance and distribution of search time are computed, and also approximated numerically. A certain accelerated Monte Carlo method is discussed in connection with the numerical approximation of the distribution of search time.

DOI

https://doi.org/10.31274/rtd-180813-8000

Publisher

Digital Repository @ Iowa State University, http://lib.dr.iastate.edu/

Copyright Owner

Cherng-Tarng (Tony) Lin

Language

en

Proquest ID

AAI8323297

File Format

application/pdf

File Size

74 pages

Share

COinS