Doctoral Dissertations

On unified computational intelligence: neural networks, dynamic programming, and applications to economics and finance

Keywords and Phrases

Computational social science

Abstract

"This dissertation introduces the concept of unified computational intelligence and encompasses algorithm design, applications, theoretical developments, and the identification of new frontiers for multidisciplinary research. It presents a new way to look at unified learning systems and presents a novel Adaptive Resonance Theory-based unified learning architecture. An application to robotic security and sensor fusion is given in detail. Work on the theoretical developments of unified computational intelligence is also presented. A new theorem in the time scales calculus is proven and theorems on the Hamilton-Jacobi-Ballman equation of dynamic programming and the dynamic programming algorithm are presented, both for various general dynamic derivatives and for the case of the quantum calculus. Neural network learning is also given the time scales treatment, where the backpropagation algorithm is proven to hold in this new calculus as well as in its quantum calculus rendition. Additionally, the idea of an ordered derivative on time scales, a concept fundamental to the backpropagation algorithm, is defined. Finally, applications of computational intelligence in the emerging field of agent-based computational social science are discussed. Multiple published articles are presented outlining the uses of computational intelligence techniques in the increasingly relevant areas of finance and economics"--Abstract, page iii.

Advisor(s)

Wunsch, Donald C.

Committee Member(s)

Beetner, Daryl G.
Venayagamoorthy, Ganesh K.
Insall, Matt
Stanley, R. Joe

Department(s)

Electrical and Computer Engineering

Degree Name

Ph. D. in Computer Engineering

Sponsor(s)

21st Century Systems Inc.
Boston University Center for Neural and Adaptive Systems
Los Alamos National Laboratory
Mary K. Finley Missouri Endowment
National Science Foundation (U.S.)
Sandia Laboratories

Publisher

Missouri University of Science and Technology

Publication Date

Fall 2009

Pagination

x, 158 pages

Note about bibliography

Includes bibliographical references (pages 146-157).

Rights

© 2009 John Seiffertt, All rights reserved.

Document Type

Dissertation - Open Access

File Type

text

Language

English

Subject Headings

Computational intelligenceDynamic programmingMultiagent systemsNeural networks (Computer science)

Thesis Number

T 9566

Print OCLC #

620702539

This document is currently not available here.

Share

 
COinS