Home
Up
About
Contents
Feedback
Search

 

General Information Program Contact

 

Program

 

(Max,+)-Stream
Contact Stream Chairpersons

 

International Federation of Operational Research Societies – XVIth Triennial Conference
 – Edinburgh, Scotland, July 8 – 12, 2002

 

Preliminary Conference Schedule
(Max,+)-Approach to Dynamic Event Systems

 

Session #1 - MB15
Semi-Plenary Address
The (Max,+) Algebra: A new Approach to Performance Evaluation of DES

Speaker
Bernd Heidergott
Technische Universiteit Delft, Netherlands
icon_email1_bw.gif (950 Byte) Heidergott@eurandom.tue.nl

Abstract
This tutorial provides an introduction to the (max,+) algebra with a view towards performance analysis and control of DES. First, we introduce the basic concepts of the (max,+) algebra. In a second part, performance analysis and control of deterministic DESs is addressed. Our main example will be public transportation networks. We will discuss the application of eigenvalues and eigenvectors in (max,+) algebra to timetable design and robustness analysis of timetables and present a simple scheme for calculating propagation of delays. Eventually, we discuss applications of (max,+) algebra to ergodic theory and stability analysis of stochastic DES.

 

Session #2 - MC5
Tutorial: Measurement and Decisions - Theory, Tools, and Applications

Presenter
Jonathan Barzilai, Dalhousie University,
icon_email1_bw.gif (950 Byte) Jonathan.Barzilai@dal.ca

Abstract
Measurement plays a central role in science. We will discuss problems with classical mathematical models of measurement and present a new, simple and powerful measurement theory. The new theory has far-reaching theoretical and practical implications. For example, we will report on a recent case where the validity of a scoring methodology used to evaluate proposals in a large defense project has been the subject of litigation. Based on this theory, we have developed a new, intuitive, flexible, easy-to-use but powerful decision methodology - Preference Function Modeling (PFM). A software package implementing PFM will be demonstrated.

Author / Speaker
Jonathan Barzilai
Affiliation
Dept. of Industrial Engineering, Dalhousie University
Address
740 Saint-Maurice, Suite 410, Montreal, Quebec H3C 1L5, Canada, icon_telephone_bw.gif (951 Byte) +1 (514) 954-3665, icon_facsimile_bw.gif (893 Byte) +1 (514) 954-9739, icon_email1_bw.gif (950 Byte) Jonathan.Barzilai@dal.ca  icon_web_bw.gif (951 Byte) www.ScientificDecisions.com

 

 

Session #2 - FC6
Preference Measurement - Utility, AHP and PFM

Chairperson
Cathal M. Brugha, University College Dublin
icon_email1_bw.gif (950 Byte) cathal.brugha@ucd.ie

 

PFM - A Powerful Measurement, Decision and Selection Methodology

Abstract
Preference Function Modeling (PFM) is a new methodology. It is founded on rigorous mathematical foundations and is powerful, easy to use, and more flexible than the AHP. We will review the theory, tools, and the practical application of this methodology.

Author / Speaker
Jonathan Barzilai
Affiliation
Dept. of Industrial Engineering, Dalhousie University
Address
740 Saint-Maurice, Suite 410, Montreal, Quebec H3C 1L5, Canada, icon_telephone_bw.gif (951 Byte) +1 (514) 954-3665, icon_facsimile_bw.gif (893 Byte) +1 (514) 954-9739, icon_email1_bw.gif (950 Byte) Jonathan.Barzilai@dal.ca  icon_web_bw.gif (951 Byte) www.ScientificDecisions.com

 

Classification of Measurement Models - Implications for Utility Theory

Abstract
We classify all measurement models as weak, proper, or strong. This classification has major implications for measurement of subjective variables and, in particular, utility. We will establish connections to one-dimensional measurement, coordinate functions, pair wise comparisons, etc. We will highlight some of the consequences of these results for utility theory.

Author / Speaker
Jonathan Barzilai
Affiliation
Dept. of Industrial Engineering, Dalhousie University
Address
740 Saint-Maurice, Suite 410, Montreal, Quebec H3C 1L5, Canada, icon_telephone_bw.gif (951 Byte) +1 (514) 954-3665, icon_facsimile_bw.gif (893 Byte) +1 (514) 954-9739, icon_email1_bw.gif (950 Byte) Jonathan.Barzilai@dal.ca  icon_web_bw.gif (951 Byte) www.ScientificDecisions.com

 

The AHP is Not a Valid Methodology

Abstract
The AHP is a variant of Miller's Hierarchical Process. It violates the fundamental principles of the theory of measurement and is not a valid methodology. Earlier analysis of the AHP by Belton & Gear mis-identified the problems and proposed a correction that does not address the underlying errors. Dyer's analysis too is limited to the symptoms of these errors. We will review errors related to scale type; measurement units; the "relative importance" interpretation; hierarchical decomposition; normalization; eigenvector, etc.

Author / Speaker
Jonathan Barzilai
Affiliation
Dept. of Industrial Engineering, Dalhousie University
Address
740 Saint-Maurice, Suite 410, Montreal, Quebec H3C 1L5, Canada, icon_telephone_bw.gif (951 Byte) +1 (514) 954-3665, icon_facsimile_bw.gif (893 Byte) +1 (514) 954-9739, icon_email1_bw.gif (950 Byte) Jonathan.Barzilai@dal.ca  icon_web_bw.gif (951 Byte) www.ScientificDecisions.com

 

 

 

 

 

 

 

 

 


Please direct questions or comments regarding proper functioning of this web site to webmaster@ceci.com

Copyright © 1999 - 2010 Choice Engineering Consultants International   Imprint & Legal Notice
Last reviewed on April 16, 2010