Principles of multiobjective optimization
Loading...
Authors
Rosenthal, Richard E.
Subjects
Optimization, Multiple Objectives, Multicriterion Decision Making
Advisors
Date of Issue
1984-08
Date
1984-08
Publisher
Monterey, California. Naval Postgraduate School
Language
en_US
Abstract
The purpose of this paper is to isolate and analyze the principal ideas of
multiobjective optimization. This is done without casting recriminatory aspersions
on single-objective optimization or championing any one multiobjective
technique. The paper first attempts to define the problem, and then discusses
the fundamental ideas, many of which stem from common sense. Each idea is examined
for strengths and weaknesses and two—efficiency and utility—are shown
worthy of extended consideration. In the light of this analysis some general
recommendations are made. Besides offering the simple advice of not dismissing single-objective optimization as a possible approach, we suggest that three
broad classes of multiobjective techniques are very promising in terms of reliably
and believably achieving a most preferred solution. These are: (i)
partial generation of the efficient set, a rubric we use for unifying a wide
spectrum of existing methods, (ii) explicit utility maximization, a much overlooked
approach combining multiattribute decision theory and mathematical programming,
and (iii) implicit utility maximization, the name we use for the
popular class of methods introduced by Geoffrion, Dyer and Feinberg and extended
significantly by many others.
Type
Technical Report
Description
Series/Report No
Department
Operations Research
Organization
Operations Research (OR)
Graduate School of Operational and Information Sciences (GSOIS)
Identifiers
NPS Report Number
NPS55-84-016
Sponsors
This work was supported by the Naval Postgraduate School Foundation
Research Program under contract with the National Research Council.
Funder
N0001484WR41001
Format
Citation
Distribution Statement
Approved for public release; distribution is unlimited.