A Geometrically-Based Method for Efficient Many-Objective Decision-Making

Abstract

Practitioners of the systems engineering discipline are increasingly asked to make decisions from large sets of alternative solutions while considering the conflicting interests of diverse system stakeholders. Formulated as many-alternative, many-objective optimization problems, a posteriori methods are often applied to these scenarios to determine the solution alternatives that are objectively best performing according to the diverse stakeholder preferences. Frequently operating under computational and temporal constraints, decision-makers are often forced to consider fewer alternatives or incorporate a smaller number of stakeholder preferences due to the inefficiencies of current a posteriori methods. Utilizing a geometric comparison to the ideal point of the solution-space, a method is proposed that seeks to reduce the computational and temporal expense of determining the set of objectively superior solutions. In a numerical comparison to current methods, the proposed was shown to exhibit improved efficiency across a range of many-objective test-classes. Equipped with these efficiency allowances, systems engineering decision-makers can consider more alternatives and a greater number of stakeholder preferences without violating computational or time restrictions. These liberties enable a more complete and tailored search of the solution-space, permitting the identification of more thoroughly vetted and scrutinized objectively superior engineering solutions.

Department(s)

Engineering Management and Systems Engineering

Keywords and Phrases

Geometric presort; Ideal point comparison; Many-objective decision-making; Pareto efficient set; Pareto frontier

International Standard Book Number (ISBN)

978-099751956-3

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2024 American Society of Engineering Management, All rights reserved.

Publication Date

01 Jan 2019

This document is currently not available here.

Share

 
COinS