The hypervolume indicator as a performance measure in dynamic optimization

Sabrina Oliveira*, Elizabeth F. Wanner, Sérgio R. de Souza, Leonardo C.T. Bezerra, Thomas Stützle

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In many real world problems the quality of solutions needs to be evaluated at least according to a bi-objective non-dominated front, where the goal is to optimize solution quality using as little computational resources as possible. This is even more important in the context of dynamic optimization, where quickly addressing problem changes is critical. In this work, we relate approaches for the performance assessment of dynamic optimization algorithms to the existing literature on bi-objective optimization. In particular, we introduce and investigate the use of the hypervolume indicator to compare the performance of algorithms applied to dynamic optimization problems. As a case study, we compare variants of a state-of-the-art dynamic ant colony algorithm on the traveling salesman problem with dynamic demands (DDTSP). Results demonstrate that our proposed approach accurately measures the desirable characteristics one expects from a dynamic optimizer and provides more insights than existing alternatives.

Original languageEnglish
Title of host publicationEvolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings
EditorsSanaz Mostaghim, Kaisa Miettinen, Kalyanmoy Deb, Erik Goodman, Carlos A. Coello Coello, Kathrin Klamroth, Patrick Reed
PublisherSpringer
Pages319-331
Number of pages13
ISBN (Print)9783030125974
DOIs
Publication statusPublished - 3 Feb 2019
Event10th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2019 - East Lansing, United States
Duration: 10 Mar 201913 Mar 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11411 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2019
CountryUnited States
CityEast Lansing
Period10/03/1913/03/19

Fingerprint

Dynamic Optimization
Performance Measures
Dynamic Optimization Problems
Ant Colony Algorithm
Performance Assessment
Dynamic Algorithms
Travelling salesman problems
Optimization Algorithm
Optimise
Resources
Traveling salesman problem
Optimization
Alternatives
Demonstrate

Keywords

  • Dynamic optimization
  • Multi-objective optimization
  • Performance assessment

Cite this

Oliveira, S., Wanner, E. F., de Souza, S. R., Bezerra, L. C. T., & Stützle, T. (2019). The hypervolume indicator as a performance measure in dynamic optimization. In S. Mostaghim, K. Miettinen, K. Deb, E. Goodman, C. A. Coello Coello, K. Klamroth, & P. Reed (Eds.), Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings (pp. 319-331). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11411 LNCS). Springer. https://doi.org/10.1007/978-3-030-12598-1_26
Oliveira, Sabrina ; Wanner, Elizabeth F. ; de Souza, Sérgio R. ; Bezerra, Leonardo C.T. ; Stützle, Thomas. / The hypervolume indicator as a performance measure in dynamic optimization. Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings. editor / Sanaz Mostaghim ; Kaisa Miettinen ; Kalyanmoy Deb ; Erik Goodman ; Carlos A. Coello Coello ; Kathrin Klamroth ; Patrick Reed. Springer, 2019. pp. 319-331 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{97355c2047954da289c697ab803cfce4,
title = "The hypervolume indicator as a performance measure in dynamic optimization",
abstract = "In many real world problems the quality of solutions needs to be evaluated at least according to a bi-objective non-dominated front, where the goal is to optimize solution quality using as little computational resources as possible. This is even more important in the context of dynamic optimization, where quickly addressing problem changes is critical. In this work, we relate approaches for the performance assessment of dynamic optimization algorithms to the existing literature on bi-objective optimization. In particular, we introduce and investigate the use of the hypervolume indicator to compare the performance of algorithms applied to dynamic optimization problems. As a case study, we compare variants of a state-of-the-art dynamic ant colony algorithm on the traveling salesman problem with dynamic demands (DDTSP). Results demonstrate that our proposed approach accurately measures the desirable characteristics one expects from a dynamic optimizer and provides more insights than existing alternatives.",
keywords = "Dynamic optimization, Multi-objective optimization, Performance assessment",
author = "Sabrina Oliveira and Wanner, {Elizabeth F.} and {de Souza}, {S{\'e}rgio R.} and Bezerra, {Leonardo C.T.} and Thomas St{\"u}tzle",
year = "2019",
month = "2",
day = "3",
doi = "10.1007/978-3-030-12598-1_26",
language = "English",
isbn = "9783030125974",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "319--331",
editor = "Sanaz Mostaghim and Kaisa Miettinen and Kalyanmoy Deb and Erik Goodman and {Coello Coello}, {Carlos A.} and Kathrin Klamroth and Patrick Reed",
booktitle = "Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings",
address = "Germany",

}

Oliveira, S, Wanner, EF, de Souza, SR, Bezerra, LCT & Stützle, T 2019, The hypervolume indicator as a performance measure in dynamic optimization. in S Mostaghim, K Miettinen, K Deb, E Goodman, CA Coello Coello, K Klamroth & P Reed (eds), Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11411 LNCS, Springer, pp. 319-331, 10th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2019, East Lansing, United States, 10/03/19. https://doi.org/10.1007/978-3-030-12598-1_26

The hypervolume indicator as a performance measure in dynamic optimization. / Oliveira, Sabrina; Wanner, Elizabeth F.; de Souza, Sérgio R.; Bezerra, Leonardo C.T.; Stützle, Thomas.

Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings. ed. / Sanaz Mostaghim; Kaisa Miettinen; Kalyanmoy Deb; Erik Goodman; Carlos A. Coello Coello; Kathrin Klamroth; Patrick Reed. Springer, 2019. p. 319-331 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11411 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - The hypervolume indicator as a performance measure in dynamic optimization

AU - Oliveira, Sabrina

AU - Wanner, Elizabeth F.

AU - de Souza, Sérgio R.

AU - Bezerra, Leonardo C.T.

AU - Stützle, Thomas

PY - 2019/2/3

Y1 - 2019/2/3

N2 - In many real world problems the quality of solutions needs to be evaluated at least according to a bi-objective non-dominated front, where the goal is to optimize solution quality using as little computational resources as possible. This is even more important in the context of dynamic optimization, where quickly addressing problem changes is critical. In this work, we relate approaches for the performance assessment of dynamic optimization algorithms to the existing literature on bi-objective optimization. In particular, we introduce and investigate the use of the hypervolume indicator to compare the performance of algorithms applied to dynamic optimization problems. As a case study, we compare variants of a state-of-the-art dynamic ant colony algorithm on the traveling salesman problem with dynamic demands (DDTSP). Results demonstrate that our proposed approach accurately measures the desirable characteristics one expects from a dynamic optimizer and provides more insights than existing alternatives.

AB - In many real world problems the quality of solutions needs to be evaluated at least according to a bi-objective non-dominated front, where the goal is to optimize solution quality using as little computational resources as possible. This is even more important in the context of dynamic optimization, where quickly addressing problem changes is critical. In this work, we relate approaches for the performance assessment of dynamic optimization algorithms to the existing literature on bi-objective optimization. In particular, we introduce and investigate the use of the hypervolume indicator to compare the performance of algorithms applied to dynamic optimization problems. As a case study, we compare variants of a state-of-the-art dynamic ant colony algorithm on the traveling salesman problem with dynamic demands (DDTSP). Results demonstrate that our proposed approach accurately measures the desirable characteristics one expects from a dynamic optimizer and provides more insights than existing alternatives.

KW - Dynamic optimization

KW - Multi-objective optimization

KW - Performance assessment

UR - http://www.scopus.com/inward/record.url?scp=85063033985&partnerID=8YFLogxK

UR - https://link.springer.com/chapter/10.1007%2F978-3-030-12598-1_26

U2 - 10.1007/978-3-030-12598-1_26

DO - 10.1007/978-3-030-12598-1_26

M3 - Conference contribution

AN - SCOPUS:85063033985

SN - 9783030125974

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 319

EP - 331

BT - Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings

A2 - Mostaghim, Sanaz

A2 - Miettinen, Kaisa

A2 - Deb, Kalyanmoy

A2 - Goodman, Erik

A2 - Coello Coello, Carlos A.

A2 - Klamroth, Kathrin

A2 - Reed, Patrick

PB - Springer

ER -

Oliveira S, Wanner EF, de Souza SR, Bezerra LCT, Stützle T. The hypervolume indicator as a performance measure in dynamic optimization. In Mostaghim S, Miettinen K, Deb K, Goodman E, Coello Coello CA, Klamroth K, Reed P, editors, Evolutionary Multi-Criterion Optimization - 10th International Conference, EMO 2019, Proceedings. Springer. 2019. p. 319-331. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-12598-1_26