2016 IEEE Symposium Series on Computational Intelligence (SSCI) 2016
DOI: 10.1109/ssci.2016.7850220
|View full text |Cite
|
Sign up to set email alerts
|

Connections of reference vectors and different types of preference information in interactive multiobjective evolutionary algorithms

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
39
0

Year Published

2019
2019
2021
2021

Publication Types

Select...
5
1

Relationship

5
1

Authors

Journals

citations
Cited by 26 publications
(39 citation statements)
references
References 39 publications
0
39
0
Order By: Relevance
“…Interactive versions of the two a posteriori algorithms (iRVEA and iNSGA-III) were also implemented and included in this study. The details of iRVEA can be found in [12] and iNSGA-III was implemented in a similar manner. RVEA and NSGA-III were chosen for this study as they have been shown to work well in problems with k > 2 [4,5].…”
Section: Methodsmentioning
confidence: 99%
See 1 more Smart Citation
“…Interactive versions of the two a posteriori algorithms (iRVEA and iNSGA-III) were also implemented and included in this study. The details of iRVEA can be found in [12] and iNSGA-III was implemented in a similar manner. RVEA and NSGA-III were chosen for this study as they have been shown to work well in problems with k > 2 [4,5].…”
Section: Methodsmentioning
confidence: 99%
“…One of the ways to incorporate a DM's preferences in decomposition-based EAs is to manipulate the spread of the RVs to account for the preferences [4,15]. In many such methods, the DM is required to provide their preferences in the form of a reference point in the objective space [12,26,27]. The components of a reference point are desirable values of each objective function, which may or may not be achievable.…”
Section: Evolutionary Algorithmsmentioning
confidence: 99%
“…Furthermore, human DM(s) took part in the remaining 16 experiments. A single human DM participated in [13], [15], [20], [24], [38], [40], [43], [46], [50], [56]- [58], [60], [61], [69], except for one experiment [47], where a group of DMs was involved in a group decision making context. In fact, in all these experiments, the author(s) acted as DMs establishing some assumptions about possible preference information.…”
Section: Type Of Dmsmentioning
confidence: 99%
“…On the other hand, by simulating a human DM's responses, author(s) could sometimes test different preference types in the same experiment. The most widely used and tested preference type was specifying a reference point ( [24], [38], [56], [57], [61], [69]). In [24], other types of preference information were tested: selecting preferred solutions, specifying non-preferred solutions and specifying preferred ranges for objective function values in addition to the reference point.…”
Section: Type Of Preferencesmentioning
confidence: 99%
“…The nondominated solutions violating some reservation level are penalized and those closer to the aspiration levels according to the Euclidean distance are rewarded. Furthermore, [9] suggests an interactive EMO method based on RVEA [2], where the DM can specify, if desired, preferred ranges for the objectives (i.e. aspiration and reservation levels), which are used to adjust the set of reference vectors in RVEA.…”
Section: Introductionmentioning
confidence: 99%