Paper Info
Title
A Simple Numerical Scheme For Generation Of Weighting Factors For Multiobjective Optimisation
Abstract
Optimisers for multi- or many-objective optimisation problems can be categorised as scalarisation and meta-heuristic approaches. Many of the approaches from both groups require to use a set of weighting vectors, which are expected to be as evenly distributed as possible. The current practice employs the normal-boundary intersection (NBI) method which has one disadvantage in that the number of sampling points must be the number ofk-combination. This work proposes a numerical scheme called clustering-based hyperplane sampling (CBHS) to deal with such a weak point. The method is based on random sampling on a hyperplane and clustering. The classical NBI method and some of its extended versions are used to examine the performance of the proposed algorithm. The comparative results reveal that CBHS is the best performer with using longer computing time. Moreover, its real advantage is the capability of generating a set of weighting vectors with any sample size.
Year
DOI
Venue
2021
10.1007/s00500-020-05249-0
SOFT COMPUTING
Keywords
DocType
Volume
Weighting factors, Scalarisation, Many-objective optimisation, Multiobjective meta-heuristics
Journal
25
Issue
ISSN
Citations 
2
1432-7643
0
PageRank 
References 
Authors
0.34
0
2
Name
Order
Citations
PageRank
Sujin Bureerat100.34
Nantiwat Pholdee2398.08