Detail publikace

Evolutionary Multiobjective Bayesian Optimization Algorithm:Experimental Study

Originální název

Evolutionary Multiobjective Bayesian Optimization Algorithm:Experimental Study

Anglický název

Evolutionary Multiobjective Bayesian Optimization Algorithm:Experimental Study

Jazyk

en

Originální abstrakt

This paper deals with the utilizing of the Bayesian optimization algorithm (BOA) for multiobjective optimization of hypergraph partitioning. The main attention is focused on the incorporation of the Pareto optimality concept. We have modified the standard algorithm BOA for one criterion optimization according to well known niching techniques to find the Pareto optimal set. This approach was compared with standard weighting techniques and the single optimization approach with the constraint. The experiments are focused mainly on the bi-objective optimization because of the visualization simplicity.

Anglický abstrakt

This paper deals with the utilizing of the Bayesian optimization algorithm (BOA) for multiobjective optimization of hypergraph partitioning. The main attention is focused on the incorporation of the Pareto optimality concept. We have modified the standard algorithm BOA for one criterion optimization according to well known niching techniques to find the Pareto optimal set. This approach was compared with standard weighting techniques and the single optimization approach with the constraint. The experiments are focused mainly on the bi-objective optimization because of the visualization simplicity.

BibTex


@inproceedings{BUT5431,
  author="Josef {Schwarz} and Jiří {Očenášek}",
  title="Evolutionary Multiobjective Bayesian Optimization Algorithm:Experimental Study",
  annote="This paper deals with the utilizing of the Bayesian optimization algorithm (BOA) for multiobjective optimization of hypergraph partitioning. The main attention is focused on the incorporation of the Pareto optimality concept. We have modified the standard algorithm BOA for one criterion optimization according to well known niching techniques to find the Pareto optimal set. This approach was compared with standard weighting techniques and the single optimization approach with the constraint. The experiments are focused mainly on the bi-objective optimization because of the visualization simplicity.",
  booktitle="Proceedings of the 35th Spring International Conference MOSIS'01, Vol. 1",
  chapter="5431",
  year="2001",
  month="january",
  pages="101--108",
  type="conference paper"
}