Detail publikace

Towards Compositional Coevolution in Evolutionary Circuit Design

Originální název

Towards Compositional Coevolution in Evolutionary Circuit Design

Anglický název

Towards Compositional Coevolution in Evolutionary Circuit Design

Jazyk

en

Originální abstrakt

A divide and conquer approach is one of the methods introduced to get over the scalability problem of the evolutionary circuit design. A complex circuit is decomposed into modules which are evolved separately and without any interaction. The benefits are in reducing the search space and accelerating the evaluation of candidate circuits. In this paper, the evolution of non-interacting modules is replaced by a coevolutionary algorithm, in which the fitness of a module depends on fitness values of other modules, i.e. the modules are adapted to work together. The proposed method is embedded into Cartesian genetic programming (CGP). The coevolutionary approach was evaluated in the design of a switching image filter which was decomposed into the filtering module and detector module. The filters evolved using the proposed coevolutionary method show a higher quality of filtering in comparison with filters utilizing independently evolved modules. Furthermore, the whole design process was accelerated 1.31 times in comparison with the standard CGP.

Anglický abstrakt

A divide and conquer approach is one of the methods introduced to get over the scalability problem of the evolutionary circuit design. A complex circuit is decomposed into modules which are evolved separately and without any interaction. The benefits are in reducing the search space and accelerating the evaluation of candidate circuits. In this paper, the evolution of non-interacting modules is replaced by a coevolutionary algorithm, in which the fitness of a module depends on fitness values of other modules, i.e. the modules are adapted to work together. The proposed method is embedded into Cartesian genetic programming (CGP). The coevolutionary approach was evaluated in the design of a switching image filter which was decomposed into the filtering module and detector module. The filters evolved using the proposed coevolutionary method show a higher quality of filtering in comparison with filters utilizing independently evolved modules. Furthermore, the whole design process was accelerated 1.31 times in comparison with the standard CGP.

BibTex


@inproceedings{BUT111610,
  author="Michaela {Drahošová} and Gergely {Komjáthy} and Lukáš {Sekanina}",
  title="Towards Compositional Coevolution in Evolutionary Circuit Design",
  annote="A divide and conquer approach is one of the methods introduced to get over the
scalability problem of the evolutionary circuit design. A complex circuit is
decomposed into modules which are evolved separately and without any interaction.
The benefits are in reducing the search space and accelerating the evaluation of
candidate circuits. In this paper, the evolution of non-interacting modules is
replaced by a coevolutionary algorithm, in which the fitness of a module depends
on fitness values of other modules, i.e. the modules are adapted to work
together. The proposed method is embedded into Cartesian genetic programming
(CGP). The coevolutionary approach was evaluated in the design of a switching
image filter which was decomposed into the filtering module and detector module.
The filters evolved using the proposed coevolutionary method show a higher
quality of filtering in comparison with filters utilizing independently evolved
modules. Furthermore, the whole design process was accelerated 1.31 times in
comparison with the standard CGP.",
  address="Institute of Electrical and Electronics Engineers",
  booktitle="2014 IEEE International Conference on Evolvable Systems Proceedings",
  chapter="111610",
  doi="10.1109/ICES.2014.7008735",
  edition="NEUVEDEN",
  howpublished="electronic, physical medium",
  institution="Institute of Electrical and Electronics Engineers",
  year="2014",
  month="december",
  pages="157--164",
  publisher="Institute of Electrical and Electronics Engineers",
  type="conference paper"
}