- author = "Jan Drchal and Miroslav Snorek",
- title = "Distance measures for HyperGP with fitness sharing",
- booktitle = "GECCO '12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference",
- year = "2012",
- editor = "Terry Soule and Anne Auger and Jason Moore and David Pelta and Christine Solnon and Mike Preuss and Alan Dorin and Yew-Soon Ong and Christian Blum and Dario Landa Silva and Frank Neumann and Tina Yu and Aniko Ekart and Will Browne and Tim Kovacs and Man-Leung Wong and Clara Pizzuti and Jon Rowe and Tobias Friedrich and Giovanni Squillero and Nicolas Bredeche and Stephen L. Smith and Alison Motsinger-Reif and Jose Lozano and Martin Pelikan and Silja Meyer-Nienberg and Christian Igel and Greg Hornby and Rene Doursat and Steve Gustafson and Gustavo Olague and Shin Yoo and John Clark and Gabriela Ochoa and Gisele Pappa and Fernando Lobo and Daniel Tauritz and Jurgen Branke and Kalyanmoy Deb",
- isbn13 = "978-1-4503-1177-9",
- pages = "545--552",
- keywords = "genetic algorithms, genetic programming, generative and developmental systems",
- month = "7-11 " # jul,
- organisation = "SIGEVO",
- address = "Philadelphia, Pennsylvania, USA",
-
DOI = "
doi:10.1145/2330163.2330241",
- publisher = "ACM",
- publisher_address = "New York, NY, USA",
-
abstract = "In this paper we propose a new algorithm called
HyperGPEFS (HyperGP with Explicit Fitness Sharing). It
is based on a HyperNEAT, which is a well-established
evolutionary method employing indirect encoding of
artificial neural networks. Indirect encoding in
HyperNEAT is realized via special function called
Compositional and Pattern Producing Network (CPPN),
able to describe a neural network of arbitrary size.
CPPNs are represented by network structures, which are
evolved by means of a slightly modified version of
another, well-known algorithm NEAT (NeuroEvolution of
Augmenting Topologies). HyperGP is a variant of
HyperNEAT, where the CPPNs are optimized by Genetic
Programming (GP). Published results reported promising
improvement in the speed of convergence.
Our approach further extends HyperGP by using fitness sharing to promote a diversity of a population. Here, we thoroughly compare all three algorithms on six different tasks. Fitness sharing demands a definition of a tree distance measure. Among other five, we propose a generalized distance measure which, in conjunction with HyperGPEFS, significantly outperforms HyperNEAT and HyperGP on all, but one testing problems. Although this paper focuses on indirect encoding, the proposed distance measures are generally applicable.",
- notes = "Also known as \cite{2330241} GECCO-2012 A joint meeting of the twenty first international conference on genetic algorithms (ICGA-2012) and the seventeenth annual genetic programming conference (GP-2012)",
