Exportar Publicação

Mariano, P., Christensen, A. L. & Gomes, J. (2014). Novelty search in competitive coevolution. In Thomas Bartz-Beielstein, Jürgen Branke, Bogdan Filipič, Jim Smith (Ed.), Parallel Problem Solving from Nature -- PPSN XIII, Conference Proceedings. Ljubljana: Springer.

P. L. Mariano et al.,  "Novelty search in competitive coevolution", in Parallel Problem Solving from Nature -- PPSN XIII, Conf. Proc., Thomas Bartz-Beielstein, Jürgen Branke, Bogdan Filipič, Jim Smith, Ed., Ljubljana, Springer, 2014, vol. 8672

@inproceedings{mariano2014_1730780502850,
	author = "Mariano, P. and Christensen, A. L. and Gomes, J.",
	title = "Novelty search in competitive coevolution",
	booktitle = "Parallel Problem Solving from Nature -- PPSN XIII, Conference Proceedings",
	year = "2014",
	editor = "Thomas Bartz-Beielstein, Jürgen Branke, Bogdan Filipič, Jim Smith",
	volume = "8672",
	number = "",
	series = "",
	doi = "10.1007/978-3-319-10762-2_23",
	publisher = "Springer",
	address = "Ljubljana",
	organization = "",
	url = "https://link.springer.com/book/10.1007/978-3-319-10762-2?page=2&oscar-books=true#toc"
}

TY  - CPAPER
TI  - Novelty search in competitive coevolution
T2  - Parallel Problem Solving from Nature -- PPSN XIII, Conference Proceedings
VL  - 8672
AU  - Mariano, P.
AU  - Christensen, A. L.
AU  - Gomes, J.
PY  - 2014
DO  - 10.1007/978-3-319-10762-2_23
CY  - Ljubljana
UR  - https://link.springer.com/book/10.1007/978-3-319-10762-2?page=2&oscar-books=true#toc
AB  - One of the main motivations for the use of competitive coevolution systems is their ability to capitalise on arms races between competing species to evolve increasingly sophisticated solutions. Such arms races can, however, be hard to sustain, and it has been shown that the competing species often converge prematurely to certain classes of behaviours. In this paper, we investigate if and how novelty search, an evolutionary technique driven by behavioural novelty, can overcome convergence in coevolution. We propose three methods for applying novelty search to coevolutionary systems with two species: (i) score both populations according to behavioural novelty; (ii) score one population according to novelty, and the other according to fitness; and (iii) score both populations with a combination of novelty and fitness. We evaluate the methods in a predator-prey pursuit task. Our results show that novelty-based approaches can evolve a significantly more diverse set of solutions, when compared to traditional fitness-based coevolution.
ER  -