Using Co-solvability to Model and Exploit Synergetic Effects in Evolution

We introduce, analyze, and experimentally verify the concept of co-solvability, meant as the ability of a solution maintained by an evolutionary run to solve (correctly process) a pair of fitness cases (tests). The method based on this concept can be considered as a second-order implicit fitness sharing, where solutions compete for the rewards granted for solving pairs of tests, rather than single tests. We prove that co-solvability fitness function is by definition synergistic and imposes selection pressure which is qualitatively different from that induced by standard fitness function or implicit fitness sharing. The results of experimental verification on eight genetic programming tasks demonstrate that evolutionary runs driven by the proposed fitness function usually converge faster to global optima than other methods.


Editor(s):
Schaefer, Robert
Cotta, Carlos
Kolodziej, Joanna
Rudolph, Gunter
Published in:
Parallel Problem Solving from Nature -- PPSN XI, 492--501
Presented at:
11th International Conference on Parallel Problem Solving From Nature, Krakow, Poland, September 11-15, 2010
Year:
2010
Publisher:
Heidelberg, Springer
ISBN:
978-3-642-15870-4
Keywords:
Laboratories:




 Record created 2010-09-01, last modified 2018-01-28

External link:
Download fulltext
n/a
Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)