- author = "Una-May O'Reilly and Trent McConaghy and Rick Riolo",
- title = "GPTP 2009: An Example of Evolvability",
- booktitle = "Genetic Programming Theory and Practice {VII}",
- year = "2009",
- editor = "Rick L. Riolo and Una-May O'Reilly and Trent McConaghy",
- series = "Genetic and Evolutionary Computation",
- address = "Ann Arbor",
- month = "14-16 " # may,
- publisher = "Springer",
- chapter = "1",
- pages = "1--18",
- keywords = "genetic algorithms, genetic programming",
- isbn13 = "978-1-4419-1653-2",
-
DOI = "
doi:10.1007/978-1-4419-1626-6_1",
-
URL = "
http://trent.st/content/2009-GPTP-aims.pdf",
- size = "18 pages",
-
abstract = "This introductory chapter gives a brief description of
genetic programming (GP); summarises current GP
algorithm aims, issues, and progress; and finally
reviews the contributions of this volume, which were
presented at the GP Theory and Practice (GPTP) 2009
workshop.
This year marks a transition wherein the aims of GP algorithms-reasonable resource usage, high quality results, and reliable convergence-are being consistently realised on an impressive variety of real-world applications by skilled practitioners in the field. These aims have been realized due to GP researchers' growing collective understanding of the nature of GP problems which require search across spaces which are massive, multi-modal, and with poor locality, and how that relates to long-discussed GP issues such as bloat and premature convergence. New ways to use and extend GP for improved computational resource usage, quality of results, and reliability are appearing and gaining momentum. These include: reduced resource usage via rationally designed search spaces and fitness functions for specific applications such as induction of implicit functions or modelling stochastic processes arising from bio-networks; improved quality of results by explicitly targeting the interpretability or trustworthiness of the final results; and heightened reliability via consistently introducing new genetic material in a structured manner or via coevolution and teaming. These new developments highlight that GP's challenges have changed from simply making it work on smaller problems, to consistently and rapidly getting high-quality results on large real-world problems. GPTP 2009 was a forum to advance GP's state of the art and its contributions demonstrate how these aims can be met on a variety of difficult problems.",
- notes = "part of \cite{Riolo:2009:GPTP}",
