1 / 7

Genetic Programming

Genetic Programming. Genetic programming (GP) is an automated method for creating a working computer program from a high-level problem statement of a problem.

herbst
Download Presentation

Genetic Programming

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Genetic Programming Genetic programming (GP) is an automated method for creating a working computer program from a high-level problem statement of a problem. The evolutionary search uses the Darwinian principle of natural selection (survival of the fittest) and analogs of various naturally occurring operations. Pyramid Search Method

  2. Wasted Computation Due To Early Convergence • Early Convergence : The optimal solutions can not be found for some runs even after millions of evaluations, thus enormous amount of computer resources are wasted for runs hunting around the suboptimal. • The graph demonstrates that many runs do not find a solution in a symbolic regression problem. One of big problems in genetic programming is the early convergence, which causes waste. Pyramid Search Method

  3. Avoiding Waste by Pyramid Method The pyramid method proposed in this paper saves. It smartly allows runs breaking through local optima to the solution. It is tested in the following four problems. MAX Prob. 5 Even Parity Symbolic Regression Object Detection Pyramid Search Method

  4. What is the Pyramid Method ? A number of populations are initialised and independently evolved for a number of generations at which point the worst performing populations are discarded. This evolve/discard process is continued until the problem is solved or one population remains. pyramid(np,ps,pr,ng) np Number of populations ps Population Size pr Pruning ratio, Fraction of remaining populations to remove ng Number of Generations between prunings Initialise np populations of size ps popsleft = np Until problem solved or a single population remains do Evolve each population for ng generations Remove the round(pr*popsleft) least fit populations popsleft = popsleft-round(pr/popsleft) end do If problem not solved then continue evolving single population until problem solved or maximum evaluations is reached Pyramid Search Method

  5. Why the Pyramid Method Works ? • Kills bad runs earlier • Auto detects bad ones by comparison • Maintains divergence • Easy for parallel processing A variant of multiple runs It Saves Your Time, Resources, etc. Why not use it ? Pyramid Search Method

  6. Questions Pyramid Search Method

  7. REFERENCES • Thomas Loveard. Genetic programming with meta-search: Searching for a successful population within the classification domain. In Conor Ryan, Terence Soule, Maarten Keijzer, Edward Tsang, Riccardo Poli, and Ernesto Costa, editors, Proceedings of the 6th European Conference on GeneticProgramming (EuroGP 2003), pages 119–129, Berlin, 2003. Springer. • Andrew Innes, Vic Ciesielski, John Mamutil, and Sabu John. Landmark detection for cephalometric radiology images using genetic programming. International Journal of Knowledge Based Intelligent Engineering Systems, 7(3):164-171, July 2003. • W. B. Langdon and R. Poli. An analysis of the MAX problem in genetic programming. In John R. Koza, Kalyanmoy Deb, Marco Dorigo, David B. Fogel, Max Garzon, HitoshiIba, and Rick L. Riolo, editors, Genetic Programming 1997: Proceedings of the Second International Conference on Genetic Programming. Morgan Kaufmann, July 1997. • J. He and X. Yao, ``From an Individual to a Population: An Analysis of the First Hitting Time of Population-Based Evolutionary Algorithms,'' IEEE Transactions on Evolutionary Computation, 6(5):495-511, October 2002. • John R. Koza. Genetic Programming II: Automatic Discovery of Reusable Programs. MIT Press, 1994. • Emin Erkan Korkmaz and G¨okt¨urk ¨ Uc¸oluk. Controlled genetic programming search for solving deceptive problems. In Erick Cant´u-Paz, editor, Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002), pages 295–300, New York, NY, July 2002. AAAI. • Sean Luke. When short runs beat long runs. In Lee Spector et al., editor, Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001), pages 74–80, San Francisco, California, USA, 2001. Morgan Kaufmann. Pyramid Search Method

More Related