Flex-GP: Genetic Programming on the Cloud

Created by W.Langdon from gp-bibliography.bib Revision:1.4333

  author =       "Dylan Sherry and Kalyan Veeramachaneni and 
                 James McDermott and Una-May O'Reilly",
  title =        "Flex-GP: Genetic Programming on the Cloud",
  booktitle =    "Applications of Evolutionary Computing,
                 EvoApplications 2012: EvoCOMNET, EvoCOMPLEX, EvoFIN,
                 EvoGAMES, EvoHOT, EvoIASP, EvoNUM, EvoPAR, EvoRISK,
                 EvoSTIM, EvoSTOC",
  year =         "2011",
  month =        "11-13 " # apr,
  editor =       "Cecilia {Di Chio} and Alexandros Agapitos and 
                 Stefano Cagnoni and Carlos Cotta and F. {Fernandez de Vega} and 
                 Gianni A. {Di Caro} and Rolf Drechsler and 
                 Aniko Ekart and Anna I Esparcia-Alcazar and Muddassar Farooq and 
                 William B. Langdon and Juan J. Merelo and 
                 Mike Preuss and Hendrik Richter and Sara Silva and 
                 Anabela Simoes and Giovanni Squillero and Ernesto Tarantino and 
                 Andrea G. B. Tettamanzi and Julian Togelius and 
                 Neil Urquhart and A. Sima Uyar and Georgios N. Yannakakis",
  series =       "LNCS",
  volume =       "7248",
  publisher =    "Springer Verlag",
  address =      "Malaga, Spain",
  publisher_address = "Berlin",
  pages =        "477--486",
  organisation = "EvoStar",
  keywords =     "genetic algorithms, genetic programming",
  isbn13 =       "978-3-642-29177-7",
  DOI =          "doi:10.1007/978-3-642-29178-4_48",
  size =         "10 pages",
  abstract =     "We describe FlexGP, which we believe to be the first
                 large-scale genetic programming cloud computing system.
                 We took advantage of existing software and selected a
                 socket-based, client-server architecture and an
                 island-based distribution model. We developed core
                 components required for deployment on Amazon's Elastic
                 Compute Cloud. Scaling the system to hundreds of nodes
                 presented several unexpected challenges and required
                 the development of software for automatically managing
                 deployment, reporting, and error handling. The system's
                 performance was evaluated on two metrics, performance
                 and speed, on a difficult symbolic regression problem.
                 Our largest successful FlexGP runs reached 350 nodes
                 and taught us valuable lessons for the next phase of
  notes =        "ECJ java. Distributed Island model, each with a
                 population of 3000, max pop > 1million. (cf
                 \cite{langdon:2008:SC} ). non-toriodal four node
                 (square) grid used to pass 40 emigrants every 4th
                 generation. (Up to 100 generations). Amazon Linux AMI
                 micro instance insufficient, used small sized instances
                 instead. Amazon AMI boot takes variable time (up to
                 several minutes). Need to add 'controller' and
                 'LogServer' on top of ECJ and Amazon (Python Bota, also
                 twisted and nmap, all 3 open source).

                 EvoPAR Part of \cite{DiChio:2012:EvoApps}
                 EvoApplications2012 held in conjunction with
                 EuroGP2012, EvoCOP2012, EvoBio'2012 and EvoMusArt2012",

Genetic Programming entries for Dylan Sherry Kalyan Veeramachaneni James McDermott Una-May O'Reilly