Home > Press > Illinois Researchers Break Billion Variable Optimization Barrier
A paper published today in the journal Complexity describes how a team of researchers in the Illinois Genetic Algorithms Laboratory (IlliGAL) at the University of Illinois at Urbana-Champaign (UIUC) has achieved efficient, scalable solutions on difficult optimization problems containing over a billion variables. The team led by noted researcher and author David E. Goldberg used specially programmed genetic algorithms (GAs)--search procedures based on natural selection and genetics--to achieve the feat, together with theories of scalability and implementation techniques developed at Illinois. Optimization uses mathematics and computation to find efficient, effective solutions to problems in science, technology, and commerce, and it is widely used in scheduling, engineering design, and business management. Procedures in common use today are limited to thousands, sometimes millions, of variables because the most powerful methods become prohibitively expensive as the size of the problem increases. The Illinois result proves that billion-variable problems can be solved effectively and practically on existing computers with known procedures.
Illinois Researchers Break Billion Variable Optimization Barrier
Urbana, IL | Posted on January 19th, 2007
The calculations were performed on subsets of the 1536-processor Turing cluster housed in UIUC's Computational Science and Engineering (CSE) program. CSE director, Michael Heath, greeted the accomplishment. "This is exactly the kind of paradigm-breaking computational result that we hoped to enable in creating the Turing cluster." UIUC material scientist, Duane Johnson suggested that the result "is a milestone in the developing world of nanotechnology, enabling the analysis and design of new molecules in ways that were not previously possible," and John Deere emerging technology guru Bill Fulkerson sees the results as heralding a new day of complex systems optimization more generally. "Gone are the days of using a toy genetic algorithm to solve a toy problem. With petascale computing and solvers like this, complex systems optimization becomes possible."
Other team members included Kumara Sastry, a PhD candidate in Industrial and Enterprise Systems Engineering and Xavier Llora, a machine learning researcher at the National Center for Supercomputing Applications (NCSA). Although the team is pleased with the billion-variable result, it is not resting on its laurels. Sastry put it this way: "One reason this result is so interesting is because it is so general. With most optimization procedures you are stuck solving a limited class of problems. This result is immediately useful to a broad array of problems, and existing theory and technique tells us how to speed results on larger, harder problems that would otherwise be prohibitively expensive or impossible." Goldberg is excited by the array of existing application areas that can benefit from the result. "Genetic algorithms have been used regularly for two decades across the spectrum of human endeavor. Science, engineering, commerce, and even the humanities and the arts have already benefited from myriad applications of genetic algorithms. The billion-variable result can be put to use immediately across the panoply of existing and yet-to-be-imagined application domains." Complexity editor-in-chief, Alfred Hübler welcomed the research as "spectacular." "Goldberg's team has achieved something special. This result advances complexity science and technology immediately and noticeably."
The work was sponsored by the Air Force Office of Scientific Research and the NSF-sponsored Materials Computation Center (MCC) at the UIUC.
The article is available at http://www3.interscience.wiley.com/cgi-bin/jhome/38804 .
About University of Illinois
At Illinois, research shapes the campus identity, stimulates classroom instruction and serves as a springboard for public engagement activities throughout the world. Opportunities abound for graduate students to develop independent projects and launch their own careers as researchers while working alongside faculty and assisting in their research. Illinois continues its long tradition of groundbreaking accomplishments with remarkable new discoveries and achievements that inspire and enrich the lives of people around the world.
For more information, please click here
University of Illinois
Copyright © PRWeb
If you have a comment, please Contact
Issuers of news releases, not 7th Wave, Inc. or Nanotechnology Now, are solely responsible for the accuracy of the content.
Preparing for Nano
Durnham University's DEEPEN project comes to a close September 26th, 2012
Technical Seminar at ANFoS 2012 August 22nd, 2012
Nanotechnology shows we can innovate without economic growth April 12th, 2012
Thailand to host NanoThailand 2012 December 18th, 2011
Rice's Naomi Halas to direct Smalley Institute: Optics pioneer will lead Rice's multidisciplinary science institute January 15th, 2015
SUNY Board Appoints Dr. Alain Kaloyeros as Founding President of SUNY Polytechnic Institute January 13th, 2015
CNSE's Smart System Technology & Commercialization Center Successfully Recertifies as ISO 9001:2008 January 12th, 2015
SUNY Poly Now Accepting Applications to the Colleges of Nanoscale Science and Engineering for Fall 2015: Full Scholarships Available to Incoming CNSE Students January 7th, 2015
Nanoparticles Increase Durability of Concrete Decorations in Cold Areas January 26th, 2015
Iranian Researchers Boost Solar Cells Efficiency Using Anti-Aggregates January 26th, 2015
Detection of Heavy Metals in Samples with Naked Eye January 26th, 2015
Engineering self-assembling amyloid fibers January 26th, 2015
OCSiAl supports NanoART Imagery Contest January 23rd, 2015
EnvisioNano: An image contest hosted by the National Nanotechnology Initiative (NNI) January 22nd, 2015
Oxford Instruments Asylum Research Announces AFM Image Contest Winners January 11th, 2015
Longhorn beetle inspires ink to fight counterfeiting November 5th, 2014