ANNIE 2003 Schedule
 
Back to ANNIE 2003
 


LUNCHEON PLENARY
 



Wednesday, Noon - 1:30 p.m.
 

(JW's Restaurant)

Hierarchical Bayesian Optimization Algorithm: Toward a New Generation of Evolutionary Algorithms

Dr. Martin Pelikan,
University of Missouri, St. Louis, MO  

Overview: A black-box optimization problem may be defined by the set of all potential solutions and a procedure for evaluating competing solutions. The separation of problem specifics from an optimizer enables a straightforward formulation of many different real-world problems within this framework. That is why the design of robust and scalable black-box optimization techniques has become one of the most important challenges in computational optimization. This talk focuses on an advanced black-box optimization algorithm called the hierarchical Bayesian optimization algorithm (hBOA), which derives inspiration from evolutionary computation on one hand and machine learning on the other. The talk motivates the algorithm, describes its basic procedure, and presents interesting results of its application to challenging real-world problems. The talk also lists important directions for future research in this area.

Biography: Martin Pelikan received his Ph.D. from the Department of Computer Science at the University of Illinois at Urbana-Champaign in 2002. Pelikan worked at some of the best institutes in the field of evolutionary computation, including the Illinois Genetic Algorithms Laboratory (IlliGAL) at the University of Illinois at Urbana-Champaign, Adaptive Systems group at the German National Center for Information Technology at Sankt Augustin, and the Swiss Federal Institute of Technology (ETH) Zurich. He pioneered one of the most promising areas of evolutionary computation, which combines evolutionary computation and machine learning to provide a robust and scalable solution to broad classes of black-box optimization problems without the need for much input from the user. Pelikan invented two of the most advanced evolutionary algorithms based on this principle: the Bayesian optimization algorithm (BOA) and the hierarchical BOA (hBOA).