About the School
Contact and People
Future Undergraduate Students
Prospective Postgraduates
Current Students
Current Postgraduates
Research
IT News
Awards
Industry Links and Prizes
School and IT Information
Other
Internal Information
Research Seminar
Direction Matters in High Dimensional OptimisationDr Cara MacNish, UWA 11am, 11th July AbstractDirectional biases are evident in many benchmarking problems for real- valued global optimisation, as well as many of the evolutionary and allied algorithms that have been proposed for solving them. It has been shown that directional biases make some kinds of problems easier to solve for similarly biased algorithms, which can give a misleading view of algorithm performance in the search for general purpose optimisation algorithms.This talk will present results of a study the effects of directional bias for high-dimensional optimisation problems. We show that the impact of directional bias is magnified as dimension increases, and can in some cases lead to differences in performance of many orders of magnitude. We present a new version of the classical evolutionary programming algorithm, which we call unbiased evolutionary programming (UEP), and show that it has markedly improved performance for high-dimensional optimisation. The above results will be preceded by a brief general introduction to optimisation, its relationship to modelling, control and learning, and the search for the holy grail. |
