Seminar 2016-11-30

Date: 2016-11-30

Time: 13:15-14:15

Speaker: Professor Peter J. Stuckey, University of Melbourne

Abstract:

Solving combinatorial optimization problems is hard. There are many techniques which we need to combine to successfully solve these
problems: including inference, search, relaxation, and decomposition. Any method to solve combinatorial optimization problems must tradeoff how much time to spend in search, and how much time to spend on methods to reduce the search. In this talk I will propose that we should spend the least amount of time possible on methods to reduce search, until there is evidence that it will be useful. I will demonstrate how this "laziness" principle can be highly effective in solving.

Bio:

Professor Peter J. Stuckey is a Professor in the Department of Computing and Information Systems in the University of Melbourne, and project leader in the Data61 CSIRO laboratory. Peter Stuckey is a pioneer in constraint programming, the science of modelling and solving complex combinatorial problems. His research interests include: discrete optimization; programming languages, in particular declarative programing languages; constraint solving  algorithms; bioinformatics; and constraint-based graphics. He enjoys problem solving in any area, having publications in e.g. databases, timetabling, and system security, and working with companies such as Oracle and Rio Tinto on problems that interest them.