This project funds the travel expenses of US based researchers to the tenth International conference on Monte Carlo and quasi-Monte Carlo methods, to be held in Sydney Australia in February 2012. The meeting will consider advanced methods for psuedo-random number generation, construction of low discrepancy point sets, and complexity of algorithms. There will be tutorial sessions on Monte Carlo, quasi-Monte Carlo and Markov chain Monte Carlo.
The Monte Carlo (MC) method is a computer based simulation using random number generators. The name was given by atomic researchers in the 1940s who likened their simulation methods to keeping score in a casino in order to learn some odds. Monte Carlo methods are used in every branch of science and engineering because they allow brute force computer power to be used on problems that are too complicated to solve mathematically. Quasi-Monte Carlo (QMC) methods replace simulated random numbers by strategically chosen ones. By leaving less to chance, large improvements in accuracy are possible. MC and QMC methods are widely applied in computer graphics to make animated movies and other images, in computational finance to control risks, in statistical inference to separate real findings from chance fluctuations, and in many other areas. Much of the top QMC work is done in Europe, Asia, Canada and Australia. This conference will bring together leading researchers from around the world to share results. This project will support travel expenses of US based researchers to participate in this exchange of knowledge. Two of the researchers to be supported are US experts giving major(plenary) talks. Of the other researchers to be supported priority will be given to academically young US researchers in the mathematical sciences, especially postdoctoral students and junior faculty, but also including PhD students.
, the Tenth International Conference on Monte Carlo and Quasi-Monte Carlo Methods in Scientific Computing, held in Sydney, Australia, from February 13 to February 17 of 2012. Scientific Merit Monte Carlo methods are used in all branches of science. They are simulations driven by random number generators. Quasi-Monte Carlo (QMC) methods involve a kind of derandomization of the Monte Carlo method. They generate points that are more evenly distributed than genuinely randomly sampled points are. The result can be greatly improved accuracy even in high dimensional settings that were long thought to be the exclusive domain of Monte Carlo methods. This conference brought together mathematicians, statisticians, and computer scientists to share results about Monte Carlo and quasi-Monte Carlo methods. The two methods are linked and many researchers randomize their QMC points to estimate error and even reduce error by cancellation. Both MC and QMC make use of deep results in mathematics to construct and investigate point sets, although the results are easily usable by non-specialists. It is interesting that the mathematics used to construct good random number generators strongly parallel that used to construct good QMC point sets. Broader impacts Much of the best work in MC and QMC is done in Europe and Asia. The conference site, UNSW in Australia, is also home to a concentration of excellent QMC researchers. This grant was designed to provide support for U.S. researchers to catch up on developments from overseas. Two of the supported U.S. researchers gave high profile (plenary) talks presenting U.S. based results to the international audience. The others were mostly new researchers: two PhD students, one postdoc, an assistant professor and an associate professor. These are researchers who might otherwise not have travel funding and hence have more difficulty keeping up with the latest developments.
