Monday, May 02, 201112:00 PM - 1:00 PMCNLS Conference Room (TA-3, Bldg 1690)|
Some computational and numerical challenges in solving recurrence equations
Ori SargsyanT-6 and CNLS
In this lecture I will illustrate some challenges in solving recurrence equations by using a particular example. The recurrence equation is for computing the density function of a particular random variable that is part of a population genetic model. Based on the recurrence equation I derive an analytic expression for the moment generating function of the random variable and an analytic expression for the density function by inverting the moment generating function. The expression for the moment generating function has computational advantage over an already-existing formula, but evaluating it numerically is challenging because of noise created by cancelation-errors. I will show that this challenge can be overcome by rewriting the analytic expression in another form and using the Wolfram Mathematica software that allows to do exact computation using fraction forms for rational numbers.
Host: Lijun Zhu, firstname.lastname@example.org