70 likes | 193 Views
Multivariate Numerical Integration. Rida Assaf Western Michigan University r ida.assaf@wmich.edu. Adaptive Multivariate integration. For a given function of the form : I f = Given the approximate Q f, calculate I f where : E f = |Q f – I f | < tolerance
E N D
Multivariate Numerical Integration Rida Assaf Western Michigan University rida.assaf@wmich.edu
Adaptive Multivariate integration • For a given function of the form : If = Given the approximate Qf, calculate If where : Ef = |Qf– If | < tolerance - f might have a lot of singularities in the interior domain.
Adaptive Multivariate integration • Calculate If using Gauss-Konrod 21 points weighted sum scheme :
Adaptive Multivariate integration • Dealing with n-dimensional functions for n 3 for n = 3 our function looks like : If =
Code • Tasks: • Implement adaptive univariate integration routine ✔ • Use the adaptive routine in all directions ✔ • Parallelize the outer Gauss-Konrod iterations✔
Updates • Updates: • Functions support Epsilon and Deltas ✔ • Distributed Memory Systems support(MPI) • Data Parallelism • Task Parallelism
Questions? Thank you!