[ANN] Tasmanian.jl Adaptive Sparse Grid Function Approximation




I wrapped the TASMANIAN library for julia in

This provides adaptive sparse grid approximations, useful for people who need to approximate high dimensional functions. The package is not registered yet as there is too little documentation, but the examples are illustrative. comments welcome!


Nice addition @floswald, thanks! Do you know if the TASMANIAN supports some of the Smolyak-type sparse grids developed at Sandia labs? https://web.stanford.edu/~paulcon/slides/Oxford_2012.pdf

It is a general sparsification scheme that I had to implement in C++ back in the days, but that I’d love to have in Julia. The idea is that given any of those orthogonal polynomial basis (e.g. Legendre, Laguerre, Hermite …), it produces an optimal approximation for a minimum number of monomials retained.


Found https://github.com/ZacLN/SparseInterp.jl


Looks great! Did you try it out?


Didn’t try, just amazed how Julians are always aware of the most advanced technology out there :slight_smile:

Tasmanian.jl is definitely a good addition that I will remember if I need sparse grids in the future.