I read a paper now (field: theoretical physics, which is also my background), where they claim that some approach to solve the problem is "numerically exact".
What does this mean?
I read a paper now (field: theoretical physics, which is also my background), where they claim that some approach to solve the problem is "numerically exact".
What does this mean?
Copyright © 2021 JogjaFile Inc.
It means that the algorithm used to solve the problem does not involve any approximation so that the final result is exact up to machine precision ($10^{-16}$ for double precision floats for example).