Abstract
If the computational cost grows rapidly by some high power of the size of the range of variables considered for some computational problem, then it is worthwhile to consider breaking up the problem to a set of smaller coordinate ranges and find a reasonably accurate way to combine these results. Here one such method is described, using cubic grids to break up the problem to smaller ones and using a set of 8 different such cubic grids which are eventually combined taking into accounts some of the expected accuracy variations which depend on the choice of each actual cubic grid.
Keywords
Get full access to this article
View all access options for this article.
