Some Algorithms for Computing Best Approximations from Convex Cones
Frank Deutsch, John H. McCabe, George M. Phillips · SIAM Journal on Numerical Analysis · 1975
It is shown that the problem of finding best approximations out of convex cones is equivalent to finding them out of linear subspaces. The subspace which works depends on the element being approximated and may be chosen to be a hyperplane or a finite-dimensional subspace. Algorithms based on these facts are given, and some applications to specific problems are made.