Two fast algorithms for projecting a point onto the canonical simplex (Q327053)

Two algorithms are considered for finding the orthogonal projection of a point onto a convex polyhedron. The first one is the scalar algorithm based on the algebraic analysis of the Kuhn-Tucker optimality conditions. The second one is the vector algorithm based on a recurrence of vector quantities. This paper presents improved versions of the description and proof of the finite convergence of the scalar and vector algorithms. Numerical results on the computational complexity of the two algorithms are presented.