Constrained Fermat-Torricelli-Weber Problem in real Hilbert Spaces (1806.04296v1)

Abstract: The Fermat-Weber location problem requires finding a point in $\mathbb{R}^n$ that minimizes the sum of weighted Euclidean distances to $m$ given points. An iterative solution method for this problem was first introduced by E. Weiszfeld in 1937. Global convergence of Weiszfeld's algorithm was proven by W. Kuhn in 1973. This paper studies Fermat-Weber location problems with closed convex constraint sets in real Hilbert spaces. In section two, we show that existence and uniqueness of solutions of the problems. Moreover, the solution is stable with respect to the perturbation of the $m$ anchor points. In the section three, we extend Weiszfeld's algorithm by adding a projection on the constraint set. The convergence of the sequence generated by the method to the optimal solution of the problem is proved.