Best approximation problems in tensor-product spaces (Q791786)

An important concept in approximation theory is the approximation of a function of several variables by ''natural'' combinations of functions of fewer variables. This paper contributes existence theorems for best approximations to a function of two variables from tensor products of functions of one variable. A variety of results are proved. The hypothesis involves concepts such as existence of continuous projections, weak closed subspaces, and continuity properties of corresponding proximity maps for spaces of one variable functions. The paper uses uniform cross-norms, characterizations of tensor products, and a result the authors have targeted for generalization by naming it the ''sitting duck theorems''.