Continuous Selections for Metric Projections and Interpolating Subspaces

The existence of continuous selections for metric projections is the theoretical foundation of the existence of stable algorithms for computing best approximation elements. In this monograph we will give various intrinsic characterizations of subspaces of C o(T) which ensure the existence of continuous metric selections. Since the Chebyshev approximation is a special case of semi-infinite optimization, we hope that our study will give some insight to stability problems in semi-infinite optimization as well as parametric optimizations.