Lipschitz Extension Constants Equal Projection Constants
نویسنده
چکیده
For a Banach space V we define its Lipschitz extension constant, LE(V ), to be the infimum of the constants c such that for every metric space (Z, ρ), every X ⊂ Z, and every f : X → V , there is an extension, g, of f to Z such that L(g) ≤ cL(f), where L denotes the Lipschitz constant. The basic theorem is that when V is finite-dimensional we have LE(V ) = PC(V ) where PC(V ) is the well-known projection constant of V . We obtain some direct consequences of this theorem, especially when V = Mn(C). We then apply known techniques for calculating projection constants, involving averaging of projections, to calculate LE((Mn(C))sa). We also discuss what happens if we also require that ‖g‖∞ = ‖f‖∞. In my exploration of the relationship between vector bundles and Gromov– Hausdorff distance [20] I need to be able to extend matrix-valued functions from a closed subset of a compact metric space to the whole metric space, with as little increase of the Lipschitz constant as possible. There is a substantial literature concerned with extending Lipschitz functions, but I have had difficulty finding there the facts which I need. The purpose of this largely expository paper is to describe and employ a very strong relationship between the Lipschitz extension problem and what is referred to as the “projection constant” for finite-dimensional Banach spaces. This permits us to bring to bear on the Lipschitz extension problem the quite substantial literature concerning projection constants. This then provides the facts which I need, as well as other interesting facts. In Section 1 we introduce what we call the Lipschitz extension constant, LE(V ), of a Banach space V . I have not found exactly this definition in the literature, although there are definitions very close to it. We also recall the well-known definition of the projection constant, PC(V ), of a Banach space V . The basic theorem is that if V is finite-dimensional, then LE(V ) = PC(V ). I have not found this theorem stated in the literature, probably because LE(V ) is not defined in the literature, but I am told that this theorem is well-known to specialists on the geometry of Banach spaces. In Section 2 we give the proof that LE(V ) ≤ PC(V ), while in Section 3 we give the proof that PC(V ) ≤ LE(V ), thus proving the basic theorem. In Section 4 we give some consequences of the basic theorem which come directly from using facts about projection constants that are available in the literature. One of these consequences is that LE(C) = 4/π. Another of these consequences is the 2000 Mathematics Subject Classification. 46B20; 26A16.
منابع مشابه
Extending Lipschitz Maps into C ( K ) - Spaces
We show that if K is a compact metric space then C(K) is a 2-absolute Lipschitz retract. We then study the best Lipschitz extension constants for maps into C(K) from a given metric space M , extending recent results of Lancien and Randrianantoanina. They showed that a finitedimensional normed space which is polyhedral has the isometric extension property for C(K)-spaces; here we show that the s...
متن کاملOn the Relative Schoenflies Theorem
We prove generalizations of the relative Schoennies extension theorem for topo-logical, quasiconformal, or bi-Lipschitz embeddings due to Gauld and VV aiss all a, and show that maximal dilatations and bi-Lipschitz constants of the extensions can be controlled.
متن کاملHolomorphic Extension on Product Lipschitz Surfaces in Two Complex Variables
In this work we prove a new Lp holomorphic extension result for functions defined on product Lipschitz surfaces with small Lipschitz constants in two complex variables. We define biparameter and partial Cauchy integral operators that play the role of boundary values for holomorphic functions on product Lipschitz domain. In the spirit of the application of David-Journé-Semmes [DJS85] and Christ’...
متن کاملA univariate global search working with a set of Lipschitz constants for the first derivative
In the paper, a global optimization problem is considered where the objective function f(x) is univariate, black-box, and its first derivative f (x) satisfies the Lipschitz condition with an unknown Lipschitz constant K. In the literature, there exist methods solving this problem by using an a priori given estimate of K, its adaptive estimates, and adaptive estimates of local Lipschitz constant...
متن کاملA Decomposition Method for MINLPs with Lipschitz Continuous Nonlinearities
Many mixed-integer optimization problems are constrained by nonlinear functions that do not possess desirable analytical properties like convexity or factorability or cannot even be evaluated exactly. This is, e.g., the case for problems constrained by differential equations or for models that rely on black-box simulation runs. For these problem classes, we present, analyze, and test algorithms...
متن کامل