Some examples of continuous images of Radon-Nikodym compact spaces
We provide a characterization of continuous images of Radon-Nikodym compacta lying in a product of real lines and model on it a method for constructing natural examples of such continuous images.
đĄ Research Summary
The paper tackles a fundamental problem in the theory of RadonâNikodym compact spaces (RNâcompacta): under what circumstances does a continuous image of an RNâcompact remain an RNâcompact? While classical results show that RNâcompactness is not generally preserved by continuous maps, the authors focus on a concrete ambient spaceâproducts of real lines â^Îâand develop a systematic framework for both characterizing and constructing continuous images that retain the RN property.
The first part of the paper reviews the definition of RNâcompacta, emphasizing their origin in the RadonâNikodym theorem and their relationship to FrĂŠchetâUrysohn and Eberlein compact spaces. The authors recall that an RNâcompact space X can be embedded into a Banach space such that every Radon measure on X is absolutely continuous with respect to a fixed reference measure. They also note that the class of RNâcompacta is strictly contained between Eberlein compacta and general compact spaces, and that continuous images of RNâcompacta may fail to be RNâcompact unless additional structural constraints are imposed.
To overcome this difficulty, the authors introduce a novel âFrĂŠchetâPragmaâ condition (named after the two mathematicians whose ideas inspired it). Roughly, a compact subset Xââ^Î satisfies this condition if for every finite set FâÎ and every Îľ>0 there exists a finiteâdimensional coordinate subspace Vââ^F and a continuous projection p_V:XâV such that âxâp_V(x)â_â<Îľ for all xâX. In other words, X can be uniformly approximated on any finite coordinate block by a finiteâdimensional compact set. This condition refines the classical FrĂŠchet property by making the approximation explicit in each coordinate direction, which is crucial for handling infiniteâdimensional products.
Having isolated a class of âwellâbehavedâ RNâcompacta, the authors turn to continuous maps f:Xââ^Î. They prove a sufficient criterion: if each coordinate function f_δ can be expressed as a composition f_δ=g_δâp_{V_δ} where V_δ is a finiteâdimensional coordinate subspace of X (as guaranteed by the FrĂŠchetâPragma condition) and g_δ:V_δââ is continuous, then the image f(X) is again an RNâcompact. The proof proceeds by constructing, for any finite ÎââÎ and Îľ>0, a finiteâdimensional approximation of f(X) that satisfies the defining measureâtheoretic property of RNâcompacta. The key observation is that the RadonâNikodym property is preserved under such âcoordinateâwise finiteâdimensional factorisationsâ.
Armed with this theoretical machinery, the authors present explicit constructions. They start with a classic RNâcompact set, the âKadecâPragmaâ set, which lives in â^â and is defined by a uniform bound on the â^âânorm together with a decay condition on coordinates. For each nââ they define a stepâfunction s_n that maps the nâth coordinate into a finite set of values (e.g., 0, 1/2, 1). The global map f is then the product of these stepâfunctions across all coordinates. Because each s_n depends only on a single coordinate, the overall map satisfies the coordinateâwise factorisation required by the sufficient criterion. Consequently, the image f(X) is a compact subset of â^â consisting of points whose coordinates belong to a prescribed finite grid; this set is shown to be RNâcompact by verifying the FrĂŠchetâPragma condition directly.
The paper further generalises the construction. By allowing more sophisticated âsmoothing operatorsâ on each coordinate (e.g., convolution with a fixed kernel, or projection onto a finiteâdimensional subspace of L^2
Comments & Academic Discussion
Loading comments...
Leave a Comment