Means on bounded uniformly continuous functions #
For a topological group G, a mean on TauCeti.UCB G is a real linear functional m which is
positive, 0 ≤ f → 0 ≤ m f, and normalized, m 1 = 1. A left-invariant mean is the object whose
existence defines amenability of G.
This file develops the basic theory of means. Continuity is not part of the definition: since
every f lies below the constant ‖f‖, positivity and normalization give |m f| ≤ ‖f‖, so a mean
is 1-Lipschitz for the supremum norm, and in particular compatible with uniform limits. The
derived continuous linear functional has operator norm one. The means form a convex subset of the
continuous dual of UCB G, and evaluation at a point of G is a mean, so this set is nonempty.
Main definitions #
TauCeti.Mean G: a positive normalized linear functional onUCB G.TauCeti.Mean.toContinuousLinearMap: the continuous linear functional underlying a mean.TauCeti.Mean.eval: evaluation at a point, as a mean.
Main results #
TauCeti.Mean.map_algebraMap: a mean takes the valuecat the constant functionc.TauCeti.Mean.ciInf_le_apply,TauCeti.Mean.apply_le_ciSup:⨅ x, f x ≤ m f ≤ ⨆ x, f x.TauCeti.Mean.abs_apply_le_norm,TauCeti.Mean.abs_apply_sub_apply_le:|m f| ≤ ‖f‖and|m f - m g| ≤ ‖f - g‖.TauCeti.Mean.lipschitzWith: a mean is1-Lipschitz; it is therefore aContinuousLinearMapClass.TauCeti.Mean.norm_toContinuousLinearMap: a mean has operator norm one.TauCeti.Mean.mem_range_toContinuousLinearMap_iff: the continuous linear functionals coming from means are exactly the positive ones taking the value1at1.TauCeti.Mean.convex_range_toContinuousLinearMap: the means form a convex set.
References #
- B. Bekka, P. de la Harpe, A. Valette, Kazhdan's Property (T), Appendix G.1.
A mean on UCB G: a positive real linear functional taking the value 1 at the constant
function 1.
- toFun : TauCeti.UCB G → ℝ
A mean takes the value
1at the constant function1.
Instances For
Two means are equal when they agree on every function.
A mean takes the value 1 at the constant function 1.
A mean built from a linear functional which is nonnegative on nonnegative functions and takes
the value 1 at 1.
Instances For
A mean takes the value c at the constant function c.
A mean of a function bounded above by c is at most c.
A mean of a function bounded below by c is at least c.
A mean of a function is at least the infimum of its values.
A mean of a function is at most the supremum of its values.
A mean of a function is bounded in absolute value by its supremum norm.
A mean changes by at most the supremum distance between its arguments.
A mean is 1-Lipschitz for the supremum norm.
A mean is continuous for the supremum norm.
The continuous linear functional underlying a mean.
Instances For
A mean has operator norm one.
A continuous linear functional on UCB G comes from a mean exactly when it is nonnegative on
nonnegative functions and takes the value 1 at 1.
The means on UCB G form a convex set of continuous linear functionals.
Evaluation at a point of G is a mean on UCB G.