The Setting
Let be a probability space and a real Hilbert space, with inner product . The mean of a function is the vector of , and its second moment is the number . The question is how far strays from a fixed vector in mean square, as measured by , once the second moment, the vector and the gap between the mean and are known.
Let be the set under study, whose points are written with in the base plane, and a bi-Lipschitz map from to a finite-dimensional real space. For each , the sheet of is the map , , on the open set of base points with . A derivative column of a sheet is one of its partial derivatives , in the horizontal and vertical directions, taken over a rectangle of its domain with the normalized Lebesgue measure on , so that the mean is . The second moment of is set against , the largest first coordinate of the closure of the pairs over all sheets and the points of at which both partial derivatives exist.
The second question is in one variable: how large can the increment of a function on the real line be between two points of an interval, when is an indefinite integral of a function and the energy of on the interval, the mean of over it, is given?
Means on a Segment
A segment is a closed interval of the real line with ✦; its length is .✦ The uniform law on is the probability measure , where is Lebesgue measure on the line.✦ The mean over of a function integrable on is (1)
A bounded measurable field on a measurable space with values in a real normed space is a strongly measurable map that is bounded: there is a constant with for all .✦
The Claim
Theorem 1. Let be a probability space, a real Hilbert space, a function such that and are integrable, and . Then satisfies the variance identity (2)✦ and for every real number the deficit of its second moment below satisfies (3)✦ If moreover for a real number , then (4)✦
Theorem 2. Let be a real Banach space, a segment, a bounded measurable field, and an indefinite integral of : for all real numbers and . Then for all the increment of is bounded by the energy of on : (5)✦
What Is Checked
The identity (2) and the bounds (3) and (4) are stated over an arbitrary probability space and a function from it to a real Hilbert space with and integrable; the bound (5) is stated over a segment with its length and its uniform law , a bounded measurable field on the line with values in a real Banach space, and an indefinite integral of .