xiv Notation measure μ, with the usual identification of functions which coincide almost everywhere. Whenever p ≥ 1, we shall denote by p its conjugate exponent: 1 p + 1 p = 1. Whenever T is a map from a measure space X, equipped with a measure μ, to an arbitrary space Y , we denote by T #μ the image measure (or push- forward) of μ by T . Explicitly, (T #μ)[B] = μ[T −1(B)], where T −1(B) = {x ∈ X T (x) ∈ B}. The set of all T : X → X such that T #μ = μ will be denoted by S(X). We shall always use push-forward in this sense: when we write T #f = g, where f and g are nonnegative functions, this means that the measure having density f is pushed foward to the measure having density g (usually the reference measure will be the Lebesgue measure). If X is a topological space, then it will be equipped with its Borel σ- algebra. We shall denote by C(X) the space of continuous functions on X by Cb(X) the space of bounded continuous functions on X and by C0(X) the space of continuous functions on X going to 0 at infinity. Sometimes these notations will be replaced by C(X R), Cb(X R), C0(X R). The space Cb(X) comes with a natural norm, u ∞ = supX |u|. Whenever A ⊂ X, we denote by Int(A) the largest open set contained in A, and by A the smallest closed set containing A. We set ∂A = A \ Int(A). By definition, the support of a measure μ on X will be the smallest closed set F ⊂ X with μ[X \ F ] = 0, and will be denoted Supp μ. On the other hand, when we say that μ is concentrated on A ⊂ X, this only means that μ[X \ A] = 0, without A being necessarily closed. If X is a metric space, we shall equip it with the topology induced by its distance, and denote by B(x, r) the ball of radius r and center x. We shall denote by Lip(X) the set of all Lipschitz functions on X we shall also denote by Pp(X) the space of Borel probability measures μ on X with finite moment of order p, in the sense that d(x0,x)p dμ(x) +∞ for some (and thus any) x0 ∈ X. When X is a Banach space and X∗ its topological dual, we shall denote by ·, · the duality bracket between X and X∗. A particular case of this is the scalar product in a Hilbert space. If ϕ is a convex function on a Banach space X, then ϕ∗ will stand for its dual convex function, in the sense of Legendre-Fenchel duality. The subdifferential of ϕ will be denoted by ∂ϕ, and identified with its graph, which is a subset of X × X∗. Basic definitions for these objects are recalled

Purchased from American Mathematical Society for the exclusive use of nofirst nolast (email unknown) Copyright 2003 American Mathematical Society. Duplication prohibited. Please report unauthorized use to cust-serv@ams.org. Thank You! Your purchase supports the AMS' mission, programs, and services for the mathematical community.