Bourbaki

Exercise 4, § 3

Integration, IX, p. 109

Notations and hypotheses are those of the preceding exercise. One calls abstract measure on (E,T)(E, \mathscr{T}) every mapping m of T\mathscr{T} into R‾+\overline{\mathbf{R}}_+ having the following property: for every sequence (An)n∈N(A_n)_{n \in \mathbf{N}} of sets belonging to T\mathscr{T}, pairwise disjoint, one has m(⋃n∈NAn)=∑n∈Nm(An)m(\bigcup_{n \in \mathbf{N}} A_n) = \sum_{n \in \mathbf{N}} m(A_n). One calls integral on (E,T)(E, \mathscr{T}) every mapping J of M\mathscr{M} into R‾+\overline{\mathbf{R}}_+ satisfying the following conditions: α)\alpha) J(λ⋅f)=λ⋅J(f)J(\lambda \cdot f) = \lambda \cdot J(f) for λ∈R‾+\lambda \in \overline{\mathbf{R}}_+ and f∈Mf \in \mathscr{M} (with the usual convention 0⋅(+∞)=00 \cdot (+\infty) = 0); β)\beta) J(∑n∈Nfn)=∑n∈NJ(fn)J\left( \sum_{n \in \mathbf{N}} f_n \right) = \sum_{n \in \mathbf{N}} J(f_n) for every sequence (fn)n∈N(f_n)_{n \in \mathbf{N}} of elements of M\mathscr{M}.

a) Let J be an integral on (E,T)(E, \mathscr{T}); for every subset A∈TA \in \mathscr{T}, set mJ(A)=J(φA)m_J(A) = J(\varphi_A); show that mJm_J is an abstract measure and that the mapping J↦mJJ \mapsto m_J is a bijection of the set of integrals on (E,T)(E, \mathscr{T}) onto the set of abstract measures on (E,T)(E, \mathscr{T}) (if m is an abstract measure, first define J(f)J(f) by linearity for the finite positive T\mathscr{T}-step functions and use Exer. 3 e)). If m is an abstract measure, we will denote by m∗m^* the corresponding integral.

b) Let m be an abstract measure on (E,T)(E, \mathscr{T}). For every positive function f on E, finite or not, set m∗(f)=inf⁡g∈M,g≥fm∗(g)m^*(f) = \inf_{g \in \mathscr{M}, g \geq f} m^*(g). Show that m∗m^* is an encumbrance on E (imitate the proof of Th. 3 of Ch. IV, §1, No. 3). One sets m∗(A)=m∗(φA)m^*(A) = m^*(\varphi_A) for every subset A of E.

Solution

No solution has been written yet.

Transcribed from the printed volume. Not translated.