Exercise Sheet 4, ST213 1) Let (Ω, F, P) be a measure space and X a non-negative Borel function. Show that Z Q(A) = 1A X dP, A∈F defines a measure on (Ω, F). 2) Let (Ω, F, P) be a measure space and Q be absolutely continuous with respect to P (we write Q P) with density D = dQ\dP. Let further X be a Borel real-valued function on Ω. Show that the integral of X with respect to Q is defined if and only if the integral of X D with respect to P is defined, in which case they are equal, i.e.: Z Z X dQ = X D dP. Show this first for X a simple Borel function. 3) Let (Ω, F, P) be a measure space, X and Y two Borel real-valued functions. R R (a) Show that if X is integrable and X = Y a.e. then Y is integrable and XdP = Y dP. (b) Let Q be a measure on (Ω, F, P) such that Q P, and A ∈ F. Show that if P(A) = 0 then Q(A) = 0, but (in general) not vice versa. 4) Let Pn , n = 1, 2, . . ., be a sequence of measures on a measurable space (Ω, F). P∞ a) Show that Q = n=1 Pn is a measure on (Ω, F). b) Let X be a non-negative Borel function. Show that Z X dQ = ∞ Z X X dPn . n=1 Remark that Q(A) might be infinite for some A ∈ F. In such a case the definition of the integral implies that Z (a1A ) dQ = a · Q(A) = a · ∞ for a ≥ 0 and the only meaningful definition of this a · ∞ in integration theory is a · ∞ = ∞ if a > 0 but a · ∞ = 0 if a = 0. Furthermore a + ∞ = ∞ for a ≥ 0 including a = ∞. NEVER calculate anything like +∞ − ∞ without extra consideration! 5) Let (Ω, F) be a measurable space. Pn (a) Let X = i=1 ai 1Ai , Ai ∈ F, ai ∈ [0, +∞), i = 1, . . . , n, n ∈ N. Show that X is a (simple, of course) Borel function. (b) Let X : Ω → R be nonnegative, Borel measurable, i ∈ N. Define: i i2 X k − 1 i i + i1{X≥i} . Xi = 1 2i {(k−1)/2 ≤X<k/2 } k=1 Show that Xi is a simple Borel function and that Xi ↑ X as i → ∞, uniformly on every set on which X is bounded. 6) Let Ω be any set and ω ∈ Ω. Define δω : P(Ω) → [0, 1] by letting δωP (A) = 1A (ω). Realise that δω is a ∞ probability measure on P(Ω), the Dirac measure at ω. Conclude that n=1 δn is a measure, the so-called counting measure, on P(N). 1