AI summary by TwinMind
Lecture 10: Simple Functions
MIT OpenCourseWare
1h 22m 55s
Review of Measurable Functions
- Measurable function f: E → extended reals if for all α ∈ ℝ, f⁻¹((α,∞)) is measurable
- Pre-image of any Borel set is measurable
- Measurability preserved under:
- Sups, lim sups, lim infs, limits
- Changing function on measure zero sets
- Linear combinations and products
Complex-Valued Measurable Functions
- f: E → ℂ is measurable if Re(f) and Im(f) are measurable
- f = Re(f) + i·Im(f)
- Properties preserved: αf, f+g, fg, |f| are measurable for measurable f,g and complex α
- Pointwise limits of measurable functions are measurable
- Follows from real/imaginary parts converging separately
- Uses previous theorems for extended real-valued functions
Simple Functions Definition
- Measurable function φ: E → ℂ is simple if range φ(E) = {a₁, a₂, ..., aₙ} (finite)
- Canonical representation: φ(x) = Σⱼ₌₁ⁿ aⱼ χ_{Aⱼ}(x)
- Aⱼ = φ⁻¹({aⱼ}) are measurable, disjoint
- ∪ⱼ Aⱼ = E
- Linear combinations, products of simple functions remain simple
Simple Function Approximation Theorem (Non-negative Functions)
- For f: E → [0,∞] measurable, ∃ sequence {φₙ} of simple functions:
- a) 0 ≤ φ₀(x) ≤ φ₁(x) ≤ φ₂(x) ≤ ... ≤ f(x) ∀x ∈ E
- b) φₙ(x) → f(x) pointwise
- c) Uniform convergence on {x: f(x) ≤ b} for any b ≥ 0
Construction of Approximating Sequence
- For n = 0,1,2,..., define:
- Eₖ,ₙ = {x ∈ E: k·2⁻ⁿ ≤ f(x) < (k+1)·2⁻ⁿ} for k = 0,...,2^(2n)-1
- Fₙ = {x ∈ E: f(x) ≥ 2ⁿ}
- φₙ(x) = Σₖ₌₀^(2^(2n)-1) k·2⁻ⁿ χ{Eₖ,ₙ}(x) + 2ⁿ χ{Fₙ}(x)
Proof Strategy
- Part (a): Show φₙ ≤ φₙ₊₁ by examining set inclusions
- Key estimate: For x ∈ {f ≤ 2ⁿ}, |f(x) - φₙ(x)| ≤ 2⁻ⁿ
- Parts (b,c) follow from this estimate and monotonicity
Extension to Complex-Valued Functions
- For f: E → ℂ measurable, decompose using positive/negative parts:
- f = Re(f)⁺ - Re(f)⁻ + i(Im(f)⁺ - Im(f)⁻)
- Apply previous theorem to each non-negative component
- Properties:
- |φₙ(x)| increasing to |f(x)|
- φₙ → f pointwise
- Uniform convergence on bounded sets
Lebesgue Integration of Simple Functions
- Definition: For φ = Σⱼ₌₁ⁿ aⱼ χ_{Aⱼ} (canonical form)
- ∫_E φ dx = Σⱼ₌₁ⁿ aⱼ m(Aⱼ)
- Set notation: L⁺(E) = {non-negative measurable functions on E}
Properties of Simple Function Integration
- Homogeneity: ∫_E cφ = c∫_E φ for c ≥ 0
- Additivity: ∫_E (φ + ψ) = ∫_E φ + ∫_E ψ
- Monotonicity: φ ≤ ψ ⟹ ∫_E φ ≤ ∫_E ψ
- Domain restriction: ∫_F φ = ∫_E φχ_F ≤ ∫_E φ for F ⊆ E measurable
Proof Techniques
- Use canonical representations with disjoint sets
- Apply additivity of Lebesgue measure
- Exploit disjoint union decompositions
