absolute convergence of integral and boundedness of derivative


Theorem.  Assume that we have an http://planetmath.org/node/11865absolutely converging integralDlmfPlanetmath

∫a∞f⁢(x)⁢𝑑x

where the real function f and its derivative f′ are continuousMathworldPlanetmath and f′ additionally bounded on the interval  [a,∞).  Then

limx→∞⁡f⁢(x)= 0. (1)

Proof.  If  c>a,  we obtain

∫acf(x)f′(x)dx=12/ac(f(x))2=(f⁢(c))2-(f⁢(a))22,

from which

(f⁢(c))2=(f⁢(a))2+2⁢∫acf⁢(x)⁢f′⁢(x)⁢𝑑x. (2)

Using the boundedness of f′ and the absolute convergenceMathworldPlanetmath, we can estimate upwards the integral

∫ac|f⁢(x)⁢f′⁢(x)|⁢𝑑x=∫ac|f⁢(x)|⁢|f′⁢(x)|⁢𝑑x≦M⁢∫ac|f⁢(x)|⁢𝑑x≦M⁢∫a∞|f⁢(x)|⁢𝑑x ∀c∈[a,∞)

whence ∫a∞|f⁢(x)⁢f′⁢(x)|⁢𝑑x is finite and thus ∫a∞f⁢(x)⁢f′⁢(x)⁢𝑑x converges absolutely.  Hence (2) implies

limc→∞⁡(f⁢(c))2=(f⁢(a))2+2⁢∫a∞f⁢(x)⁢f′⁢(x)⁢𝑑x,

i.e. limx→∞⁡(f⁢(x))2 exists as finite, therefore also

limx→∞⁡|f⁢(x)|:=A.

Antithesis:  A>0.  It implies that there is an x0(≧a) such that

|f⁢(x)|≧A2 ∀x≧x0.

If now  b>x0, then we had

∫x0b|f⁢(x)|⁢𝑑x≧A2⁢(b-x0)⟶∞ as⁢b→∞.

This means that ∫x0∞|f⁢(x)|⁢𝑑x and consequently also ∫a∞|f⁢(x)|⁢𝑑x would be divergent.  Since it is not true, we infer that  A=0,  i.e. that the assertion (1) is true.

Title absolute convergence of integral and boundedness of derivative
Canonical name AbsoluteConvergenceOfIntegralAndBoundednessOfDerivative
Date of creation 2013-03-22 19:01:28
Last modified on 2013-03-22 19:01:28
Owner pahio (2872)
Last modified by pahio (2872)
Numerical id 7
Author pahio (2872)
Entry type Theorem
Classification msc 40A10
Related topic NecessaryConditionOfConvergence