תרגיל 1
נתון האינטגרל המוכלל
∫ 0 1 ln x x α d x \int_{0}^1 \frac{\ln x}{x^\alpha}dx 0 ∫ 1 x α ln x d x
הוכיחו כי האינטגרל מתבדר עבור כל α ≥ 1 \alpha\geq 1 α ≥ 1 .
הוכחה: יהי α ∈ R \alpha \in \mathbb{R} α ∈ R .
אם α = 1 \alpha=1 α = 1 מתקיים:
∫ 0 1 ln x x α d x = lim ε → 0 + ∫ ε 1 ln x x d x = ↓ שיטת ההצבה ( t = ln x d t = 1 x d x ) lim ε → 0 + ∫ ln ε 0 t d t = lim ε → 0 + t 2 2 ∣ ln ε 0 = lim ε → 0 + ( − ln 2 ε 2 ) = − ∞ \int_{0}^1 \frac{\ln x}{x^\alpha}dx=\lim_{ \epsilon \to 0^+ } \int_{\epsilon}^1 \frac{\ln x}{x}dx \underset{\begin{matrix}
\downarrow \\
\R{\text{שיטת ההצבה}} \\
\begin{pmatrix}
t=\ln x \\
dt=\frac{1}{x}dx
\end{pmatrix}
\end{matrix}}{=} \lim_{ \epsilon \to 0^+ } \int_{\ln\epsilon}^{0} tdt=\lim_{ \epsilon \to 0^+ } \left. \frac{t^2}{2}\right\vert_{\ln\epsilon}^0 =\lim_{ \epsilon \to 0^+ } \left( -\frac{\ln^2\epsilon}{2} \right)=-\infty 0 ∫ 1 x α ln x d x = ε → 0 + lim ε ∫ 1 x ln x d x ↓ שיטת ההצבה ( t = ln x d t = x 1 d x ) = ε → 0 + lim l n ε ∫ 0 t d t = ε → 0 + lim 2 t 2 l n ε 0 = ε → 0 + lim ( − 2 ln 2 ε ) = − ∞
⇐ לכן, לפי הגדרה, האינטגרל מתבדר עבור α = 1 \alpha=1 α = 1 .
אחרת מתקיים:
∫ 0 1 ln x x α d x = lim ε → 0 + ∫ ε 1 ln x x α d x = ↓ שיטת ההצבה ( t = ln x d t = 1 x d x x = e t ) lim ε → 0 + ∫ ln ε 0 t e ( α − 1 ) t d t = lim ε → 0 + ∫ ln ε 0 e ( 1 − α ) t t d t = ↓ אינטגרציה בחלקים ( u = t ⟹ u ′ = 1 v ′ = e ( 1 − α ) t ⟹ v = e ( 1 − α ) t 1 − α ) lim ε → 0 + ( e ( 1 − α ) t 1 − α t ∣ ln ε 0 − 1 1 − α ∫ ln ε 0 e ( 1 − α ) t d t ) = lim ε → 0 + ( e ( 1 − α ) t 1 − α t ∣ ln ε 0 − e ( 1 − α ) t ( 1 − α ) 2 ∣ ln ε 0 ) = lim ε → 0 + ( e ( 1 − α ) ln ε 1 − α − 1 ( 1 − α ) 2 + e ( 1 − α ) ln ε ( 1 − α ) 2 ) = lim ε → 0 + ( e ( 1 − α ) ln ε 1 − α + e ( 1 − α ) ln ε ( 1 − α ) 2 ) − 1 ( 1 − α ) 2 \begin{align*}
\int_{0}^1 \frac{\ln x}{x^\alpha}dx&=\lim_{ \epsilon \to 0^+ } \int_{\epsilon}^1 \frac{\ln x}{x^\alpha}dx \underset{\begin{matrix}
\downarrow \\
\R{\text{שיטת ההצבה}} \\
\begin{pmatrix}
t=\ln x \\
dt=\frac{1}{x}dx \\
x=e^t
\end{pmatrix}
\end{matrix}}{=} \lim_{ \epsilon \to 0^+ } \int_{\ln\epsilon}^0 \frac{t}{e^{(\alpha-1)t}}dt=\lim_{ \epsilon \to 0^+ } \int_{\ln\epsilon}^0 e^{(1-\alpha)t}tdt \\
&\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\underset{\begin{matrix}
\downarrow \\
\R{\text{אינטגרציה בחלקים}} \\
\begin{pmatrix}
u=t\implies u'=1 \\
v'=e^{(1-\alpha)t}\implies v=\frac{e^{(1-\alpha)t}}{1-\alpha}
\end{pmatrix}
\end{matrix}}{=}\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! \lim_{ \epsilon \to 0^+ } \left( \left. \frac{e^{(1-\alpha)t}}{1-\alpha}t \right\vert_{\ln\epsilon}^0 -\frac{1}{1-\alpha}\int_{\ln\epsilon}^0 e^{(1-\alpha)t}dt \right) \\
&=\lim_{ \epsilon \to 0^+ } \left( \left. \frac{e^{(1-\alpha)t}}{1-\alpha}t \right\vert_{\ln\epsilon}^0 - \left. \frac{e^{(1-\alpha)t}}{(1-\alpha)^2} \right\vert_{\ln\epsilon}^0 \right) \\
&=\lim_{ \epsilon \to 0^+ } \left( \frac{e^{(1-\alpha)\ln\epsilon}}{1-\alpha}-\frac{1}{(1-\alpha)^2}+\frac{e^{(1-\alpha)\ln\epsilon}}{(1-\alpha)^2} \right) \\
&=\lim_{ \epsilon \to 0^+ } \left( \frac{e^{(1-\alpha)\ln\epsilon}}{1-\alpha}+\frac{e^{(1-\alpha)\ln\epsilon}}{(1-\alpha)^2} \right) -\frac{1}{(1-\alpha)^2}
\end{align*} 0 ∫ 1 x α ln x d x = ε → 0 + lim ε ∫ 1 x α ln x d x ↓ שיטת ההצבה t = ln x d t = x 1 d x x = e t = ε → 0 + lim l n ε ∫ 0 e ( α − 1 ) t t d t = ε → 0 + lim l n ε ∫ 0 e ( 1 − α ) t t d t ↓ אינטגרציה בחלקים ( u = t ⟹ u ′ = 1 v ′ = e ( 1 − α ) t ⟹ v = 1 − α e ( 1 − α ) t ) = ε → 0 + lim 1 − α e ( 1 − α ) t t l n ε 0 − 1 − α 1 l n ε ∫ 0 e ( 1 − α ) t d t = ε → 0 + lim ( 1 − α e ( 1 − α ) t t l n ε 0 − ( 1 − α ) 2 e ( 1 − α ) t l n ε 0 ) = ε → 0 + lim ( 1 − α e ( 1 − α ) l n ε − ( 1 − α ) 2 1 + ( 1 − α ) 2 e ( 1 − α ) l n ε ) = ε → 0 + lim ( 1 − α e ( 1 − α ) l n ε + ( 1 − α ) 2 e ( 1 − α ) l n ε ) − ( 1 − α ) 2 1
אם α > 1 \alpha> 1 α > 1 אז מתקיים:
( 1 − α ) ln ε = − ( α − 1 ) ln ε = ( α − 1 ) ln 1 ε (1-\alpha)\ln\epsilon=-(\alpha-1)\ln \epsilon=(\alpha-1)\ln \frac{1}{\epsilon} ( 1 − α ) ln ε = − ( α − 1 ) ln ε = ( α − 1 ) ln ε 1
הפונקציה e ( α − 1 ) k 1 − α \frac{e^{(\alpha-1)k}}{1-\alpha} 1 − α e ( α − 1 ) k היא אלמנטרית ולכן רציפה ובנוסף מקיימת e ( α − 1 ) k 1 − α ⟶ k → ∞ − ∞ \frac{e^{(\alpha-1)k}}{1-\alpha}\underset{k\to \infty}{\longrightarrow}-\infty 1 − α e ( α − 1 ) k k → ∞ ⟶ − ∞ שכן α > 1 \alpha>1 α > 1 ו-e ℓ k ⟶ k → ∞ ∞ e^{\ell k}\underset{k\to \infty}{\longrightarrow}\infty e ℓ k k → ∞ ⟶ ∞ כאשר ℓ > 0 \ell >0 ℓ > 0 .
בנוסף, ln 1 ε ⟶ ε → 0 + ∞ \ln \frac{1}{\epsilon}\underset{\epsilon\to0^+}{\longrightarrow}\infty ln ε 1 ε → 0 + ⟶ ∞ ולכן מתקיים:
∫ 0 1 ln x x α d x = lim ε → 0 + ( e ( 1 − α ) ln ε 1 − α + e ( 1 − α ) ln ε ( 1 − α ) 2 ) − 1 ( 1 − α ) 2 = − ∞ \int_{0}^1 \frac{\ln x}{x^\alpha}dx=\lim_{ \epsilon \to 0^+ } \left( \frac{e^{(1-\alpha)\ln\epsilon}}{1-\alpha}+\frac{e^{(1-\alpha)\ln\epsilon}}{(1-\alpha)^2} \right) -\frac{1}{(1-\alpha)^2}=-\infty 0 ∫ 1 x α ln x d x = ε → 0 + lim ( 1 − α e ( 1 − α ) l n ε + ( 1 − α ) 2 e ( 1 − α ) l n ε ) − ( 1 − α ) 2 1 = − ∞
לכן, לכל α ≥ 1 \alpha\geq 1 α ≥ 1 מתקיים כי ∫ 0 1 ln x x α d x \displaystyle \int_{0}^1 \frac{\ln x}{x^\alpha}dx 0 ∫ 1 x α ln x d x מתבדר.
∎
השתמשו בהגדרת אינטגרל מוכלל ובאינטגרציה בחלקים כדי להוכיח כי האינטגרל מתכנס עבור כל 0 < α < 1 0<\alpha<1 0 < α < 1 .
הוכחה:
יהי 0 < α < 1 0<\alpha<1 0 < α < 1 . אזי:
∫ 0 1 ln x x α d x = lim ε → 0 + ∫ ε 1 ln x x α d x = ↓ אינטגרציה בחלקים ( u = ln x ⟹ u ′ = 1 x v ′ = 1 x α ⟹ v = 1 ( 1 − α ) x α − 1 ) lim ε → 0 + ( ln x ( 1 − α ) x α − 1 ∣ ε 1 − ∫ ε 1 1 ( 1 − α ) x α d x ) = lim ε → 0 + ( ln x ( 1 − α ) x α − 1 ∣ ε 1 − 1 ( 1 − α ) 2 x α − 1 ∣ ε 1 ) = lim ε → 0 + ( − ln ε ( 1 − α ) ε α − 1 − 1 ( 1 − α ) 2 + 1 ( 1 − α ) 2 ε α − 1 ) = lim ε → 0 + ( 1 − ( 1 − α ) ln ε ( 1 − α ) 2 ε α − 1 − 1 ( 1 − α ) 2 ) = lim ε → 0 + 1 − ( 1 − α ) ln ε ( 1 − α ) 2 ε α − 1 − 1 ( 1 − α ) 2 = ↓ α − 1 < 0 lim ε → 0 + ε 1 − α 1 − ( 1 − α ) ln ε ( 1 − α ) 2 − 1 ( 1 − α ) 2 = 1 ( 1 − α ) 2 lim ε → 0 + 1 − ( 1 − α ) ln ε 1 ε 1 − α − 1 ( 1 − α ) 2 = ↓ לופיטל 1 ( 1 − α ) 2 lim ε → 0 + ( 1 − α ) 1 ε − ( 1 − α ) ε − α ε ( 1 − α ) 2 − 1 ( 1 − α ) 2 = 1 1 − α lim ε → 0 + 1 ε α − 1 ε 1 − α + α 2 − 1 ( 1 − α ) 2 = 1 1 − α lim ε → 0 + ε α 2 − α α − 1 − 1 ( 1 − α ) 2 = ↓ α 2 > α ⇓ α 2 − α > 1 + הפונקציה אלמנטרית − 1 ( α − 1 ) 2 ⋅ 0 − 1 ( 1 − α ) 2 = − 1 ( 1 − α ) 2 \begin{align*}
\int_{0}^1 \frac{\ln x}{x^\alpha}dx&=\lim_{ \epsilon \to 0^+ } \int_{\epsilon}^1 \frac{\ln x}{x^\alpha}dx \underset{\begin{matrix}
\downarrow \\
\R{\text{אינטגרציה בחלקים}} \\
\begin{pmatrix}
u=\ln x\implies u'=\frac{1}{x} \\
v'=\frac{1}{x^\alpha}\implies v=\frac{1}{(1-\alpha)x^{\alpha-1}}
\end{pmatrix}
\end{matrix}}{=} \lim_{ \epsilon \to 0^+ } \left( \left. \frac{\ln x}{(1-\alpha)x^{\alpha-1}} \right\vert_{\epsilon}^1-\int_{\epsilon}^1 \frac{1}{(1-\alpha)x^\alpha}dx \right) \\
&=\lim_{ \epsilon \to 0^+ } \left( \left. \frac{\ln x}{(1-\alpha)x^{\alpha-1}} \right\vert_{\epsilon}^1-\left. \frac{1}{(1-\alpha)^2x^{\alpha-1}} \right\vert_{\epsilon}^1 \right) \\
&=\lim_{ \epsilon \to 0^+ } \left( \frac{-\ln\epsilon}{(1-\alpha)\epsilon^{\alpha-1}}-\frac{1}{(1-\alpha)^2}+\frac{1}{(1-\alpha)^2\epsilon^{\alpha-1}} \right)=\lim_{ \epsilon \to 0^+ } \left( \frac{1-(1-\alpha)\ln\epsilon}{(1-\alpha)^2\epsilon^{\alpha-1}}-\frac{1}{(1-\alpha)^2} \right) \\
&=\lim_{ \epsilon \to 0^+ } \frac{1-(1-\alpha)\ln\epsilon}{(1-\alpha)^2\epsilon^{\alpha-1}}-\frac{1}{(1-\alpha)^2} \underset{\begin{matrix}
\downarrow \\
\alpha-1<0
\end{matrix}}{=} \lim_{ \epsilon \to 0^+ } \epsilon^{1-\alpha} \frac{1-(1-\alpha)\ln\epsilon}{(1-\alpha)^2}-\frac{1}{(1-\alpha)^2} \\
&=\frac{1}{(1-\alpha)^2}\lim_{ \epsilon \to 0^+ } \frac{1-(1-\alpha)\ln\epsilon}{\frac{1}{\epsilon^{1-\alpha}}}-\frac{1}{(1-\alpha)^2} \!\!\underset{\begin{matrix}
\downarrow \\
\R{\text{לופיטל}}
\end{matrix}}{=}\!\! \frac{1}{(1-\alpha)^2}\lim_{ \epsilon \to 0^+ } (1-\alpha) \frac{\frac{1}{\epsilon}}{-\frac{(1-\alpha)\epsilon^{-\alpha}}{\epsilon^{(1-\alpha)^2}}}-\frac{1}{(1-\alpha)^2} \\
&= \frac{1}{1-\alpha} \lim_{ \epsilon \to 0^+ } \frac{\frac{1}{\epsilon}}{\frac{\alpha-1}{\epsilon^{1-\alpha+\alpha^2}}}-\frac{1}{(1-\alpha)^2}=\frac{1}{1-\alpha}\lim_{ \epsilon \to 0^+ } \frac{\epsilon^{\alpha^2-\alpha}}{\alpha-1}-\frac{1}{(1-\alpha)^2} \\
&\!\!\!\!\!\!\!\!\!\!\! \underset{\begin{matrix}
\downarrow \\
\alpha^2> \alpha \\
\Downarrow \\
\alpha^2-\alpha>1 \\
+ \\
\R{\text{הפונקציה אלמנטרית}}
\end{matrix}}{=} \!\!\!\!\!\!\!\!\!\!\!\! \frac{-1}{(\alpha-1)^2} \cdot 0-\frac{1}{(1-\alpha)^2}=-\frac{1}{(1-\alpha)^2}
\end{align*} 0 ∫ 1 x α ln x d x = ε → 0 + lim ε ∫ 1 x α ln x d x ↓ אינטגרציה בחלקים ( u = ln x ⟹ u ′ = x 1 v ′ = x α 1 ⟹ v = ( 1 − α ) x α − 1 1 ) = ε → 0 + lim ( 1 − α ) x α − 1 ln x ε 1 − ε ∫ 1 ( 1 − α ) x α 1 d x = ε → 0 + lim ( ( 1 − α ) x α − 1 ln x ε 1 − ( 1 − α ) 2 x α − 1 1 ε 1 ) = ε → 0 + lim ( ( 1 − α ) ε α − 1 − ln ε − ( 1 − α ) 2 1 + ( 1 − α ) 2 ε α − 1 1 ) = ε → 0 + lim ( ( 1 − α ) 2 ε α − 1 1 − ( 1 − α ) ln ε − ( 1 − α ) 2 1 ) = ε → 0 + lim ( 1 − α ) 2 ε α − 1 1 − ( 1 − α ) ln ε − ( 1 − α ) 2 1 ↓ α − 1 < 0 = ε → 0 + lim ε 1 − α ( 1 − α ) 2 1 − ( 1 − α ) ln ε − ( 1 − α ) 2 1 = ( 1 − α ) 2 1 ε → 0 + lim ε 1 − α 1 1 − ( 1 − α ) ln ε − ( 1 − α ) 2 1 ↓ לופיטל = ( 1 − α ) 2 1 ε → 0 + lim ( 1 − α ) − ε ( 1 − α ) 2 ( 1 − α ) ε − α ε 1 − ( 1 − α ) 2 1 = 1 − α 1 ε → 0 + lim ε 1 − α + α 2 α − 1 ε 1 − ( 1 − α ) 2 1 = 1 − α 1 ε → 0 + lim α − 1 ε α 2 − α − ( 1 − α ) 2 1 ↓ α 2 > α ⇓ α 2 − α > 1 + הפונקציה אלמנטרית = ( α − 1 ) 2 − 1 ⋅ 0 − ( 1 − α ) 2 1 = − ( 1 − α ) 2 1
לכן, לפי הגדרה, האינטגרל מתכנס עבור 0 < α < 1 0<\alpha<1 0 < α < 1 .
∎
תרגיל 2
נניח ש-f ( x ) f(x) f ( x ) רציפה ואי-שלילית בקרן [ 1 , ∞ ) [1,\infty) [ 1 , ∞ ) וכי \displaystyle \int_{1}^\infty f(x)dx\end{gather}<\infty .
הוכיחו או הפריכו כי בהכרח מתקיים lim x → ∞ f ( x ) = 0 \lim_{ x \to \infty }f(x)=0 lim x → ∞ f ( x ) = 0 אם נתון ש-f ( x ) > 0 f(x)>0 f ( x ) > 0 לכל x x x .
הטענה לא נכונה .
נגדיר את הפונקציה f ( x ) = 1 x 2 + A ( x ) f(x)=\frac{1}{x^2}+A(x) f ( x ) = x 2 1 + A ( x ) כאשר A ( x ) A(x) A ( x ) היא פונקציית האוהלים הסטנדרטית.
הפונקציה חיובית לכל x ∈ [ 1 , ∞ ) x \in[1,\infty) x ∈ [ 1 , ∞ ) שכן פונקצית האוהלים אי שלילית ו-1 x 2 \frac{1}{x^2} x 2 1 חיובית. בנוסף, הפונקציה רציפה שכן היא חיבור של פונקציות רציפות.
∫ 1 ∞ f ( x ) d x = lim M → ∞ ∫ 1 M f ( x ) d x = lim M → ∞ ∫ 1 M 1 x 2 d x + lim M → ∞ ∫ 1 M A ( x ) d x \int_{1}^\infty f(x)dx=\lim_{ M \to \infty } \int_{1}^M f(x)dx=\lim_{ M \to \infty } \int_{1}^M \frac{1}{x^2}dx+\lim_{ M \to \infty } \int_{1}^M A(x)dx 1 ∫ ∞ f ( x ) d x = M → ∞ lim 1 ∫ M f ( x ) d x = M → ∞ lim 1 ∫ M x 2 1 d x + M → ∞ lim 1 ∫ M A ( x ) d x
אינטגרל פונקציית האוהלים מתכנס כפי שנראה בהרצאה ולפי מסקנה האינטגרל של 1 x 2 \frac{1}{x^2} x 2 1 מתכנס ולכן האינטגרל מתכנס.
אך לאור פונקציית האוהלים, הפונקציה f f f אינה מתכנסת ב-∞ \infty ∞ .
נניח ש-f ( x ) f(x) f ( x ) רציפה ואי-שלילית בקרן [ 1 , ∞ ) [1,\infty) [ 1 , ∞ ) וכי \begin{gather}\int_{1}^\infty f(x)dx\end{gather}<\infty .
הוכיחו או הפריכו כי בהכרח מתקיים \begin{gather}\lim_{ x \to \infty }f(x)=0 אם נתון ש-f ( x ) f(x) f ( x ) מונוטונית.
הטענה נכונה .
הוכחה:
נתון כי f f f רציפה ומונוטונית ולכן, לפי משפט, היא מתכנסת במובן הרחב או לגבול סופי.
נתון כי f f f אי שלילית ולכן, לפי יחס סדר גבולות lim x → ∞ f ( x ) = K ≥ 0 \lim_{ x \to \infty }f(x)=K\geq 0 lim x → ∞ f ( x ) = K ≥ 0 .
נניח בשלילה כי K > 0 K>0 K > 0 . אזי, לפי יחס סדר גבולות, קיים C > 0 C>0 C > 0 שעבורו קיים L > 1 L>1 L > 1 כך שלכל x ≥ L x\geq L x ≥ L מתקיים f ( x ) > C f(x)> C f ( x ) > C .
∫ 1 ∞ f ( x ) d x = lim M → ∞ ∫ 1 M f ( x ) d x = ∫ 1 L f ( x ) d x + lim M → ∞ ∫ L M f ( x ) d x \int_{1}^\infty f(x)dx=\lim_{ M \to \infty } \int_{1}^M f(x)dx= \int_{1}^L f(x)dx+\lim_{ M\to \infty } \int_{L}^M f(x)dx 1 ∫ ∞ f ( x ) d x = M → ∞ lim 1 ∫ M f ( x ) d x = 1 ∫ L f ( x ) d x + M → ∞ lim L ∫ M f ( x ) d x
לכל x ≥ L x\geq L x ≥ L מתקיים f ( x ) > C f(x)> C f ( x ) > C ולכן, לפי מונוטוניות האינטגרל ויחס סדר גבולות מתקיים:
∞ = lim M → ∞ K ( M − L ) = ↓ N L lim M → ∞ ∫ L M C d x < lim M → ∞ ∫ L M f ( x ) d x \infty =\lim_{ M \to \infty } K(M-L) \underset{\begin{matrix}
\downarrow \\
NL
\end{matrix}}{=}\lim_{ M \to \infty } \int_{L}^M Cdx< \lim_{ M \to \infty } \int_{L}^M f(x)dx ∞ = M → ∞ lim K ( M − L ) ↓ N L = M → ∞ lim L ∫ M C d x < M → ∞ lim L ∫ M f ( x ) d x
לכן, מתקיים כי ∞ < ∫ 1 ∞ f ( x ) d x \displaystyle\infty < \int_{1}^\infty f(x)dx ∞ < 1 ∫ ∞ f ( x ) d x בסתירה לכך ש-\displaystyle \int_{1}^\infty f(x)dx\end{gather}<\infty .
לכן, lim x → ∞ f ( x ) = 0 \lim_{ x \to \infty }f(x)=0 lim x → ∞ f ( x ) = 0 .
∎
נניח ש-f ( x ) f(x) f ( x ) רציפה ואי-שלילית בקרן [ 1 , ∞ ) [1,\infty) [ 1 , ∞ ) וכי \begin{gather}\int_{1}^\infty f(x)dx\end{gather}<\infty .
הוכיחו או הפריכו כי בהכרח מתקיים \begin{gather}\lim_{ x \to \infty }f(x)=0 אם נתון ש-f ( x ) f(x) f ( x ) רציפה במ"ש.
הדרכה: הטענה נכונה. כדי להוכיח אותה, הניחו בשלילה שהגבול אינו אפס ובנו סדרת נקודות x n x_{n} x n השואפת ל-∞ \infty ∞ שמתאימה להנחת השלילה. השתמשו ברציפות במ"ש בכל קטע [ x n , x n + 1 ] [x_{n},x_{n}+1] [ x n , x n + 1 ] ובהנחת השלילה כדי להגיע לסתירה עם הנתון.
הטענה נכונה .
הוכחה:
נניח בשלילה כי lim x → ∞ f ( x ) ≠ 0 \lim_{ x \to \infty }f(x)\neq 0 lim x → ∞ f ( x ) = 0 .
לכן, קיימות ∞ \infty ∞ נקודות של f f f שאינן בסביבה של 0 0 0 . נתון כי הפונקציה אי שלילית ולכן, כל נקודה שאינה בסביבה של 0 0 0 בהכרח מקיימת f > 0 f>0 f > 0 .
יהי ε 0 > 0 \epsilon_{0}>0 ε 0 > 0 כך שקיימות ∞ \infty ∞ נק' x x x של f f f שאינן בסביבה של 0 0 0 המקיימות f ( x ) > ε 0 f(x)>\epsilon_{0} f ( x ) > ε 0 . זה בהכרח מתקיים לפי הנ"ל.
נבנה מהנקודות האלו סדרה לפי סדר הופעתם בפונקציה(כלומר, לכל x 1 > x 2 x_{1}>x_{2} x 1 > x 2 יתקיים ש-x 1 x_{1} x 1 יופיע אחרי x 2 x_{2} x 2 בסדרה x n x_{n} x n )
מכיוון שזוהי סדרה מונוטונית עולה ולא חסומה מתקיים lim n → ∞ x n = ∞ \lim_{ n \to \infty }x_{n}=\infty lim n → ∞ x n = ∞ ובנוסף מתקיים f ( x n ) > ε 0 f(x_{n})> \epsilon_{0} f ( x n ) > ε 0 לכל n ∈ N n \in \mathbb{N} n ∈ N .
נתון כי f f f רציפה במ"ש ב-R \mathbb{R} R . לכן, לכל ε > 0 \epsilon>0 ε > 0 קיימת δ > 0 \delta>0 δ > 0 כך שלכל x , y ∈ R x,y \in \mathbb{R} x , y ∈ R המקיימים ∣ x − y ∣ < δ |x-y|<\delta ∣ x − y ∣ < δ מתקיים: ∣ f ( x ) − f ( y ) ∣ < ε |f(x)-f(y)|<\epsilon ∣ f ( x ) − f ( y ) ∣ < ε .
זה מתקיים בפרט עבור y = x n , ε = ε 0 2 , δ > 1 y=x_{n},\epsilon= \frac{\epsilon_{0}}{2},\delta>1 y = x n , ε = 2 ε 0 , δ > 1 לכל n ∈ N n \in \mathbb{N} n ∈ N ואז:
− ε 0 2 < f ( x ) − f ( x n ) < ε 0 2 ⟹ f ( x n ) − ε 0 2 < f ( x ) < f ( x n ) + ε 0 2 -\frac{\epsilon_{0}}{2}< f(x)-f(x_{n})< \frac{\epsilon_{0}}{2} \implies f(x_{n})-\frac{\epsilon_{0}}{2}< f(x)< f(x_{n})+\frac{\epsilon_{0}}{2} − 2 ε 0 < f ( x ) − f ( x n ) < 2 ε 0 ⟹ f ( x n ) − 2 ε 0 < f ( x ) < f ( x n ) + 2 ε 0
⟹ ε 0 − ε 0 2 < f ( x n ) − ε 0 2 < f ( x ) < f ( x n ) + ε 0 2 \implies {\epsilon_{0}}-\frac{\epsilon_{0}}{2}<f(x_{n})-\frac{\epsilon_{0}}{2}<f(x)< f(x_{n})+\frac{\epsilon_{0}}{2} ⟹ ε 0 − 2 ε 0 < f ( x n ) − 2 ε 0 < f ( x ) < f ( x n ) + 2 ε 0
נביט ב-x ∈ [ x n , x n + 1 ] x \in[x_{n},x_{n}+1] x ∈ [ x n , x n + 1 ] לכל n ∈ N n \in \mathbb{N} n ∈ N . בקטע זה מתקיים: f ( x ) > ε 0 2 f(x)> \frac{\epsilon_{0}}{2} f ( x ) > 2 ε 0 .
∫ 1 ∞ f ( x ) d x = lim M → ∞ ∫ 1 M f ( x ) d x \int_{1}^\infty f(x)dx =\lim_{ M \to \infty } \int_{1}^M f(x)dx 1 ∫ ∞ f ( x ) d x = M → ∞ lim 1 ∫ M f ( x ) d x
לפי אדטיביות האינטגרל המסויים נוכל לפרק את האינטגרל לכל החלקים של הסדרה כך:
lim M → ∞ ∫ 1 M f ( x ) d x = lim M → ∞ ( ∫ 1 x 1 f ( x ) d x + ∫ x 1 x 1 + 1 f ( x ) d x + ∫ x 1 + 1 x 2 f ( x ) d x + ∫ x 2 x 2 + 1 f ( x ) d x + ⋯ ) \lim_{ M \to \infty } \int_{1}^M f(x)dx=\lim_{ M \to \infty } \left( \int_{1}^{x_{1}}f(x)dx+\int_{x_{1}}^{x_{1}+1}f(x)dx+\int_{x_{1}+1}^{x_{2}}f(x)dx+\int_{x_{2}}^{x_{2}+1}f(x)dx+\cdots \right) M → ∞ lim 1 ∫ M f ( x ) d x = M → ∞ lim 1 ∫ x 1 f ( x ) d x + x 1 ∫ x 1 + 1 f ( x ) d x + x 1 + 1 ∫ x 2 f ( x ) d x + x 2 ∫ x 2 + 1 f ( x ) d x + ⋯
נזיז את כל אלו שגבולותיהם אינם [ x n , x n + 1 ] [x_{n},x_{n}+1] [ x n , x n + 1 ] להתחלה ונשאיר את השאר בסוף. נסמן ב-K K K את ערך חיבור האינטגרלים הזה ואז, לפי מונוטויות האינטגרל המסויים ויחס סדר גבולות מתקיים:
lim M → ∞ ( K + ∫ x 1 x 1 + 1 f ( x ) d x + ∫ x 2 x 2 + 1 f ( x ) d x + ⋯ ) > lim M → ∞ ( K + ∫ x 1 x 1 + 1 ε 0 2 d x + ∫ x 2 x 2 + 1 ε 0 2 d x + ⋯ ) = ↓ N L \lim_{ M \to \infty } \left( K+\int_{x_{1}}^{x_{1}+1}f(x)dx+\int_{x_{2}}^{x_{2}+1}f(x)dx+\cdots \right)>\lim_{ M \to \infty } \left( K+ \int_{x_{1}}^{x_{1}+1} \frac{\epsilon_{0}}{2}dx+\int_{x_{2}}^{x_{2}+1} \frac{\epsilon_{0}}{2}dx+\cdots \right) \underset{\begin{matrix}
\downarrow \\
NL
\end{matrix}}{=} M → ∞ lim K + x 1 ∫ x 1 + 1 f ( x ) d x + x 2 ∫ x 2 + 1 f ( x ) d x + ⋯ > M → ∞ lim K + x 1 ∫ x 1 + 1 2 ε 0 d x + x 2 ∫ x 2 + 1 2 ε 0 d x + ⋯ ↓ N L =
lim M → ∞ ( K + ε 0 2 ( x 1 + 1 − x 1 ) + ε 0 2 ( x 2 + 1 − x 2 ) + ⋯ ) = lim M → ∞ ( K + ε 0 2 + ε 0 2 + ⋯ ) \lim_{ M \to \infty } \left( K+\frac{\epsilon_{0}}{2}(x_{1}+1-x_{1})+\frac{\epsilon_{0}}{2}(x_{2}+1-x_{2})+\cdots \right)=\lim_{ M \to \infty } \left( K+ \frac{\epsilon_{0}}{2}+\frac{\epsilon_{0}}{2}+\cdots \right) M → ∞ lim ( K + 2 ε 0 ( x 1 + 1 − x 1 ) + 2 ε 0 ( x 2 + 1 − x 2 ) + ⋯ ) = M → ∞ lim ( K + 2 ε 0 + 2 ε 0 + ⋯ )
יש לנו n n n איברים שהם ε 0 2 \frac{\epsilon_{0}}{2} 2 ε 0 . בנוסף K ≥ 0 K\geq 0 K ≥ 0 שכן הפונקציה אי שלילית ולכן:
≥ lim M → ∞ lim n → ∞ ∑ i = 1 n ε 0 2 = lim n → ∞ ε 0 2 ∑ i = 1 n 1 = lim n → ∞ n ε 0 2 = ∞ \geq \lim_{ M \to \infty } \lim_{ n \to \infty } \sum_{i=1}^n\frac{\epsilon_{0}}{2}=\lim_{ n \to \infty } \frac{\epsilon_{0}}{2} \sum_{i=1}^n 1=\lim_{ n \to \infty } n \frac{\epsilon_{0}}{2}=\infty ≥ M → ∞ lim n → ∞ lim i = 1 ∑ n 2 ε 0 = n → ∞ lim 2 ε 0 i = 1 ∑ n 1 = n → ∞ lim n 2 ε 0 = ∞
לכן, לפי מבחן ההשוואה לפונקציות אי שליליות, האינטגרל המוכלל מתבדר בסתירה לכך שנתון שהוא מתכנס.
לכן, lim x → ∞ f ( x ) = 0 \lim_{ x \to \infty }f(x)=0 lim x → ∞ f ( x ) = 0 .
∎
נניח ש-f ( x ) f(x) f ( x ) רציפה ואי-שלילית בקרן [ 1 , ∞ ) [1,\infty) [ 1 , ∞ ) וכי \displaystyle \int_{1}^\infty f(x)dx\end{gather}<\infty .
הוכיחו או הפריכו כי בהכרח מתקיים \begin{gather}\lim_{ x \to \infty }f(x)=0 אם נתון ש-f f f גזירה ברציפות ו-∫ 1 ∞ f ′ d x \int_{1}^\infty f'dx 1 ∫ ∞ f ′ d x מתכנס.
הטענה נכונה .
הוכחה:
נתון כי f ′ f' f ′ רציפה ובנוסף כי ∫ 1 ∞ f ′ ( x ) d x \displaystyle \int_{1}^\infty f'(x)dx 1 ∫ ∞ f ′ ( x ) d x מתכנס. כלומר, קיים L ∈ R L \in \mathbb{R} L ∈ R כך ש-∫ 1 ∞ f ′ ( x ) = L \displaystyle \int_{1}^\infty f'(x)=L 1 ∫ ∞ f ′ ( x ) = L לכן:
∫ 1 ∞ f ′ ( x ) d x = lim M → ∞ ∫ 1 M f ′ ( x ) d x = ↓ N L lim M → ∞ ( f ( M ) − f ( 1 ) ) = lim M → ∞ f ( M ) − f ( 1 ) = L \int_{1}^\infty f'(x)dx=\lim_{ M \to \infty } \int_{1}^M f'(x)dx \underset{\begin{matrix}
\downarrow \\
NL
\end{matrix}}{=}\lim_{ M \to \infty } (f(M)-f(1))=\lim_{ M \to \infty } f(M)-f(1)=L 1 ∫ ∞ f ′ ( x ) d x = M → ∞ lim 1 ∫ M f ′ ( x ) d x ↓ N L = M → ∞ lim ( f ( M ) − f ( 1 )) = M → ∞ lim f ( M ) − f ( 1 ) = L
לכן, lim x → ∞ f ( x ) = L + f ( 1 ) \lim_{ x \to \infty }f(x)=L+f(1) lim x → ∞ f ( x ) = L + f ( 1 ) . נסמן L + f ( 1 ) = C ∈ R L+f(1)=C \in \mathbb{R} L + f ( 1 ) = C ∈ R .
נתון כי f f f אי שלילית ולכן, לפי יחס סדר גבולות, C ≥ 0 C\geq 0 C ≥ 0 .
נניח בשלילה כי C > 0 C>0 C > 0 . אזי, קיים 0 < K < C 0<K<C 0 < K < C עבורו קיים M > 0 M>0 M > 0 כך שעבור כל x ≥ M x\geq M x ≥ M מתקיים f ( x ) > K f(x)>K f ( x ) > K .
∫ 1 ∞ f ( x ) d x = ∫ 1 M f ( x ) d x + ∫ M ∞ f ( x ) d x = ∫ 1 M f ( x ) d x + lim S → ∞ ∫ M S f ( x ) d x \int_{1}^\infty f(x)dx=\int_{1}^M f(x)dx+\int_{M}^\infty f(x)dx=\int_{1}^M f(x)dx+\lim_{ S \to \infty } \int_{M}^S f(x)dx 1 ∫ ∞ f ( x ) d x = 1 ∫ M f ( x ) d x + M ∫ ∞ f ( x ) d x = 1 ∫ M f ( x ) d x + S → ∞ lim M ∫ S f ( x ) d x
נתון כי f ( x ) > K f(x)>K f ( x ) > K לכל x ≥ M x\geq M x ≥ M ולכן, לפי יחס סדר גבולות ומונוטוניות האינטגרל המסויים מתקיים:
∞ = lim S → ∞ K ( S − M ) = ↓ N L lim S → ∞ ∫ M S K d x < lim S → ∞ ∫ M S f ( x ) d x \infty=\lim_{ S \to \infty } K(S-M)\underset{\begin{matrix}
\downarrow \\
NL
\end{matrix}}{=}\lim_{ S \to \infty } \int_{M}^S Kdx < \lim_{ S \to \infty } \int_{M}^S f(x)dx ∞ = S → ∞ lim K ( S − M ) ↓ N L = S → ∞ lim M ∫ S K d x < S → ∞ lim M ∫ S f ( x ) d x
לכן מתקיים כי ∫ 1 ∞ f ( x ) d x > ∞ \displaystyle \int_{1}^\infty f(x)dx> \infty 1 ∫ ∞ f ( x ) d x > ∞ בסתירה לכך ש-\displaystyle \int_{1}^\infty f(x)dx\end{gather}<\infty ולכן, C = 0 C=0 C = 0 . כלומר:
lim x → ∞ f ( x ) = 0 \lim_{ x \to \infty } f(x)=0 x → ∞ lim f ( x ) = 0
∎
תרגיל 3
הוכיחו את קריטריון קושי להתכנסות אינטגרלים מוכללים:
תהי f : [ a , ∞ ] → R f:[a,\infty]\to \mathbb{R} f : [ a , ∞ ] → R פונקציה אינטגרבילית רימן בקטע [ a , b ] [a,b] [ a , b ] לכל b > a b>a b > a .
אז, ∫ a ∞ f \displaystyle \int_{a}^\infty f a ∫ ∞ f מתכנס ⟺ \iff ⟺ לכל ε > 0 \epsilon>0 ε > 0 קיים M > a M>a M > a כך שלכל x , y > M x,y > M x , y > M מתקיים ∣ ∫ x y f ( t ) d t ∣ < ε \displaystyle \left| \int_{x}^y f(t)dt \right|<\epsilon x ∫ y f ( t ) d t < ε .
הוכחה:
⟸ \impliedby ⟸ : יהי ε > 0 \epsilon>0 ε > 0 .
נתון כי ∫ a ∞ f \displaystyle \int_{a}^\infty f a ∫ ∞ f מתכנס ולכן קיים L ∈ R L \in \mathbb{R} L ∈ R כך ש-∫ a ∞ f = L \displaystyle \int_{a}^\infty f=L a ∫ ∞ f = L . כלומר, lim K → ∞ ∫ a K f = L \displaystyle\lim_{ K \to \infty } \int_{a}^K f=L K → ∞ lim a ∫ K f = L .
לפי הגדרת הגבול, לכל ε 1 > 0 \epsilon_{1}>0 ε 1 > 0 קיימת δ > 0 \delta>0 δ > 0 כך שלכל K > δ K>\delta K > δ מתקיים:
∣ ∫ a K f − L ∣ < ε 1 ⟹ L − ε 1 < ∫ a K f < ε 1 + L \left| \int_{a}^K f-L \right|<\epsilon_{1} \implies L-\epsilon_{1}<\int_{a}^K f< \epsilon_{1}+L a ∫ K f − L < ε 1 ⟹ L − ε 1 < a ∫ K f < ε 1 + L
זה מתקיים בפרט עבור ε 1 = ε 2 > 0 \epsilon_{1}=\frac{\epsilon}{2}>0 ε 1 = 2 ε > 0 .
נקח x , y > δ x,y>\delta x , y > δ אזי:
∣ ∫ x y f ∣ = ↓ אדיטיביות ∣ ∫ a x f − ∫ a y f ∣ = ∣ ∫ a x f − L + L − ∫ a y f ∣ ≤ ↓ אי שוויון המשולש ∣ ∫ a x f − L ∣ + ∣ ∫ a y f − L ∣ < 2 ε 1 = ε \left| \int_{x}^y f \right|\underset{\begin{matrix}
\downarrow \\
\R{\text{אדיטיביות}}
\end{matrix}}{=}\left| \int_{a}^x f-\int_{a}^y f \right|=\left| \int_{a}^x f-L +L-\int_{a}^y f \right|\underset{\begin{matrix}
\downarrow \\
\R{\text{אי שוויון המשולש}}
\end{matrix}}{\leq} \left| \int_{a}^x f-L \right|+\left| \int_{a}^y f-L \right|< 2\epsilon_{1} =\epsilon x ∫ y f ↓ אדיטיביות = a ∫ x f − a ∫ y f = a ∫ x f − L + L − a ∫ y f ↓ אי שוויון המשולש ≤ a ∫ x f − L + a ∫ y f − L < 2 ε 1 = ε
⟹ \implies ⟹ : נתון כי לכל ε > 0 \epsilon>0 ε > 0 קיים M > a M>a M > a כך שלכל x , y > M x,y >M x , y > M מתקיים ∣ ∫ x y f ( t ) d t ∣ < ε \displaystyle \left| \int_{x}^y f(t)dt \right|<\epsilon x ∫ y f ( t ) d t < ε .
נגדיר F ( x ) = ∫ a x f ( t ) d t \displaystyle F(x)=\int_{a}^x f(t)dt F ( x ) = a ∫ x f ( t ) d t ואז:
∣ F ( x ) − F ( y ) ∣ = ∣ ∫ a x f ( t ) d t − ∫ a y f ( t ) d t ∣ = ∣ ∫ x y f ( t ) d t ∣ < ε \left| F(x)-F(y) \right|=\left| \int_{a}^x f(t)dt-\int_{a}^y f(t)dt \right|=\left| \int_{x}^y f(t)dt \right| < \epsilon ∣ F ( x ) − F ( y ) ∣ = a ∫ x f ( t ) d t − a ∫ y f ( t ) d t = x ∫ y f ( t ) d t < ε
לפי תנאי קושי לפונקציות, F ( x ) F(x) F ( x ) מתכנסת. כלומר, קיים L ∈ R L \in \mathbb{R} L ∈ R כך ש:
lim x → ∞ F ( x ) = L ⟹ lim x → ∞ ∫ a x f ( t ) d t = L \lim_{ x \to \infty } F(x)=L \implies \lim_{ x \to \infty } \int_{a}^x f(t)dt = L x → ∞ lim F ( x ) = L ⟹ x → ∞ lim a ∫ x f ( t ) d t = L
לכן, לפי הגדרה, ∫ a ∞ f ( t ) d t \displaystyle \int_{a}^\infty f(t)dt a ∫ ∞ f ( t ) d t מתכנס.
∎
תהי f f f פונקציה חיובית מונוטונית יורדת ב-[ 2 , ∞ ) [2,\infty) [ 2 , ∞ ) . הוכיחו או הפריכו:
אם \begin{gather}\int_{2}^\infty f מתכנס אז lim x → ∞ x f ( x ) = 0 \lim_{ x \to \infty }xf(x)=0 lim x → ∞ x f ( x ) = 0 .
אם lim x → ∞ x f ( x ) = 0 \lim_{ x \to \infty }xf(x)=0 lim x → ∞ x f ( x ) = 0 אז ∫ 2 ∞ f \displaystyle \int_{2}^\infty f 2 ∫ ∞ f מתכנס.
הטענה (1) נכונה .
הוכחה:
נתון כי ∫ 2 ∞ f ( x ) d x \displaystyle \int_{2}^\infty f(x)dx 2 ∫ ∞ f ( x ) d x מתכנס. לכן, קיים L ∈ R L \in \mathbb{R} L ∈ R כך ש-lim M → ∞ ∫ 2 M f ( x ) d x = L \displaystyle \lim_{ M \to \infty }\int_{2}^M f(x)dx=L M → ∞ lim 2 ∫ M f ( x ) d x = L .
L = lim x → ∞ ∫ 2 x f ( t ) d t = lim x → ∞ x ∫ 2 x f ( t ) d t x = ↓ לופיטל lim x → ∞ ∫ 2 x f ( t ) d t + x f ( x ) 1 = L + lim x → ∞ x f ( x ) L=\lim_{ x \to \infty } \int_{2}^x f(t)dt= \lim_{ x \to \infty } \frac{\displaystyle x\int_{2}^x f(t)dt}{x} \underset{\begin{matrix}
\downarrow \\
\R{\text{לופיטל}}
\end{matrix}}{=}\lim_{ x \to \infty } \frac{\displaystyle \int_{2}^x f(t)dt+xf(x)}{1}=L+\lim_{ x \to \infty } xf(x) L = x → ∞ lim 2 ∫ x f ( t ) d t = x → ∞ lim x x 2 ∫ x f ( t ) d t ↓ לופיטל = x → ∞ lim 1 2 ∫ x f ( t ) d t + x f ( x ) = L + x → ∞ lim x f ( x )
⟹ lim x → ∞ x f ( x ) = 0 \implies \lim_{ x \to \infty } xf(x)=0 ⟹ x → ∞ lim x f ( x ) = 0
∎
הטענה (2) לא נכונה .
נסתכל על הפונקציה: f ( x ) = 1 x ln x f(x)=\frac{1}{x\ln x} f ( x ) = x l n x 1 .
נשים לב כי ln x \ln x ln x שואף ל-∞ \infty ∞ כאשר x → ∞ x\to \infty x → ∞ ולכן: lim x → ∞ x f ( x ) = 0 \lim_{ x \to \infty }xf(x)=0 lim x → ∞ x f ( x ) = 0 אך:
∫ 2 ∞ 1 x ln x d x = lim M → ∞ ∫ 2 M d x x ln x = ↓ שיטת ההצבה ( t = ln x d t = 1 x d x ) lim M → ∞ ∫ ln 2 ln M d t t = lim M → ∞ ln t ∣ ln 2 ln M = lim M → ∞ ( ln ln M − ln ln 2 ) = \int_{2}^\infty \frac{1}{x\ln x}dx=\lim_{ M \to \infty } \int_{2}^M \frac{dx}{x\ln x} \underset{\begin{matrix}
\downarrow \\
\R{\text{שיטת ההצבה}} \\
\begin{pmatrix}
t=\ln x \\
dt=\frac{1}{x}dx
\end{pmatrix}
\end{matrix}}{=}\lim_{ M \to \infty } \int_{\ln 2}^{\ln M} \frac{dt}{t}=\lim_{ M \to \infty } \left. \begin{matrix}
\\ \\
\end{matrix} \ln t \right\vert_{\ln 2}^{\ln M}=\lim_{ M \to \infty } (\ln \ln M-\ln \ln 2)= 2 ∫ ∞ x ln x 1 d x = M → ∞ lim 2 ∫ M x ln x d x ↓ שיטת ההצבה ( t = ln x d t = x 1 d x ) = M → ∞ lim l n 2 ∫ l n M t d t = M → ∞ lim ln t l n 2 l n M = M → ∞ lim ( ln ln M − ln ln 2 ) =
= lim M → ∞ ln ln M − ln ln 2 = ∞ − ln ln 2 = ∞ =\lim_{ M \to \infty } \ln \ln M-\ln \ln2=\infty-\ln \ln2=\infty = M → ∞ lim ln ln M − ln ln 2 = ∞ − ln ln 2 = ∞
ולכן האינטגרל אינו מתכנס בסתירה לטענה.
תרגיל 4
נסמן:
I 1 = ∫ 0 ∞ cos x 1 + x d x , I 2 = ∫ 0 ∞ sin x ( 1 + x ) 2 d x I_{1}=\int_{0}^\infty \frac{\cos x}{1+x}dx, \qquad I_{2}=\int_{0}^\infty \frac{\sin x}{(1+x)^2}dx I 1 = 0 ∫ ∞ 1 + x cos x d x , I 2 = 0 ∫ ∞ ( 1 + x ) 2 sin x d x
הוכיחו כי I 1 = I 2 I_{1}=I_{2} I 1 = I 2 .
הוכחה:
I 1 = ∫ 0 ∞ cos x 1 + x d x = lim M → ∞ ∫ 0 M cos x 1 + x d x = ↓ אינטגרציה בחלקים ( u = 1 1 + x ⟹ u ′ = − 1 ( 1 + x ) 2 v ′ = cos x ⟹ v = sin x ) lim M → ∞ ( sin x 1 + x ∣ 0 M + ∫ 0 M sin x ( 1 + x ) 2 d x ) = I_{1}=\int_{0}^\infty \frac{\cos x}{1+x}dx=\lim_{ M \to \infty } \int_{0}^M \frac{\cos x}{1+x}dx \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\underset{\begin{matrix}
\downarrow \\
\R{\text{אינטגרציה בחלקים}} \\
\begin{pmatrix}
u=\frac{1}{1+x} \implies u'=-\frac{1}{(1+x)^2} \\
v'=\cos x \implies v=\sin x
\end{pmatrix}
\end{matrix}}{=}\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! \lim_{ M \to \infty } \left(\left. \frac{\sin x}{1+x} \right\vert_{0}^M+ \int_{0}^M \frac{\sin x}{(1+x)^2}dx \right)= I 1 = 0 ∫ ∞ 1 + x cos x d x = M → ∞ lim 0 ∫ M 1 + x cos x d x ↓ אינטגרציה בחלקים ( u = 1 + x 1 ⟹ u ′ = − ( 1 + x ) 2 1 v ′ = cos x ⟹ v = sin x ) = M → ∞ lim 1 + x sin x 0 M + 0 ∫ M ( 1 + x ) 2 sin x d x =
= lim M → ∞ sin M 1 + M + lim M → ∞ ∫ 0 M sin x ( 1 + x ) 2 d x =\lim_{ M \to \infty } \frac{\sin M}{1+M}+\lim_{ M \to \infty } \int_{0}^M \frac{\sin x}{(1+x)^2}dx = M → ∞ lim 1 + M sin M + M → ∞ lim 0 ∫ M ( 1 + x ) 2 sin x d x
לפי משפט הסנדוויץ עם הפונקציות − 1 1 + M , 1 1 + M -\frac{1}{1+M}, \frac{1}{1+M} − 1 + M 1 , 1 + M 1 , מתקיים lim M → ∞ sin M 1 + M = 0 \lim_{ M \to \infty } \frac{\sin M}{1+M}=0 lim M → ∞ 1 + M s i n M = 0 ולכן:
I 1 = lim M → ∞ ∫ 0 M sin x ( 1 + x ) 2 d x = ∫ 0 ∞ sin x ( 1 + x ) 2 d x = I 2 I_{1}=\lim_{ M \to \infty } \int_{0}^M \frac{\sin x}{(1+x)^2}dx=\int_{0}^\infty \frac{\sin x}{(1+x)^2}dx=I_{2} I 1 = M → ∞ lim 0 ∫ M ( 1 + x ) 2 sin x d x = 0 ∫ ∞ ( 1 + x ) 2 sin x d x = I 2
∎
הראו כי I 2 I_{2} I 2 מתכנס בהחלט בעוד I 1 I_{1} I 1 מתכנס בתנאי.
נסתכל תחילה על I 2 I_{2} I 2 :
הפונקציה ∣ sin x ∣ ( 1 + x ) 2 < 1 ( 1 + x ) 2 \frac{|\sin x|}{(1+x)^2}< \frac{1}{(1+x)^2} ( 1 + x ) 2 ∣ s i n x ∣ < ( 1 + x ) 2 1 ומתקיים:
∫ 0 ∞ 1 ( 1 + x ) 2 d x = lim M → ∞ ∫ 0 M 1 ( 1 + x ) 2 d x = ↓ שיטת ההצבה ( t = 1 + x d t = d x ) lim M → ∞ ∫ 1 M + 1 1 t 2 d t = ↓ N L lim M → ∞ − 1 1 + x ∣ 1 M + 1 = \int_{0}^\infty \frac{1}{(1+x)^2}dx=\lim_{ M \to \infty } \int_{0}^M \frac{1}{(1+x)^2}dx \underset{\begin{matrix}
\downarrow \\
\R{\text{שיטת ההצבה}} \\
\begin{pmatrix}
t=1+x \\
dt=dx
\end{pmatrix}
\end{matrix}}{=} \lim_{ M \to \infty } \int_{1}^{M+1} \frac{1}{t^2}dt \underset{\begin{matrix}
\downarrow \\
NL
\end{matrix}}{=}\lim_{ M \to \infty } \left. -\frac{1}{1+x} \right\vert_{1}^{M+1}= 0 ∫ ∞ ( 1 + x ) 2 1 d x = M → ∞ lim 0 ∫ M ( 1 + x ) 2 1 d x ↓ שיטת ההצבה ( t = 1 + x d t = d x ) = M → ∞ lim 1 ∫ M + 1 t 2 1 d t ↓ N L = M → ∞ lim − 1 + x 1 1 M + 1 =
= lim M → ∞ ( 1 2 − 1 M + 2 ) = 1 2 − 0 = 1 2 =\lim_{ M \to \infty } \left( \frac{1}{2}-\frac{1}{M+2} \right)=\frac{1}{2}-0=\frac{1}{2} = M → ∞ lim ( 2 1 − M + 2 1 ) = 2 1 − 0 = 2 1
בנוסף שתי הפונקציות אי שליליות ולכן, לפי מבחן ההשוואה לפונקציות אי שלילות מתקיים כי I 2 I_{2} I 2 מתכנס בהחלט.
לכן, לפי משפט, I 2 I_{2} I 2 מתכנס ולכן, לפי סעיף א', I 1 I_{1} I 1 גם כן מתכנס.
נראה כעת כי I 1 I_{1} I 1 מתכנס בתנאי:
lim x → ∞ ∣ cos x ∣ x + 1 ∣ cos x ∣ x = lim x → ∞ x x + 1 = lim x → ∞ 1 1 = 1 \lim_{ x \to \infty } \frac{\frac{|\cos x|}{x+1}}{\frac{|\cos x|}{x}}=\lim_{ x \to \infty } \frac{x}{x+1}=\lim_{ x \to \infty } \frac{1}{1}=1 x → ∞ lim x ∣ c o s x ∣ x + 1 ∣ c o s x ∣ = x → ∞ lim x + 1 x = x → ∞ lim 1 1 = 1
אנו יודעים כי 0 ≤ ∣ cos x ∣ ≤ 1 0\leq |\cos x|\leq 1 0 ≤ ∣ cos x ∣ ≤ 1 ולכן cos 2 x ≤ ∣ cos x ∣ \cos^2x \leq |\cos x| cos 2 x ≤ ∣ cos x ∣ . לכן, 0 ≤ cos 2 x x ≤ ∣ cos x ∣ x 0\leq \frac{\cos^2x}{x}\leq \frac{|\cos x|}{x} 0 ≤ x c o s 2 x ≤ x ∣ c o s x ∣ . בנוסף מתקיים:
∫ 0 ∞ cos 2 x x d x = ∫ 0 ∞ 1 + cos 2 x 2 x d x = ∫ 0 ∞ ( 1 2 x + cos 2 x 2 x ) d x = \int_{0}^\infty \frac{\cos^2x}{x}dx=\int_{0}^\infty \frac{1+\cos 2x}{2x}dx=\int_{0}^\infty \left( \frac{1}{2x}+ \frac{\cos 2x}{2x} \right)dx= 0 ∫ ∞ x cos 2 x d x = 0 ∫ ∞ 2 x 1 + cos 2 x d x = 0 ∫ ∞ ( 2 x 1 + 2 x cos 2 x ) d x =
= lim ε → 0 + ∫ ε 1 ( 1 2 x + cos 2 x 2 x ) d x + lim M → ∞ ∫ 1 M ( 1 2 x + cos 2 x 2 x ) d x =\lim_{ \epsilon \to 0^+ } \int_{\epsilon}^1 \left( \frac{1}{2x}+\frac{\cos2x}{2x} \right)dx+ \lim_{ M \to \infty } \int_{1}^M \left( \frac{1}{2x}+\frac{\cos2x}{2x} \right)dx = ε → 0 + lim ε ∫ 1 ( 2 x 1 + 2 x cos 2 x ) d x + M → ∞ lim 1 ∫ M ( 2 x 1 + 2 x cos 2 x ) d x
נתחיל מהחלק הימני:
lim M → ∞ ( ∫ 1 M 1 2 x d x + ∫ 1 M cos 2 x 2 x d x ) = 1 2 lim M → ∞ ∫ 1 M 1 x d x + 1 2 lim M → ∞ ∫ 1 M cos 2 x 2 x d x \lim_{ M \to \infty } \left( \int_{1}^M \frac{1}{2x}dx+\int_{1}^M \frac{\cos2x}{2x}dx \right)=\frac{1}{2}\lim_{ M \to \infty } \int_{1}^M \frac{1}{x}dx+\frac{1}{2}\lim_{ M \to \infty } \int_{1}^M \frac{\cos2x}{2x}dx M → ∞ lim 1 ∫ M 2 x 1 d x + 1 ∫ M 2 x cos 2 x d x = 2 1 M → ∞ lim 1 ∫ M x 1 d x + 2 1 M → ∞ lim 1 ∫ M 2 x cos 2 x d x
הפונקציה 1 x \frac{1}{x} x 1 מונוטונית יורדת ורציפה ו-lim x → ∞ 1 x = 0 \lim_{ x \to \infty } \frac{1}{x}=0 lim x → ∞ x 1 = 0 והפונקציה ∫ 0 x cos 2 t d t \displaystyle \int_{0}^x \cos 2t\,dt 0 ∫ x cos 2 t d t חסומה שכן מתקיים:
∣ ∫ 0 x cos 2 t d t ∣ = ↓ N L ∣ cos 2 x 2 − 1 2 ∣ ≤ ↓ אי שוויון המשולש ∣ cos 2 x ∣ + 1 2 < 2 2 = 1 \left| \int_{0}^x \cos 2t\,dt \right| \underset{\begin{matrix}
\downarrow \\
NL
\end{matrix}}{=}\left| \frac{\cos2x}{2}-\frac{1}{2} \right|\underset{\begin{matrix}
\downarrow \\
\R{\text{אי שוויון המשולש}}
\end{matrix}}{\leq} \frac{|\cos2x|+1}{2}< \frac{2}{2}=1 0 ∫ x cos 2 t d t ↓ N L = 2 cos 2 x − 2 1 ↓ אי שוויון המשולש ≤ 2 ∣ cos 2 x ∣ + 1 < 2 2 = 1
לכן, לפי מבחן דיריכלה, הפונקציה מתכנסת.
בנוסף, נשים לב כי לפי מסקנה, lim M → ∞ ∫ 1 M 1 x d x = ∫ 1 ∞ 1 x d x = ∞ \displaystyle \lim_{ M \to \infty }\int_{1}^M \frac{1}{x}dx=\int_{1}^\infty \frac{1}{x}dx=\infty M → ∞ lim 1 ∫ M x 1 d x = 1 ∫ ∞ x 1 d x = ∞ ולכן מתקיים:
lim M → ∞ ( ∫ 1 M 1 2 x d x + ∫ 1 M cos 2 x 2 x d x ) = ∞ + L = ∞ \lim_{ M \to \infty } \left( \int_{1}^M \frac{1}{2x}dx+\int_{1}^M \frac{\cos 2x}{2x}dx \right)=\infty+L=\infty M → ∞ lim 1 ∫ M 2 x 1 d x + 1 ∫ M 2 x cos 2 x d x = ∞ + L = ∞
לכן, ∫ 0 ∞ cos 2 x x d x \displaystyle \int_{0}^\infty \frac{\cos^2x}{x}dx 0 ∫ ∞ x cos 2 x d x מתבדרת ולכן, לפי מבחן ההשוואה הגבולי, I 1 I_{1} I 1 מתכנס בתנאי!