Convergence tests for improper integrals — comparison: Visual walkthrough

4.2.12 · D2 · Maths › Calculus II — Integration › Convergence tests for improper integrals — comparison

Yeh parent topic ka visual companion hai. Steps ko order mein padho — har picture argument ki agli line carry karti hai. Hum do star players se milenge — ek curve jise hum $f$ kahenge aur ek oonchi "lid" curve jise hum $g$ kahenge — Step 1 aur Step 4 mein theek se; abhi sirf itna jaano ki $f$ hamaari curve hai aur $g$ ek helper hai jo uske upar baith'ta hai.

Step 1 — "Curve ke neeche area" ka matlab kya hota hai

KYA. Koi function lo aur usse $f$ kaho — yeh hamaari curve hai, jiska area hume ultimately jaanna hai. $x = a$ se $x = t$ tak ke window par, symbol $\int_{a}^{t} f (x) d x$ ka matlab hai curve $y = f (x)$ aur flat floor $y = 0$ ke beech ka area, left wall $x = a$ se right wall $x = t$ tak.

Term by term:

$f (x)$ — horizontal position $x$ par hamaari curve ki height.
$d x$ — ek infinitely thin slice of width, region ka ek sliver.
$\int_{a}^{t}$ — "har sliver ko jodte jao" jaise $x$ , $a$ se $t$ tak sweep karta hai.

YAHAN se kyun shuru karein. Hum kisi argument par tab hi trust kar sakte hain jab symbols ke neeche ke words honest hon. Baaki sab kuch sirf "yeh area us area se chhhota hai" hai, toh pehle yeh agree karna zaroori hai ki integral ek area hi hai.

PICTURE. Neela sliver ek $f (x) d x$ hai. Saare slivers ko dono walls ke beech stack karo aur tumhe shaded region milti hai.

Figure — Convergence tests for improper integrals — comparison

Step 2 — $\infty$ hume limit use karne par majboor kyun karta hai

KYA. Hum area hamesha ke liye right tak chahte hain: $\int_{a}^{\infty} f d x$ . Lekin $\infty$ koi number nahi hai — tum koi wall infinity par khadi nahi kar sakte. Toh hum sirf ek legal kaam karte hain: right wall ko ek finite jagah $t$ par rakho, area measure karo, phir $t$ ko dhire dhire aur right slide karo aur dekho area kya karta hai: $\int_{a}^{\infty} f (x) d x = lim_{t \to \infty} \int_{a}^{t} f (x) d x .$

$t$ — movable right wall (ek real, finite number).
$lim_{t \to \infty}$ — "wall ko right ki taraf run karne do; area kis number ke paas jaata hai?"

Agar woh running area ek finite number par settle ho jaata hai, toh hum kehte hain integral converges karta hai. Agar woh bina ruke badhta rehta hai, toh woh diverges karta hai.

LIMIT kyun aur sirf " $\infty$ plug in" kyun nahi. Tum physically ek infinite bar par ek shot mein integrate nahi kar sakte; tum sirf finite bars par integrate kar sakte ho. Limit woh bridge hai "finite window jo main compute kar sakta hoon" se "infinite window jo main chahta hoon" tak.

PICTURE. Right wall ko teen snapshots mein right jaate dekho; running area $F (t)$ badhta hai.

Step 3 — Running area ko naam do, aur notice karo ki yeh sirf UPAR jaata hai

KYA. Running area ko ek naam do: $F (t) = \int_{a}^{t} f (x) d x .$ Ab maano $f (x) \geq 0$ har jagah hai (curve kabhi floor ke neeche nahi jaati). Toh jab tum wall ko thoda sa right push karte ho, tum sirf area add kar sakte ho — kabhi subtract nahi — kyunki har naya sliver height $\geq 0$ ka hai. Toh: $t_{2} > t_{1} ⟹ F (t_{2}) \geq F (t_{1}) .$ Aisi function jo kabhi decrease nahi karti usse monotone increasing kehte hain.

$F (t)$ — wall $t$ tak collect kiya gaya total area.
$f$ par " $\geq 0$ " — yeh promise ki koi bhi sliver kabhi area nahi kaatega.

ISKA itna importance kyun hai. Ek increasing quantity ki sirf do possible destinies hain: ya toh woh hamesha ke liye badhti rehti hai (diverges), ya phir woh ek ceiling se takraati hai aur level off ho jaati hai (converges). Koi oscillating nahi, koi wandering nahi.

Yahan Monotone Convergence Theorem kaam kar raha hai, aur yahan uska statement hai taaki tumhe page chhodni na pade: koi bhi quantity jo increasing hai aur kisi fixed ceiling ke neeche rehti hai, woh zaroor ek finite limit ki taraf jaayegi.

Definition "Increasing + capped ⇒ has a limit"

sach kyun hai (completeness of $R$ ) $F (t)$ ke collect kiye gaye saare values imagine karo. Woh badhte rehte hain lekin har ek ceiling $M$ ke neeche baith'ta hai. Real numbers mein ek special property hoti hai — completeness — jo guarantee karti hai ki upper bound wale numbers ke kisi bhi set ka ek least upper bound $L$ hota hai (sabse nichli possible ceiling, koi gaps allowed nahi). Ab $F$ us $L$ ki taraf badhta hai: woh $L$ ko pass nahi kar sakta (yeh ek ceiling hai), aur woh $L$ se neeche hamesha ke liye ruk bhi nahi sakta (warna thoda sa chhhota number ek aur bhi tight ceiling hoga, jo $L$ ke least hone ka contradiction hoga). "Pass nahi kar sakta" aur "short reh nahi sakta" ke beech squeeze hoke, $F (t)$ ke paas koi choice nahi balki $L$ ki taraf home in kare — yahi uska finite limit hai. Completeness ke bina (sirf fractions imagine karo, gaps ke saath) values kisi aise hole ki taraf crowd kar sakti thin jo number hi nahi hota; yeh gap-free real line hi hai jo limit ko genuinely exist karati hai. Yahi woh machine hai jise hum Step 5 mein cash in karte hain — aur yahi wajah hai ki hum $f \geq 0$ demand karte hain, taaki $F$ genuinely increasing ho.

PICTURE. $F (t)$ ki staircase sirf upar step kar sakti hai jaise $t$ badhta hai.

Step 4 — Ek doosri curve upar rakh do aur inequality padho

KYA. Ek doosri, oonchi function $g$ — helper "lid" — laao, jiske liye $0 \leq f (x) \leq g (x) for all x \geq a .$ Iska matlab hai: har horizontal position par, $g$ -curve, $f$ -curve se kam se kam utni hi oonchi hai, aur dono floor ke upar rehti hain.

Ab lid ke running area ko apna naam do, bilkul waisa hi jaisa humne Step 3 mein $F$ ko naam diya tha: $Let G (t) = \int_{a}^{t} g (x) d x .$

$0 \leq f$ — $f$ kabhi floor ke neeche nahi jaata (Step 3 se, isliye $F$ badhta hai).
$f \leq g$ — $g$ ek lid hai jo $f$ ke upar har jagah baith'ti hai.
$G (t)$ — wall $t$ tak lid $g$ ke neeche total area (aur, kyunki $g \geq 0$ hai, yeh bhi badhta hai).

Kyunki $g$ har $x$ par oonchi hai, kisi bhi wall tak uska area utna hi bada hai: $F (t) = \int_{a}^{t} f \leq \int_{a}^{t} g = G (t) .$

KYUN. Do mushkil areas ko directly compare karna painful hai. Lekin heights compare karna trivial hai — tum sirf dekhte ho kaunsi curve upar hai. Height-by-height " $f \leq g$ " turant area-by-area " $F \leq G$ " de deta hai. Humne ek mushkil sawaal (ek integral) ko ek aasaan sawaal (formulas ki inequality) se replace kar diya.

PICTURE. $g$ plum lid hai; $f$ orange region hai jo uske neeche rehti hai.

Step 5 — Finite lid hamare area ke liye ceiling ban jaati hai

KYA. Ab maano hum pehle se jaante hain ki lid ka total area finite hai: $\int_{a}^{\infty} g d x = M (koi finite number) .$ Kyunki $G (t)$ bhi increasing hai aur $M$ ki taraf jaata hai, toh har $t$ ke liye $G (t) \leq M$ hai. Isse Step 4 ke saath chain karo: $F (t) \leq G (t) \leq M .$ Toh $F (t)$ ek increasing quantity hai (Step 3) jo kabhi $M$ se zyada nahi hoti — yeh upar se ek ceiling se bounded hai.

$M$ — lid ke neeche finite total area.
$F (t) \leq M$ — hamaara area kabhi ceiling $M$ se nahi nikalta.

KYUN yeh argument close karta hai. Monotone Convergence Theorem ke mutabiq (jise humne Step 3 mein completeness se power diya tha), ek increasing quantity jo ek ceiling ke neeche trapped hai infinity ki taraf nahi ja sakti — uska ek least upper bound hota hai aur woh us par home in karti hai. Isliye $lim_{t \to \infty} F (t)$ exist karta hai aur finite hai: $0 \leq f \leq g and \int_{a}^{\infty} g converges ⟹ \int_{a}^{\infty} f converges .$ Yahi Direct Comparison Test hai, area se, ek limit se, "increasing" se, aur "ek ceiling ke neeche trapped" se build kiya gaya — aur kuch nahi.

PICTURE. Orange staircase $F (t)$ badhti hai lekin dashed ceiling $M$ ke neeche pin hai; usse level off karna hi padega.

Step 6 — Mirror case: divergence upar ki taraf push karta hai

KYA. Roles flip karo. Maano bottom curve ka pehle se infinite area hai: $\int_{a}^{\infty} f d x = \infty, 0 \leq f \leq g .$ Kyunki $F (t) \leq G (t)$ (Step 4) aur $F (t) \to \infty$ , badi quantity $G (t)$ uske saath upar khinchi jaati hai: $G (t) \geq F (t) \to \infty ⟹ \int_{a}^{\infty} g diverges .$

KYUN dono directions chahiye. Convergence proofs upar se squeeze karte hain (ek finite lid dhundho). Divergence proofs neeche se push karte hain (neeche ek infinite floor dhundho). Same picture, opposite directions mein padha gaya — jo bilkul wahi mnemonic hai "squeeze to please, push to die."

PICTURE. Neeche ek divergent $f$ oonchi $g$ ko bhi blow up hone par majboor karti hai.

Step 7 — Degenerate trap: "infinite se chhhota" kyun KUCH BHI prove nahi karta

KYA. Tempting wala galat move: " $f \leq g$ aur $\int g = \infty$ , toh $\int f = \infty$ bhi." Dekho yeh kaise fail hota hai. Lo $f = 1/ x^{2}$ aur $g = 1/ x$ , $[1, \infty)$ par. Sach mein $f \leq g$ hai. Lekin $\int_{1}^{\infty} \frac{d x}{x ^{2}} = 1 (finite), \int_{1}^{\infty} \frac{d x}{x} = \infty.$ Ek finite area ek infinite area ke neeche baith'a hai. Ek bottomless well ke neeche hona tumhe tumhari apni depth ke baare mein kuch nahi bataata.

$f \leq g$ yahan SACH hai — lekin useless hai, kyunki yeh galat direction mein point karta hai.

KYUN. Sirf do arrows hain jo information carry karte hain: $(small \leq convergent) \Rightarrow convergent, (big \geq divergent) \Rightarrow divergent .$ Baaki do combinations ("small $\leq$ divergent", "big $\geq$ convergent") koi conclusion nahi dete. Hamesha check karo ki arrow known fate se unknown fate ki taraf sahi direction mein point kar raha hai.

PICTURE. Finite orange area $1/ x^{2}$ , infinite teal area $1/ x$ ke neeche nest karti hai — proof ki naive implication galat hai.

One-picture summary

KYA. Ek diagram mein saate steps ek saath hain: hamaari nonnegative curve $f$ (orange, area shaded), uske upar ek lid $g$ (plum), aur uske neeche ek floor curve (teal, dashed). Har pictorial element ek step hai jo visible bana diya gaya hai.

Orange region — $F (t) = \int_{a}^{t} f$ , Steps 1–3 se increasing running area.
Plum lid $g$ — Steps 4–5 se finite ceiling: $f$ ko uske neeche squeeze karo ⇒ converges.
Teal dashed floor — Step 6 se infinite floor: $f$ ko uske upar push karo ⇒ diverges.

KYUN yeh teen elements aur koi nahi. Har legal comparison bilkul do moves mein se ek hai, aur picture dono simultaneously dikhati hai: convergence jeetne ke liye ek finite lid se upar bound karo, ya divergence force karne ke liye ek infinite floor se neeche bound karo. Step 7 ka trap bilkul yahi hai jo hota hai agar tum arrows mix up karo — ek finite curve khushi se ek infinite ke neeche rehti hai, kuch prove kiye bina.

Recall Feynman retelling — poora walkthrough plain words mein

Ek integral sirf woh area hai jo ek curve ke neeche trapped hai (Step 1). Hum infinity par wall nahi bana sakte, toh hum ek finite jagah $t$ par wall banate hain aur usse right slide karte hain, area dekhte hain — woh sliding hi limit hai (Step 2); bilkul yahi sliding, infinity ki jagah ek blow-up point ki taraf aimed, vertical-asymptote integrals bhi handle karti hai. Agar curve kabhi floor ke neeche nahi jaati, toh area sirf badhta hai jaise wall move karti hai; woh kabhi shrink nahi karta (Step 3). Ek rising quantity jo ek ceiling ke neeche rehti hai uska ek least ceiling hota hai jis par woh home in karti hai — yahi real line ki completeness hai, aur yahi woh hai jo limit ko truly exist karaati hai. Ab ek oonchi curve $g$ , hamaari curve $f$ ke upar lid ki tarah rakh do: kyunki $g$ har jagah oonchi hai, uska area har jagah bada hai (Step 4). Agar us lid ka finite total area $M$ hai, toh hamaara badhta hua area hamesha ke liye $M$ ke neeche stuck hai — toh usse ek finite value par settle karna hi padega. Yahi convergence hai (Step 5). Picture ko ulta padho aur tumhe divergence milti hai: agar koi curve hamare neeche pehle se infinite area rakhti hai, toh hum bhi infinity tak khinche jaate hain (Step 6). Ek trap: kisi infinite se chhhota hona kuch prove nahi karta — ek saaf finite area jaise $1/ x^{2}$ khushi se bottomless $1/ x$ ke neeche rehta hai (Step 7). Sirf do directions khabar laati hain: finite hone ke liye ek finite lid ke neeche squeeze karo, infinite hone ke liye ek infinite floor ke upar push karo. Aur yeh sab tabhi kaam karta hai jab dono curves floor ke upar rehti hain.

Connections

Parent topic
The p-integral and p-series — lid $g$ almost hamesha ek power $x^{- p}$ hoti hai.
Monotone Convergence Theorem — Step 3 aur Step 5 ke peeche ka engine.
Improper integrals — infinite discontinuities (type 2) — yahi argument vertical asymptotes bhi cover karta hai.
Comparison test for infinite series — sums ke liye bilkul same picture.
Absolute vs conditional convergence — jab $f$ sign change kare tab kya karo.
Integral Test for series