Uniform convergence of function sequences: Foundations

4.10.24 · D1 · Maths › Advanced Topics (Elite Level) › Uniform convergence of function sequences

Yeh page kuch bhi assume nahi karta. Parent note Uniform convergence padhne se pehle, tumhe har wo symbol khud ka banana hoga jo wahan aata hai. Hum har ek ko ek picture se build karte hain.

1. Numbers jo hum allow karte hain: $R$ aur interval $[a, b]$

Kuch aur naam lene se pehle, un numbers ka playground fix karo jisme hum kaam karte hain.

$R$ ko ek aisa poora horizontal ruler samjho jo dono taraf hamesha ke liye extend karta hai, aur $[a, b]$ uska ek highlighted hissa hai jiske dono siro par solid dots hain.

YEH TOPIC ISKO KYU CHAHTA HAI: jo bhi input hum kabhi bhi kisi function ko denge, woh ek real number hai, aur jis domain ko hum dekhte hain woh almost hamesha ya poora $R$ hota hai ya ek segment $[a, b]$ . Hume pehle yeh do symbols apna banana hoga, kyunki yeh jab hi hum kisi function ke rehne ki jagah describe karte hain toh sab se pehle aate hain.

2. Function $f$ kya hota hai?

Ise ek grid par ek curve ke roop mein imagine karo. Horizontal axis mein input $x$ hai; vertical axis mein output $f (x)$ hai. Poori curve hi function hai — har horizontal position ke liye ek height.

Figure — Uniform convergence of function sequences

YEH TOPIC ISKO KYU CHAHTA HAI: hamare paas ek saath kaafi saari curves hongi, aur hum unki heights compare karenge. Isliye pehle bilkul comfortable ho jao ki "ek function" = "ek curve = ek rule jo har $x$ ke liye ek height assign karta hai".

3. Common domain $E$

$E$ ko horizontal axis ka woh hissa samjho jise hum dekhne ki permission rakhte hain. Jab baad mein hum "for all $x \in E$ " likhte hain, toh symbol $\in$ ka matlab hai "belongs to", isliye " $x \in E$ " padha jaata hai " $x$ allowed inputs mein se ek hai".

YEH TOPIC ISKO KYU CHAHTA HAI: uniform convergence sirf kisi formula ki property nahi hai — yeh depend karta hai kin inputs ko tum allow karte ho par. Parent note dikhata hai $x^{n}$ uniform hai $[0, a]$ par lekin $[0, 1]$ par nahi: same formula, alag $E$ , alag jawab. Isliye $E$ ek first-class symbol hona chahiye, jise kuch bhi measure karne se pehle naam diya jaaye.

4. $f_{n}$ mein subscript ka kya matlab hai?

Ek flip-book imagine karo: har page $n$ par thodi alag curve hai. Jab tum tezi se flip karte ho ( $n$ badhta hai), curves ek final shape mein settle ho jaati hain. Woh final shape limit function $f$ hai (koi subscript nahi — woh destination hai, list ka member nahi).

YEH TOPIC ISKO KYU CHAHTA HAI: uniform vs pointwise ek statement hai is flip-book ke settle hone ke baare mein, isliye counting index $n$ pehle chahiye baki sab se.

5. Absolute value $∣ \cdot ∣$ — gap measure karna

Ab crucial use: $∣ f_{n} (x) - f (x) ∣$ vertical distance hai curve number $n$ aur target curve ke beech, ek single input $x \in E$ par measured. Yeh hamesha $\geq 0$ hota hai (distance negative nahi ho sakti), aur exactly $0$ hota hai jab do curves us $x$ par touch karti hain.

YEH TOPIC ISKO KYU CHAHTA HAI: convergence yahi gap shrink hona hai. Neeche har definition $∣ f_{n} (x) - f (x) ∣$ ke chhota hone ke baare mein ek sentence hai.

6. Error budget $ε$ (epsilon)

Target curve ke around $ε$ half-height ka ek patla horizontal band socho — $f (x) - ε$ se $f (x) + ε$ tak ek fuzzy ribbon. " $∣ f_{n} (x) - f (x) ∣ < ε$ " kehna matlab hai: us $x$ par, curve $f_{n}$ ribbon ke andar hai.

YEH TOPIC ISKO KYU CHAHTA HAI: convergence ek challenge game hai. Challenger ek budget $ε$ name karta hai; humein eventually uske andar fit hona hoga. Chhota $ε$ = patli ribbon = zyada mushkil challenge.

7. Threshold $N$ aur arrow $⟹$

Symbol $⟹$ ka matlab hai "lead karta hai / guarantee karta hai". Isliye $n \geq N ⟹ ∣ f_{n} (x) - f (x) ∣ < ε$ plain words mein padha jaata hai: "jaise hi hum flip-book ke page $N$ ya baad mein hain, curve $f_{n}$ us $x$ par ribbon ke andar baithti hai."

YEH TOPIC ISKO KYU CHAHTA HAI: pointwise aur uniform convergence ke beech ek hi farq hai ki $N$ ko $x$ par depend karne ki permission hai ya nahi. Jab tak $N$ crystal clear nahi hai, tum yeh farq nahi dekh sakte.

8. Quantifiers $\forall$ aur $\exists$ (apne domains ke saath)

$\forall x \in E \exists N \in N$ : "har allowed input $x$ ke liye, tum phir koi page number $N$ dhundh sakte ho" — $N$ $x$ ko dekhne ke baad choose hota hai, isliye woh $x$ par depend kar sakta hai. (Har runner ko apni deadline milti hai.)
$\exists N \in N \forall x \in E$ : "ek aisa page number $N$ hai jo phir har allowed input $x$ ke liye kaam karta hai" — $N$ $x$ se pehle choose hota hai, isliye woh depend nahi kar sakta $x$ par. (Poori bheed ke liye ek deadline.)

YEH TOPIC ISKO KYU CHAHTA HAI: parent note ka EK punchline hai "ek hi farq quantifier order ka hai". Yeh page par sabse deep symbol hai — ise tab tak dubara padho jab tak click na ho jaaye.

9. Numerical limit ka fully quantified $ε$ – $N$ definition

Functions ki limit handle karne se pehle, ordinary numbers ki limit pakki karo — kyunki uniform convergence exactly isi template ko re-wrap karke banaya jaata hai.

YEH TOPIC ISKO KYU CHAHTA HAI: parent note ka $M_{n} \to 0$ ek numerical limit hai; aur continuity, integrals aur Cauchy tests sabhi is shape ke statements mein reduce hote hain. Yahan quantifier order apna karna matlab hai ki parent page par har $\to$ precisely padh sako.

10. Supremum $sup$ — tightest ceiling (aur jab woh finite hai)

Pehle, ek aur symbol: is section mein $g$ ek generic function on $E$ ke liye hai, yaani ek rule $g$ jo koi bhi input $x \in E$ leta hai aur ek real number $g (x) \in R$ return karta hai (exactly §2 ki "machine", bas ek temporary naam diya gaya taaki hum uske output values ke collection ke baare mein baat kar sakein). Thodi der mein hum $g (x) = ∣ f_{n} (x) - f (x) ∣$ , gap, lete hain.

"Maximum" kyun nahi kehte? Kyunki kabhi kabhi koi maximum hota hi nahi. Example: $x \in [0, 1)$ ke liye $x^{n}$ values $1$ ki taraf climb karti hain lekin $1$ ke equal kabhi nahi hoti (kyunki $x < 1$ ). Koi largest value nahi hai — phir bhi $1$ tightest ceiling hai. Isliye $sup = 1$ chahe max exist nahi karta. Yahi subtlety $x^{n}$ counterexample ko kaam karaati hai.

Ab poore topic ki star quantity — $g (x) = ∣ f_{n} (x) - f (x) ∣$ lo: $M_{n} := sup_{x \in E} ∣ f_{n} (x) - f (x) ∣.$ Words mein: $E$ par kahin bhi curve $n$ aur target ke beech ka sabse bada gap. $E$ ke poore domain par slide karo aur sabse bura (sabse lamba) vertical gap record karo — woh sabse bura gap $M_{n}$ hai.

YEH TOPIC ISKO KYU CHAHTA HAI: $sup$ woh tool hai jo infinitely many points ke baare mein ek statement ko ek checkable number mein convert karta hai — lekin sirf tab jab woh number finite ho, isliye boundedness equipment ka hissa hai.

11. Limit arrow $\to$ $M_{n}$ par apply karna

§9 mein humne $a_{n} \to L$ ka fully quantified meaning pehle hi diya tha. Ise specific sequence $a_{n} = M_{n}$ aur target $L = 0$ par apply karo: $M_{n} \to 0 matlab \forall ε > 0 \exists N \in N \forall n \geq N : M_{n} < ε .$ Isliye $M_{n} \to 0$ ka matlab hai $E$ par worst gap kuch nahi mein shrink ho jaata hai. Woh single line hi uniform convergence hai (parent §2). Alag se, $x^{n} \to 0$ ek fixed $x \in [0, 1)$ ke liye matlab hai: $1$ se neeche wale number ko baar baar khud se multiply karna use $0$ ki taraf crush kar deta hai.

YEH TOPIC ISKO KYU CHAHTA HAI: har convergence claim " $\to$ kuch" mein khatam hoti hai. Yeh poore subject ki verb hai — aur ab tum ise quantifiers ki ek precise chain ki tarah padh sakte ho.

12. Continuous aur integrable

YEH TOPIC IN SHABDON KO KYU CHAHTA HAI: payoff theorems (parent §4) kehti hain ki uniform convergence continuity ko preserve karti hai aur tumhe limit ko integral ke saath swap karne deti hai. In shabdon ko pahchanna zaroori hai taaki dekh sako ki uniform convergence "takleef ke kabil" kyun hai.

Prerequisite map

Upar se neeche padho: real numbers aur interval playground fix karte hain; ek common domain $E$ par functions ki pictures gap, epsilon-ribbon aur $N$ -deadline ko feed karti hain, jo quantifier duel aur limits ke liye $ε$ – $N$ template ko feed karte hain; finite supremum "for all $x$ " ko ek single number mein turn karta hai jiska limit uniform convergence define karta hai, jo payoff theorems unlock karta hai.

Equipment checklist

Recall Self-test — kya tum parent note padhne se pehle har ek ka jawaab de sakte ho?

$R$ aur $[a, b]$ ka kya matlab hai? ::: $R$ = number line par har point; $[a, b]$ = $a$ se $b$ tak ke saare reals, dono endpoints sameta. Letter $E$ kya represent karta hai, aur ise ignore kyun nahi kiya ja sakta? ::: Sabhi $f_{n}$ aur $f$ dwara shared inputs ka common domain; uniformity depend karta hai kin inputs ko tum allow karte ho ( $x^{n}$ uniform hai $[0, a]$ par lekin $[0, 1]$ par nahi). $f_{n}$ mein subscript $n$ ka kya matlab hai? ::: List mein kaunsi curve hai yeh count karne wala name-tag; yeh $n$ se multiplication NAHI hai. $∣ f_{n} (x) - f (x) ∣$ geometrically kya measure karta hai? ::: Curve $n$ aur target curve ke beech vertical distance, single input $x \in E$ par. $ε$ kya role play karta hai? ::: Ek chosen positive error budget; yeh target ke around $ε$ half-height ki ek ribbon draw karta hai jiske andar $f_{n}$ ko eventually baithna hoga. $n \geq N ⟹ \dots$ kya promise karta hai? ::: Page $N$ se aage, stated smallness guaranteed hai; $N$ deadline hai. $a_{n} \to L$ ka fully quantified meaning likho. ::: $\forall ε > 0 \exists N \in N \forall n \geq N : ∣ a_{n} - L ∣ < ε$ . Bare $\forall x, \exists N$ ki jagah $\forall x \in E$ aur $\exists N \in N$ kyun likhna chahiye? ::: Yeh kehne ke liye ki har variable kahan rehta hai; aur kyunki $\forall x \exists N$ (N, x par depend kar sakta hai) $\exists N \forall x$ (sabke liye ek N) se alag hai — pointwise vs uniform. $sup_{x \in E} g (x)$ mein $g$ kya hai aur "maximum" ki jagah $sup$ kyun? ::: $g$ ek generic function on $E$ hai (har $x \in E$ ko ek real $g (x)$ par map karta hai); $sup$ isliye use hota hai kyunki sabse badi value kabhi reach nahi ho sakti (jaise $[0, 1)$ par $x^{n}$ $1$ ke paas aata hai lekin kabhi nahi pahunchta), phir bhi $sup$ tightest ceiling naam karta hai — provided gap $E$ par bounded ho, warna $M_{n} = + \infty$ . $M_{n}$ ek phrase mein kya hai, aur $M_{n} \to 0$ ka kya matlab hai? ::: $M_{n}$ = $f_{n}$ aur $f$ ke beech $E$ par sabse bura (finite) gap; $M_{n} \to 0$ ka matlab hai woh worst gap kuch nahi mein shrink ho jaata hai — yahi IS uniform convergence. Uniform convergence tumhe safely kya karne deti hai jo pointwise nahi kar sakti? ::: Limit mein continuity rakho aur $lim$ ko $\int$ ke saath swap karo.

Connections

Parent: Uniform convergence — yahan jao jab upar ke har symbol obvious lage.
Pointwise convergence — woh kamzor cousin jise yeh foundations bhi define karti hain.
Weierstrass M-test aur Cauchy sequences in metric spaces — jahan $sup$ aur $N$ dubara aate hain.
Continuity preserved under uniform limits aur Interchange of limit and integral — §12 ka payoff vocabulary.

#flashcards/maths