Visual walkthrough — Laurent series — principal part, annulus of convergence

4.10.4 · D2 · Maths › Advanced Topics (Elite Level) › Laurent series — principal part, annulus of convergence

Shuru karne se pehle, ek promise: har symbol ko pehli baar plain words mein meaning aur ek picture milegi. Kuch bhi assume nahi kiya gaya.

Step 1 — $z$ kya hai, aur "$\oint$ around $a$" ka matlab kya hai?

KYA. Ek complex number $z$ bas ek flat map (ek plane) par ek point hai. Uski horizontal position "real part" hai, uski vertical position "imaginary part" hai. Hum ek special centre point fix karte hain aur use $a$ kehte hain. $a$ se $z$ tak ka chhota arrow $z-a$ likha jaata hai.

KYUN. Is poori theory mein sab kuch iske baare mein hai ki ek function point $a$ ke around circle karne par kaise behave karta hai. Toh sabse pehli cheez jo humein chahiye woh hai $a$ ke around ek loop mein walk karne ka tarika. Symbol ${\oint }_{\gamma }$ ka matlab hai "add up karo (integrate karo) jab tum ek baar, counter-clockwise, ek closed path $\gamma$ (ek loop) ke puri tarah around travel karte ho."

PICTURE. Figure mein loop $\gamma$ dekho. Centre dot $a$ hai. Moving dot $z$ hai. Green arrow $z-a$ spin karta hai jab $z$ loop walk karta hai.

Step 2 — Woh ek integral jo sab kuch power karta hai

KYA. Hum har whole-number power $k$ ke liye ek single, deceptively simple loop integral compute karte hain: ${\oint }_{\gamma }(z-a{)}^{k}dz.$ Yahan $k$ koi bhi integer hai ($\dots ,-2,-1,0,1,2,\dots$), aur $(z-a{)}^k$ green arrow ko $k$-th power tak raise kiya gaya hai.

KYUN. Yahi woh sieve hai jo hum baad mein use karenge. Agar hum dikhaa saken ki yeh integral almost har $k$ ke liye zero hai aur exactly ek special $k$ ke liye nonzero hai, toh poori series ko integrate karne se har term kill ho jaayega except ek — aur exactly aise hi hum ek single coefficient ko isolate karenge.

PICTURE. Dekho ki $(z-a{)}^k$ kya trace karta hai jab $z$ ek baar around jaata hai. Kyunki $z-a=\rho e^{i\theta }$ hai, hum paate hain $(z-a{)}^k={\rho }^k{e}^{ik\theta }$: fixed length ${\rho }^k$ ka ek arrow jo $(k)$ times spin karta hai jab $\theta$ ek baar around jaata hai. Figure kuch values of $k$ ke liye traced circle dikhata hai.

$z=a+\rho e^{i\theta }$ substitute karo, toh $dz=i\rho e^{i\theta }d\theta$ (loop ke along ek tiny step): ${\oint }_{\gamma }(z-a{)}^{k}dz={\int }_0^{2\pi }\underset{(z-a{)}^k}{\underbrace{{\rho }^k{e}^{ik\theta }}}\cdot \underset{dz}{\underbrace{i\rho e^{i\theta }}}d\theta =i{\rho }^{k+1}{\int }_0^{2\pi }{e}^{i(k+1)\theta }d\theta .$

• ${\rho }^k{e}^{ik\theta }$ — power khud: length ${\rho }^k$, angle $k\theta$.
• $i\rho e^{i\theta }d\theta$ — loop ke along tiny push $dz$ ki direction aur length.
• ${\rho }^{k+1}$ — dono radii ${\rho }^k$ aur $\rho$ combine hoke; ek harmless constant.
• $e^{i(k+1)\theta }$ — woh akela part jo abhi bhi spin karta hai; uski spinning hi sab kuch decide karti hai.

Step 3 — Kyun sieve almost hamesha zero hai

KYA. ${\int }_0^{2\pi }{e}^{i(k+1)\theta }d\theta$ evaluate karo.

KYUN. Yeh ek number poora outcome decide karta hai, isliye hum ise carefully dekhte hain aur do cases mein split karte hain.

PICTURE. $e^{i(k+1)\theta }$ ek unit arrow hai jo $(k+1)$ complete turns karta hai jab $\theta$ $0\to 2\pi$ jaata hai. Ek arrow ko add up karna jo full circles spin karta hai matlab har direction mein equally pointing arrows ko add karna — woh perfectly cancel ho jaate hain, zero dete hain. Lekin agar arrow kabhi nahi turn karta (poore time same way point karta hai), toh saare pieces stack up ho jaate hain aur ek bada nonzero total dete hain. Figure in dono ko contrast karta hai: cancelling petals vs. ek solid stack.

Step 4 — Maano series exist karti hai, phir ek "shifter" se multiply karo

KYA. Maano (parent note ka Annulus of convergence guarantee karta hai yeh) ki ring $r<|z-a|<R$ par humara function ek two-sided series ke barabar hai $f(z)={\sum }_{m=-\infty }^{\infty }{c}_m(z-a{)}^m.$ Hum bas ek number chahte hain, maano $c_n$. Dono sides ko $(z-a{)}^{-n-1}$ se multiply karo: $\frac{f(z)}{(z-a{)}^{n+1}}={\sum }_{m=-\infty }^{\infty }{c}_m(z-a{)}^{m-n-1}.$

KYUN. Hum deliberately har power ko shift kar rahe hain. Woh term jiske baare mein hum care karte hain (woh $m=n$ term) ban jaata hai $(z-a{)}^{n-n-1}=(z-a{)}^{-1}$ — exactly woh special power jo sieve rakhta hai! Har doosra term koi aur power ban jaata hai, jise sieve destroy kar dega. Yahi plan precisely bana hai.

PICTURE. Numbered dials ki ek row imagine karo, har term $m$ ke liye ek. $(z-a{)}^{-n-1}$ se multiply karna poori row ko slide karta hai taaki dial $m=n$ magic "$-1$" slot par land kare. Figure woh slide dikhata hai.

• $c_m$ — $m$-th power ka unknown coefficient (jo hum dhundh rahe hain).
• $(z-a{)}^{m-n-1}$ — shifted power; uska exponent $m-n-1$ hai.
• Special case $m=n$ exponent $-1$ deta hai.

Step 5 — Integrate karo: sieve exactly ek term rakhta hai

KYA. Dono sides ko annulus ke andar loop $\gamma$ ke around integrate karo (series wahan converge karti hai isliye term by term integral jaata hai): ${\oint }_{\gamma }\frac{f(z)}{(z-a{)}^{n+1}}dz={\sum }_{m=-\infty }^{\infty }{c}_m{\oint }_{\gamma }(z-a{)}^{m-n-1}dz.$

KYUN. Ab Step 3 sieve ko har inner integral par apply karo. Har ek zero hai jab tak exponent $m-n-1=-1$ na ho, yaani jab tak $m=n$ na ho. Toh poora infinite sum collapse hokar ek single surviving term ban jaata hai.

PICTURE. Saare dials dark ho jaate hain except woh jo "$-1$" slot par baitha hai, jo $2\pi i$ value ke saath light up karta hai. Baaki sab kuch contribute nahi karta. Figure woh survivor dikhata hai.

${\oint }_{\gamma }\frac{f(z)}{(z-a{)}^{n+1}}dz=c_n\cdot \underset{=2\pi i}{\underbrace{{\oint }_{\gamma }(z-a{)}^{-1}dz}}=c_n\cdot 2\pi i.$

$2\pi i$ se divide karo:

Step 6 — Free bonus: uniqueness, aur residue

KYA. Formula har $c_n$ ko directly $f$ se deta hai — koi choices nahi ki gayi. Toh coefficients forced hain.

KYUN matter karta hai. Yeh parent note ka yeh claim prove karta hai ki Laurent series har annulus ke liye unique hai: agar do series dono ring par $f$ ke barabar hain, toh formula unke coefficients ko same numbers par pin karta hai. (Different annuli $\Rightarrow$ different loop $\gamma$ available $\Rightarrow$ possibly different coefficients — koi contradiction nahi.)

PICTURE. $n=-1$ set karo. Tab $(z-a{)}^{n+1}=(z-a{)}^0=1$ hai, aur formula $f$ ka plain loop integral ban jaata hai: $c_{-1}=\frac{1}{2\pi i}{\oint }_{\gamma }f(z)dz={\mathrm{Res}}_{z=a}f.$ Woh single coefficient residue hai — akela cheez jo tab survive karta hai jab tum $f$ ko ek singularity ke around integrate karte ho — yahi exactly wajah hai ki Residue theorem aur Cauchy integral formula dono isi ek calculation se nikl aate hain. Figure $n=-1$ row ko spotlight karta hai.

• $n=-1$ — woh choice jo denominator remove kar deta hai.
• ${\oint }_{\gamma }fdz$ — literally function ka loop integral.
• Result $=2\pi i\cdot c_{-1}$ — ek loop ek number measure karta hai: residue.

Step 7 — Degenerate & edge cases (reader ko kabhi stranded mat chhhodo)

KYA & KYUN. Hum formula ko uski boundaries par test karte hain taaki koi scenario tumhe surprise na kare.

Case A — $f$ analytic hai (koi singularity nahi). Tab $f$ andar smooth hai, annulus ek full disk mein swell ho sakta hai ($r\to 0$), aur har negative $n$ ke liye loop andar kuch bhi singular na hone ki wajah se ek point tak shrink ho sakta hai, isliye ${\oint }_{\gamma }\frac{f}{(z-a{)}^{n+1}}dz=0$. Saare negative coefficients vanish ho jaate hain: Laurent series ek Taylor series mein degenerate ho jaati hai. Principal part $=0$ $\Rightarrow$ removable.

Case B — order $m$ ka pole. $f$ ko $(z-a{)}^m$ se multiply karo ek analytic function paane ke liye; uska lowest surviving negative coefficient $c_{-m}\ne 0$ hai aur uske neeche kuch nahi hai. Principal part "scream" powers $\frac{1}{z-a},\dots ,\frac{1}{(z-a{)}^m}$ ka ek finite stack hai.

Case C — essential singularity. Negative half kabhi nahi rukta; infinitely many $c_{-n}\ne 0$ hain. Formula har $n$ ke liye abhi bhi hold karta hai — bas kabhi last term produce nahi karta.

Case D — loop $\gamma$ ka choice. Kaun sa loop? Annulus ke andar koi bhi loop jo $a$ ke around ek baar counter-clockwise circle karta hai woh same answer deta hai, kyunki do aise loops ke beech $f$ analytic hai aur integral nahi badalta (deformation). Agar loop ne $a$ ke around do baar circle kiya hota, tum $2\times$ paate; agar clockwise hota, toh minus sign. Figure teeno dikhata hai.

Ek-picture summary

Ek single sweep mein poora derivation: $f$ ko shifter $(z-a{)}^{-n-1}$ se multiply karo taaki woh term jo tum chahte ho magic "$-1$" slot par land kare → loop ke around integrate karo → sieve har doosre term ko zero karta hai → ek survivor emerge karta hai, $2\pi i$ se scaled → $c_n$ read karne ke liye use divide karo.

Active recall