Residues and poles: Visual walkthrough — Hinglish deep dive

4.10.5 · D2 · Maths › Advanced Topics (Elite Level) › Residues and poles

Yeh ek visual, zero-se-shuru walkthrough hai parent result Residues and poles ki. Hum prove karne waale hain, har step par ek tasveer ke saath, ki yeh complicated-dikhne wala integral $\oint_{C} f (z) d z = 2 π i Res (f, z_{0})$ ek akele number mein kyun simat jaata hai. Pehle se koi fluency zaroori nahin — har symbol ko hum usi waqt build karte hain jab woh aata hai.

Hum dekh kya rahe hain (setup seedhi baaton mein)

Kisi bhi algebra se pehle, aaiye picture par agree kar lete hain.

$z$ ek complex number hai. Ise ek flat sheet par ek point ki tarah socho ("complex plane"): horizontal doori uska real part hai, vertical doori uska imaginary part. Toh $z$ bas graph paper par ek arrow-tip ki location hai.
$f (z)$ ek machine hai: isko point $z$ do, yeh ek aur complex number ugal deta hai. Kuch bhi mysterious nahin.
Ek contour $C$ us sheet par khiincha gaya ek closed loop hai — tum kahin se shuru karte ho, ghumte ho, aur waapis aa jaate ho. Hum ek circle use karenge kyunki circles angle se describe karne ke liye sabse aasaan loops hain.
$\oint_{C} f (z) d z$ ka matlab hai: loop ke around ek baar chalo, $f (z)$ ko har tiny step $d z$ ke saath add karte jao jo tum lete ho. Integral sign par chhota circle $\oint$ bas yeh kehta hai "path ek closed loop hai".
Ek singularity $z_{0}$ woh point hai jahan $f$ bigad jaata hai (infinity tak blow up ho jaata hai). Hamaari poori kahaani iss baare mein hai ki ek loop ko kya feel hota hai jab woh aisa point circle karta hai.

Step 1 — Singularity par zoom in karo aur $f$ ko simple pieces ke sum ke roop mein likho

KYA. Trouble-spot $z_{0}$ ke bahut kareeb, hum complicated $f$ ko pure powers of $(z - z_{0})$ ki ek infinite sum se replace karte hain. Yeh hai Laurent series:

$f (z) = saari powers add karo n = - \infty \sum \infty ek fixed number a_{n} hum z_{0} se kitne door hain, power n tak uthaya (z - z_{0})^{n}$

Symbols ko unki jagah par padhna:

$(z - z_{0})$ woh displacement arrow hai bad point se waahan tak jahan hum abhi khade hain.
$n$ saare whole numbers par run karta hai — positive aur negative dono. Negative $n$ ka matlab hai $(z - z_{0})^{- 1} = \frac{1}{z - z _{0}}$ , jo $z_{0}$ ke paas jaane par bahut bada ho jaata hai. Woh negative-power terms hi "blow-up" pieces hain.
$a_{n}$ bas woh fixed weight hai jo har power se attached hai. Show ka star $a_{- 1}$ hoga.

KYun. Powers of $(z - z_{0})$ hi woh ek cheez hain jise hum circle ke around cleanly integrate karna jaante hain. Toh hum mushkil $f$ ko easy pieces ke dher ke badle mein trade karte hain, aur har piece ko alag-alag integrate karte hain.

TASVEER. Har power $(z - z_{0})^{n}$ sheet par ek chhota "flow field" hai. Step 1 ki figure mein teen sample terms ek function mein stack ki hui hain.

Step 2 — Loop ko angle mein badlo: circle ko parametrize karo

KYA. Hum $C$ ko $z_{0}$ par centred radius $ρ$ ka ek circle choose karte hain. Us circle par har point hai

$z = z_{0} + ρ e^{i θ}, θ \in [0, 2 π]$

Term by term:

$ρ$ (Greek "rho") fixed radius hai — loop centre se kitni door baithi hai.
$e^{i θ}$ ek unit arrow hai jo angle $θ$ par point kar raha hai. Jaise $θ$ $0 \to 2 π$ sweep karta hai, yeh arrow ek baar poora ghoom jaata hai. (Yahi $e^{i θ}$ ka poora matlab hai: ek direction dial.)
Toh $z_{0} + ρ e^{i θ}$ = "centre se shuru karo, direction $θ$ mein $ρ$ ke barabar step out karo" — bilkul circle ka rim.

Tiny step $d z$ differentiate karne par ban jaata hai, $d z = i ρ e^{i θ} d θ$ jahan $d θ$ angle ka woh sliver hai jitna hum aage badhte hain, aur factor $i ρ e^{i θ}$ kehta hai ki step circle ke tangent point karta hai ( $i$ ek $9 0^{\circ}$ turn hai — radius ke perpendicular, yaani rim ke along).

KYun. Ek bhatakne waale path par integral awkward hota hai. Jab sab kuch single dial-variable $θ$ mein likha jaata hai, toh loop integral $0$ se $2 π$ tak ek ordinary integral ban jaata hai — school-level calculus.

TASVEER. Step 2 ki figure radius arrow $ρ e^{i θ}$ aur us se right angle par tangent step $d z$ dikhati hai.

Step 3 — EK power term integrate karo aur exponentials ko spin karte dekho

KYA. Ek single term $(z - z_{0})^{n}$ lo aur parametrization push through karo. Kyunki $(z - z_{0}) = ρ e^{i θ}$ :

= i\,\rho^{\,n+1}\int_0^{2\pi} e^{\,i(n+1)\theta}\,d\theta$$ Har piece ne kya kiya: - $\rho^n e^{in\theta}$ rim par woh term khud hai. - $i\rho e^{i\theta}\,d\theta$ step $dz$ hai. - Multiply karne par, do exponentials merge ho jaate hain: $e^{in\theta}\cdot e^{i\theta}=e^{i(n+1)\theta}$. **$\rho$ ki powers combine hokar $\rho^{n+1}$ ban jaati hain**, aur jo bhi constant hai woh integral ke bahar slip kar jaata hai. **KYun.** Hum sabse clean possible object ke saath reh jaate hain: $\int_0^{2\pi} e^{i k\theta}\,d\theta$ jahan $k=n+1$. Yahi ek integral poora theorem decide karta hai. **TASVEER.** Step 3 ki figure merged spinning arrow $e^{i(n+1)\theta}$ dikhati hai jo $\theta$ ke aage badhne par ek circle trace karta hai. ![[deepdives/dd-maths-4.10.05-d2-s04.png]] --- ## Step 4 — Magic cancellation: ek full spin average hokar zero ho jaata hai (jab tak spin na ho) **KYA.** $\displaystyle\int_0^{2\pi} e^{ik\theta}\,d\theta$ evaluate karo integer $k$ ke liye. $$\int_0^{2\pi} e^{ik\theta}\,d\theta= \begin{cases} 2\pi & k=0\\[4pt] 0 & k\neq 0 \end{cases}$$ Har case kyun, *geometrically*: - Agar $k\neq 0$: arrow $e^{ik\theta}$ origin ke around $|k|$ complete baar spin karta hai jaise $\theta$ $0\to 2\pi$ jaata hai. Arrows ko add karna jo **equally har direction mein** point karte hain ek perfect cancellation deta hai — sum circle ka centre hai, yaani $0$. Har push ka ek equal-and-opposite push half turn baad hota hai. - Agar $k=0$: toh $e^{i\cdot 0\cdot\theta}=e^0=1$. Arrow *kabhi move nahin karta* — woh poore time sidhaa right point karta rehta hai. Constant $1$ ko $2\pi$ length ke interval par add karne se $2\pi$ milta hai. Kuch cancel nahin hota kyunki kuch spin hi nahin kiya. **KYun.** Yahi poore subject ka dil hai: **spinning cheezein cancel ho jaati hain, non-spinning cheezein accumulate hoti hain.** Integral ek lie detector hai jo har genuinely-spinning term ke liye $0$ return karta hai. **TASVEER.** Step 4 ki figure dono cases side by side rakhti hai: ek spinning arrow jiske tips centre par cancel ho jaate hain (left) versus ek frozen arrow jo pile up karta hai (right). ![[deepdives/dd-maths-4.10.05-d2-s05.png]] > [!formula] Woh ek decisive result > $$\oint_C (z-z_0)^n\,dz=\begin{cases}2\pi i & n=-1\\[3pt]0 & n\neq -1\end{cases}$$ > Exponent $k=n+1$ hai. Yeh zero hai — ek hi surviving case — **exactly tab jab $n=-1$** ho. (Bacha hua $i$ $dz$ ke andar ke $i$ se aata hai.) --- ## Step 5 — Kaun sa term bachta hai? Sirf $n=-1$. Residue se milo. **KYA.** Ab hum individual results ko wapas sum karte hain. Poori Laurent series ko loop mein daalo: $$\oint_C f(z)\,dz=\sum_{n=-\infty}^{\infty} a_n \underbrace{\oint_C (z-z_0)^n\,dz}_{\;=\,0\text{ har }n\ne -1\text{ ke liye}}=a_{-1}\cdot(2\pi i)$$ Ise padhna: - $n\neq -1$ waale har term ne $a_n\cdot 0=0$ contribute kiya. Gone. - Akele term $n=-1$ ne $a_{-1}\cdot 2\pi i$ contribute kiya. Toh poori, possibly-infinite function loop ke dwara **ek number times $2\pi i$** mein squeeze ho jaati hai: $$\boxed{\;\oint_C f(z)\,dz=2\pi i\,a_{-1}=2\pi i\,\operatorname{Res}(f,z_0)\;}$$ **KYun.** Isi liye $a_{-1}$ ko singled out karke naam diya gaya tha. Physically: yeh us ek flow pattern ki strength hai, $\tfrac{1}{z-z_0}$, jo *exactly ek baar per loop* circulate karta hai aur isliye kabhi cancel nahin hota. **TASVEER.** Step 5 ki figure terms ka tower dikhati hai, sab struck out siwaaye glowing $n=-1$ survivor ke. ![[deepdives/dd-maths-4.10.05-d2-s06.png]] > [!mistake] "Lekin swirling wala power kaun sa hai?" > Yeh sochtna tempting lagta hai ki sabse bada blow-up, $(z-z_0)^{-m}$, dominate karta hai. **Galat:** even ek strong term jaise $(z-z_0)^{-2}$ *spin* karta hai ($k=-1\ne 0$) aur poori tarah cancel ho jaata hai. Sirf $(z-z_0)^{-1}$ ka $k=0$ hota hai. Blow-up ki strength irrelevant hai; **sirf once-around swirl bachta hai.** --- ## Step 6 — Kai singularities: loop ko har ek par shrink karo **KYA.** Maan lo $C$ kayi bad points $z_1,\dots,z_k$ enclose karta hai. [[Cauchy's Integral Theorem]] ke zariye, hum big loop ko deform kar sakte hain — integral change kiye bina, jab tak hum kisi singularity ko cross na karein — kai tiny circles mein, ek har singularity ko hug karta hua, thin corridors se juda hua. Corridors **opposite directions mein do baar** trace ki jaati hain, isliye unke contributions cancel ho jaate hain. Jo bachta hai woh ek chhota loop per singularity hai, har ek apna $2\pi i\operatorname{Res}$ deta hai: $$\oint_C f(z)\,dz=2\pi i\sum_{j=1}^{k}\operatorname{Res}(f,z_j)$$ - $\sum_j$ **sirf $C$ ke andar waali singularities** ki swirl strengths add karta hai. - $C$ ke *bahar* ki singularity kuch bhi contribute nahin karti — loop kabhi usse circle nahin karta. **KYun.** Loops ke beech ka plane singularity-free hai, isliye wahan integrals vanish ho jaate hain; sirf andar trapped swirls register karte hain. **TASVEER.** Step 6 ki figure: ek bada loop teen chhote loops mein deform hua jo cancelling corridors se connected hain. ![[deepdives/dd-maths-4.10.05-d2-s07.png]] > [!mistake] Ek pole count karna jo contour ke bahar hai > Ek bilkul sahi residue *useless* hai agar woh singularity enclosed nahin hai. **Fix:** pehle $C$ draw karo, har singularity mark karo, sirf woh rakho jo loop ke andar hain, phir sum karo. --- ## Step 7 — Degenerate aur edge cases (taaki kuch surprise na kare) Har scenario jo reader ko mil sakta hai: > [!example] Andar koi singularity nahin ($k=0$) > Sum empty hai, toh $\oint_C f\,dz=0$. Yeh bas Cauchy's theorem hai — loop ko har jagah calm water feel hota hai. **Tasveer:** ek loop jahan $f$ bilkul theek hai woh kuch return nahin karta. > [!example] Removable singularity ($a_{-1}=0$ aur koi negative terms nahin) > Example $\tfrac{\sin z}{z}$ at $0$: Laurent series mein *koi* negative powers nahin hain, toh $a_{-1}=0$ aur loop integral $0$ hai. Point bura lag raha tha lekin swirl strength zero hai — ek "fake" whirlpool. > [!example] Higher-order pole, jaise $\tfrac{1}{(z-z_0)^2}$ > Yahan strongest term ka $n=-2$ hai, jo $k=-1\ne 0$ deta hai, isliye woh **cancel** ho jaata hai. Agar $a_{-1}=0$ bhi ho, toh poora loop integral $0$ hai — ek double pole ka bhi zero residue ho sakta hai. Pole ki order aur residue **independent** hain. > [!example] Essential singularity, jaise $e^{1/z}$ at $0$ > Infinitely many negative powers, phir bhi sirf $n=-1$ bachta hai. $e^{1/z}=\sum_{n\ge 0}\tfrac{1}{n!}z^{-n}$ se, $z^{-1}$ coefficient $\tfrac{1}{1!}=1$ hai, toh $\operatorname{Res}=1$ aur $\oint = 2\pi i$. **Koi pole formula apply nahin hota** — tum $a_{-1}$ seedha series se padh lete ho. > [!mistake] Radius-dependence anxiety > Gaur karo ki final answer mein **koi $\rho$ nahin tha** — $\rho^{n+1}$ surviving term ke liye vanish ho gaya ($n=-1\Rightarrow \rho^0=1$). Toh loop ka size matter nahin karta, bas yeh ki woh $z_0$ enclose kare. Agar tumhare answer mein abhi bhi $\rho$ hai, toh tumne ek aisa term rakha hai jo cancel ho jaana chahiye tha. --- ## Ek-tasveer summary Upar sab kuch ek single frame mein: Laurent tower loop mein feed hoti hai; loop ek filter hai; har spinning term zero mein drain ho jaata hai; akela $\tfrac{1}{z-z_0}$ swirl pass through karta hai, $2\pi i$ se scale hokar. ![[deepdives/dd-maths-4.10.05-d2-s08.png]] > [!recall]- Feynman retelling — poori walkthrough kisi dost ko explain karo > Ek pond ki picture karo jis mein ek jagah whirlpool hai. Tum jaanna chahte ho whirlpool kitna strong hai, lekin centre tak pahunch nahin sakte — wahan bahut gadbad hai. Toh tum uske around ek bada circle paddle karte ho aur net spin measure karte ho jo tumhe milta hai. > > Yeh trick hai: whirlpool ke paas ka messy paani kai simple flow patterns mein split ho sakta hai, har ek centre ke around ek alag number of times per lap swirling karta hua. Jab tum ek full lap paddle karte ho, koi bhi pattern jo *whole* number of times swirl karta hai ($1,2,-2,\dots$) tumhe equally har direction mein push karta hai — woh pushes perfectly cancel ho jaati hain aur tumhe unse kuch nahin milta. Sirf **ek** special pattern, jo exactly ek baar per lap circulate karta hai ($\tfrac{1}{z-z_0}$), kabhi cancel nahin hota. Uski strength woh number hai jise hum **residue** kehte hain. > > Toh ek lap around exactly $2\pi i$ times us residue return karta hai — tumhare circle ka size matter nahin karta, bas yeh ki tum whirlpool ke around ek baar gaye. Aur agar kayi whirlpools hain, toh tum bas unki saari strengths add kar lete ho. Yahi poora Residue Theorem hai. > [!mnemonic] > **"Sirf once-around swirl bachta hai."** Baaki sab spin hokar zero ho jaata hai; $a_{-1}$ survivor hai; ek lap tumhe $2\pi i\times$ woh deta hai. --- ## Connections - [[Laurent series]] — woh tower of terms supply kiya jise humne filter kiya. - [[Cauchy's Integral Theorem]] — woh reason ki spinning terms aur empty regions zero dete hain (Steps 4 & 6). - [[Cauchy's Integral Formula]] — special case $\oint \tfrac{f(z)}{z-z_0}dz = 2\pi i\,f(z_0)$. - [[Contour integration]] — jahan yeh machine real integrals par kaam mein lagayi jaati hai. - [[Argument principle]] / [[Rouché's theorem]] — in residues ka use karke zeros aur poles count karo. - [[Singularities]] — Step 7 ke cases ke peeche ki classification.