Legendre's equation and Legendre polynomials (intro): Visual walkthrough

4.6.20 · D2 · Maths › Ordinary Differential Equations › Legendre's equation and Legendre polynomials (intro)

Hum sab kuch banate hain. Agar koi symbol aata hai, toh pehle usse draw kiya gaya tha.

Step 1 — Equation ko ek sentence ki tarah padho

KYA HAI. $y = y (x)$ ek unknown function hai jo hum dhundhna chahte hain. $y^{'}$ uski slope hai (kitni tez upar jaati hai), $y^{''}$ uski curvature hai (slope khud kitni tez change hoti hai). $n$ ek number hai jo hum pehle se choose karte hain — isse ek dial ki tarah socho. Poori line kehti hai: har point $x$ par, yeh teen pieces zero ho jaane chahiye.

YEH TARIKE SE KYO PADHEIN. Notice karo ki $y^{''}$ ke aage jo coefficient hai woh $(1 - x^{2})$ hai. $x = 0$ par yeh $1$ ke barabar hai — bilkul theek. $x = \pm 1$ par yeh $0$ ho jaata hai — curvature term gayab ho jaata hai, aur equation wahan "control kho deti hai". Woh do jagah, $x = \pm 1$ , physics mein sphere ke poles hain. Is topic mein jo bhi dramatic hota hai, woh sab us vanishing ki wajah se hota hai.

PICTURE. $(1 - x^{2})$ ka curve interval ke upar — ek dome jo exactly donoN siron par zero ko touch karta hai.

Figure — Legendre's equation and Legendre polynomials (intro)

Kaun se

x

-values

y^{''}

ke coefficient ko vanish karte hain?

x = \pm 1

, woh do poles.

Step 2 — Power series guess karo, aur batao kyun hum aisa karne ke liye allowed hain

KYA HAI. Har $a_{m}$ bas ek number hai (ek knob). Knobs ghuma do aur tumhe alag alag functions milenge. Hamara kaam: woh knob-settings dhundho jo equation ko satisfy karayen.

SERIES KYO, AUR YAHAN KYO. Ek "healthy" point ke paas ek smooth function ko hamesha aisi sum ki tarah likha ja sakta hai — yeh Power Series / Frobenius Method ka idea hai. Hum ek plain Taylor series use karne ke liye allowed hain (koi fractional powers nahi) precisely kyunki $x = 0$ ek ordinary point hai — $y^{''}$ coefficient $(1 - x^{2})$ wahan nonzero hai (Step 1). Agar hum $x = 1$ ke around expand karte, toh humhe heavier Frobenius machinery chahiye hoti.

YEH TOOLS KYO — ek sum ke derivatives. Equation mein plug karne ke liye hume $y^{'}$ aur $y^{''}$ chahiye. Ek power term differentiate karna ek rule hai jo humhe chahiye: $\frac{d}{d x} x^{m} = m x^{m - 1}$ . Toh $y^{'} = \sum_{m} m a_{m} x^{m - 1}, y^{''} = \sum_{m} m (m - 1) a_{m} x^{m - 2} .$ Har derivative power ko lower karta hai aur purani power se multiply karta hai — yeh yaad rakho, yahi poora engine hai.

PICTURE. Basic bricks $1, x, x^{2}, x^{3}$ ka ek stack, har ek apne knob $a_{m}$ se scaled, jo milke ek wiggly curve banaate hain.

Step 3 — Substitute karo, aur har term ko same power par line up karo

KYA HAI. Teeno pieces ko equation mein daalo: $(1 - x^{2}) \sum m (m - 1) a_{m} x^{m - 2} - 2 x \sum m a_{m} x^{m - 1} + n (n + 1) \sum a_{m} x^{m} = 0.$

LINE UP KYO KARO. Powers ka sum har $x$ ke liye zero ho, iske liye zaroori hai ki har ek individual power $x^{m}$ ka weight apne aap zero ho. (Tum ek $x^{2}$ ko $x^{5}$ se cancel nahi kar sakte — woh independent bricks hain.) Toh hume har sum ko dobara likhna hoga taaki woh "coefficient times $x^{m}$ " padhe, same power, phir woh coefficient read karein.

HUM HAR PIECE KE SATH KYA KARTE HAIN — term by term.

$\sum m (m - 1) a_{m} x^{m - 2}$ : power $x^{m - 2}$ hai, do bahut kam. Dummy index ko do aage jump karwa ke relabel karo ( $m \to m + 2$ ) taaki power $x^{m}$ ban jaaye: $\sum_{m} m (m - 1) a_{m} x^{m - 2} = \sum_{m} new front factor (m + 2) (m + 1) a_{m + 2} x^{m} .$
$- x^{2} \sum m (m - 1) a_{m} x^{m - 2} = - \sum m (m - 1) a_{m} x^{m}$ — pehle se $x^{m}$ hai, chhod do.
$- 2 x \sum m a_{m} x^{m - 1} = - \sum 2 m a_{m} x^{m}$ — pehle se $x^{m}$ hai.
$n (n + 1) \sum a_{m} x^{m}$ — pehle se $x^{m}$ hai.

PICTURE. Terms ke chaar ribbons, teeno pehle se " $x^{m}$ " shelf par park hain, pehla wala do shelves slide karke unke saath join ho raha hai.

Step 4 — $x^{m}$ ka coefficient read karo: recurrence ka janam

KYA HAI. $x^{m}$ ko multiply karne wali sab cheez collect karo aur zero set karo: $from y^{''} (m + 2) (m + 1) a_{m + 2} from - x^{2} y^{''} - m (m - 1) a_{m} from - 2 x y^{'} - 2 m a_{m} from + n (n + 1) y + n (n + 1) a_{m} = 0.$

KYO. Yeh Step 3 ka "weight of $x^{m}$ is zero" rule hai, concrete bana diya. Ab teeno $a_{m}$ terms collect karo: $- m (m - 1) - 2 m + n (n + 1) = - [m (m - 1) + 2 m] + n (n + 1) = n (n + 1) - m (m + 1),$ kyunki $m (m - 1) + 2 m = m^{2} + m = m (m + 1)$ .

PICTURE. Ek conveyor belt: $a_{m}$ feed karo, ratio-box use multiply karta hai, out aata hai $a_{m + 2}$ — line mein do steps aage, kabhi ek nahi.

Step 5 — Do independent chains: even aur odd

KYA HAI. General solution hai $y = a_{0} built from a_{0} [even series] + a_{1} built from a_{1} [odd series] .$

YEH KYO MATTER KARTA HAI. Kyunki chains alag hain, ek solution purely even ho sakta hai ( $a_{1} = 0$ set karo) ya purely odd ( $a_{0} = 0$ set karo). Yahi parity fact $P_{n} (- x) = (- 1)^{n} P_{n} (x)$ source par show up kar raha hai.

PICTURE. Do seedhiyaan side by side — ek green even ladder ( $a_{0}, a_{2}, a_{4}, \dots$ ) aur ek orange odd ladder ( $a_{1}, a_{3}, a_{5}, \dots$ ) — har rung use jo do upar hai use feed karti hai.

Step 6 — Miracle: integer $n$ ke liye, ek chain mar jaati hai

KYA HOTA HAI. Agar $m = n$ numerator ko zero karta hai, toh $a_{n + 2} = 0$ . Aur kyunki chain aage multiply karta hai, $a_{n + 4} = 0$ , $a_{n + 6} = 0$ , … — us parity chain mein $a_{n}$ ke baad har coefficient vanish ho jaata hai. Infinite series ruk jaati hai. Woh exactly degree $n$ ki polynomial ban jaati hai.

YEH POORA POINT KYO HAI. Integer dial $n$ ke liye, do chains mein se ek ek sahi degree- $n$ polynomial mein truncate ho jaati hai — wahi ==Legendre polynomial $P_{n} (x)$ == hai. Non-integer $n$ ke liye, numerator kabhi zero nahi hit karta, series hamesha chalti rehti hai, aur (yeh dikhaya ja sakta hai) woh poles $x = \pm 1$ par blow up karti hai. Physics poles par finite temperature/potential demand karta hai, toh woh force karta hai ki $n \in {0, 1, 2, \dots}$ ho. Yahi selection rule hai.

PICTURE. Coefficient chain aage badh rahi hai, har bar pichle se chhota, phir — $m = n$ par — snap karke zero ho jaata hai aur hamesha flat rehta hai.

Series tab terminate kyun hoti hai jab

n

ek non-negative integer hota hai?

m = n

par numerator

m (m + 1) - n (n + 1) = 0

ho jaata hai, toh

a_{n + 2} = 0

aur saare upar wale same-parity coefficients mar jaate hain — ek degree-

n

polynomial bachi rehti hai.

Step 7 — Crank ghuma: scratch se $P_{2}$ banao

NORMALIZE KYO KARO. Ek free knob $a_{0}$ bachi rehti hai — ODE akele uska size fix nahi kar sakta. Hum ise universal convention $P_{n} (1) = 1$ se pin karte hain: $a_{0} (1 - 3 \cdot 1^{2}) = a_{0} (- 2) = 1 \Rightarrow a_{0} = - \frac{1}{2} .$ Isliye $P_{2} (x) = - \frac{1}{2} (1 - 3 x^{2}) = \frac{1}{2} (3 x^{2} - 1) . ✓$

PICTURE. Parabola $\frac{1}{2} (3 x^{2} - 1)$ $[- 1, 1]$ ke upar: centre par $- \frac{1}{2}$ tak neeche jaati hai, donoN siroN $x = \pm 1$ par exactly $1$ hit karti hai (normalization kaam karte hue).

Step 8 — Edge case jo hum bhool nahi sakte: doosra solution

$Q_{n}$ KYA KARTA HAI. Woh terminate nahi karta, aur jaise $x \to \pm 1$ , woh bina bound ke badhta hai — poles par ek logarithmic blow-up. Concretely $Q_{0} (x) = \frac{1}{2} ln \frac{1 + x}{1 - x}$ , jo $x \to \pm 1$ as $\pm \infty$ tak shoot karta hai.

HUM USE KYO PHENK DETE HAIN. Temperature, potential, aur gravity sphere ke poles par finite rehni chahiye. $Q_{n}$ yeh violate karta hai. Toh do mathematical solutions mein se, physics bounded $P_{n}$ rakhti hai aur wild $Q_{n}$ discard kar deti hai. Yeh wahi "bounded-at-poles" rule hai jisne $n$ ko integer hone par force kiya tha.

PICTURE. $[- 1, 1]$ par do curves: shant, finite $P_{0} = 1$ (flat) aur runaway $Q_{0}$ jo siron par $\pm \infty$ ki taraf dive karta hai.

Ek-picture summary

Sab kuch ek canvas par: dome $(1 - x^{2})$ poles par vanish hoti hai → series guess → two-step recurrence → even/odd chains → $m = n$ par truncation → polynomial $P_{n}$ , apne wild twin $Q_{n}$ ke saath ek taraf phenka hua.

Recall Feynman retelling — poora walkthrough seedhe shabdon mein

Hum ek aisi function chahte the jo $- 1$ se $+ 1$ tak ek line par ek fussy rule maane. Function directly guess karne ki jagah humne kaha "yeh simplest building bricks ka kuch mix hai — ek flat bit, ek slope bit, ek bend bit, aur aise hi" aur amounts ko knobs chhod diye. Us mix ko rule mein plug karke aur demand karke ki woh har point ke liye hold kare, ek sahi law force hua: do steps aage wala knob current ka ek fixed multiple hai. Toh flat/bend/quartic knobs ek family banate hain, aur slope/cubic knobs doosri. Ab magic: jab bhi hamara chosen dial $n$ ek whole number hota hai, woh multiplier exactly step $n$ par exactly zero hit karta hai — toh knobs ki chain achanak ruk jaati hai. Endless mix ek choti, neat polynomial mein collapse hoti hai: ek Legendre polynomial. Aur solutions ki doosri family? Woh kabhi nahi rukti, aur line ke donoN siron par — sphere ke poles par — blast ho jaati hai. Kyunki real temperatures aur voltages poles par infinite nahi ho sakti, nature polite polynomial rakh leti hai aur cheekh maarte wale ko phenk deti hai. "Poles par finite raho" ka woh ek rule hi secretly demand karta hai ki $n$ pehli jagah se ek whole number ho.

Connections

Power Series / Frobenius Method — Steps 2–4 ke peechhe ki machinery.
Sturm-Liouville Theory — kyun surviving $P_{n}$ mutually orthogonal nikalta hai.
Laplace's Equation in Spherical Coordinates — jahan dial $n$ aur $x = cos θ$ paida hote hain.
Fourier Series — wahi "orthogonal building blocks mein expand karo" spirit.
Generating Functions — same $P_{n}$ tak ek slick alternate route.
Associated Legendre Functions & Spherical Harmonics — agla chapter, jab sphere equator ke around bhi vary karta hai.