Foundations — Finding eigenspaces

4.5.30 · D1 · Maths › Linear Algebra (Full) › Finding eigenspaces

Pehle aapko un khaas arrows ki talash mein jaane se pehle, us alphabet mein fluent hona padega jis par parent note baat karta hai. Neeche har symbol aur idea hai jo uspar depend karta hai, bilkul zero se build kiya gaya hai, har ek apni jagah earn karta hai usse pehle ki agli cheez aaye.

1. Vectors — origin par tail wale arrows

Bold letter $\mathbf{v}$ ka matlab hai "poora arrow", koi single number nahi. Plain numbers $3$ aur $1$ andar uske components hain — arrow kitna right aur kitna upar jaata hai.

Arrows ki zaroorat kyun hai? Kyunki parent note jo kuch bhi karta hai — stretching, rotating, ek line par rehna — yeh sab directions aur lengths ke baare mein statements hain, aur ek arrow ek direction with a length ka sabse clean picture hai.

Special arrow $0=\left(\begin{array}{c}0\\ 0\end{array}\right)$ zero vector hai: length zero ka ek arrow, kisi taraf point nahi karta. Ise dhyan mein rakho — yeh parent note ka sabse subtle rule create karta hai ("$0$ kabhi eigenvector nahi hota").

2. Vectors ko add karna — parallelogram rule

Abhi yeh introduce kyun karein? Kyunki baad ke ideas — "kuch arrows ke saare combinations" (properly Section 12 mein span aur subspace ke roop mein define kiye gaye) — literally arrows ko stretch aur add karke bante hain. Aap "in arrows ke saare combinations" ke baare mein baat nahi kar sakte jab tak aap unhe add nahi kar sakte, isliye addition pehle aata hai.

3. Scalars aur scalar multiplication — ek arrow ko stretch karna

Picture hi poori story hai:

• $\lambda =2$ → arrow do guna lamba, same direction.
• $\lambda =\frac{1}{2}$ → arrow aadha lamba, same direction.
• $\lambda =-1$ → arrow flip hokar exactly ulti taraf point karta hai.
• $\lambda =0$ → arrow origin par collapse ho jaata hai ($0$).

Notice karo: scalar multiplication arrow ko kabhi rotate nahi karta (negative $\lambda$ par flip ko chodkar, jo phir bhi same line par hai). Yahi wajah hai ki eigenvectors "apni khud ki line par tikay rehte hain."

4. Matrices — woh machine jo arrows ko move karti hai

Ek matrix sirf storage nahi hai — yeh ek function hai jo ek vector khaata hai aur ek vector produce karta hai. Rule (matrix–vector multiplication) ek $2\times 2$ ke liye:

Chalte hain machine ko apne example $A$ par kaam karte hue dekhte hain: $A\left(\begin{array}{c}3\\ 1\end{array}\right)=\left(\begin{array}{c}2\cdot 3+1\cdot 1\\ 1\cdot 3+2\cdot 1\end{array}\right)=\left(\begin{array}{c}7\\ 5\end{array}\right).$

Output arrow $\left(\genfrac{}{}{0ex}{}{7}{5}\right)$ alag direction mein point karta hai $\left(\genfrac{}{}{0ex}{}{3}{1}\right)$ se — machine ne use rotate kar diya. Zyaadaatar arrows rotate ho jaate hain. Parent note poochhta hai: woh rare arrows kaun se hain jo same way point karte hue bahar aate hain, sirf rescale kiye hue? Woh eigenvectors hain.

5. Matrices ko add aur subtract karna

Yahi operation $A-\lambda I$ ke andar chupi hai: aap ek grid ko doosre se position by position subtract karoge. Example: $\left(\begin{array}{cc}2& 1\\ 1& 2\end{array}\right)-\left(\begin{array}{cc}3& 0\\ 0& 3\end{array}\right)=\left(\begin{array}{cc}-1& 1\\ 1& -1\end{array}\right).$

6. Identity matrix $I$ — "kuch nahi karne waali" machine

Hume $I$ kyun chahiye? Yahan ek short derivation hai jo ise existence mein force karti hai. Hum non-turning condition $A\mathbf{v}=\lambda \mathbf{v}$ ko "kuch matrix times $\mathbf{v}$ equals $0$" ke roop mein rewrite karna chahte hain, kyunki (jaise Section 7 dikhata hai) woh shape woh hai jo hum solve karna jaante hain.

Isliye $A-\lambda I$ sense karta hai jabki "$A-\lambda$" nahi karta. Yahi exact derivation parent note karta hai; hum ise yahan reproduce karte hain taaki yeh page akele khad ho sake.

7. Homogeneous systems aur null space — jahan arrows zero par crush ho jaate hain

Jo equation hum ne abhi derive ki, $(A-\lambda I)\mathbf{v}=0$, ek homogeneous system hai: ek matrix times ek unknown arrow equals zero vector. ("Homogeneous" ka matlab sirf yeh hai ki right side sab zeros hai.)

Do arrows hamesha yahan free mein rehte hain: $0$ khud (har matrix $0\to 0$ bhejtaa hai), aur sambhavtah doosron ki poori line ya plane.

Toh eigenspace $E_{\lambda }$ exactly ek null space hai — isliye agar aap null space dhundhna jaante hain, toh aap eigenspace dhundhna bhi jaante hain. Aisi systems solve karne ki poori machinery ke liye, Null space and solving homogeneous systems dekhein.

8. Row-reduction — woh tool jo in systems ko solve karta hai

Actually ek null space dhundhne ke liye aapko ek mechanical skill chahiye: ek matrix ko ek simpler, staircase shape mein turn karna bina uska solution set badley.

Hum ise immediately Section 11 mein use karenge ek matrix ko measure karne ke liye.

9. Determinant — ek single number ka test "kya yeh machine kuch crush karti hai?"

Uska picture mein matlab: $|\det M|$ machine ka area-scaling factor hai. Agar ek machine areas ko double karti hai, $\det =\pm 2$. Agar machine 2D arrows ko ek single line par flatten karti hai (area $0$ ho jaata hai), toh $\det =0$ — machine singular kehlaati hai.

10. Characteristic polynomial — jahan $\det (A-\lambda I)$ $\lambda$ mein ek equation ban jaata hai

Jab aap actually $\det (A-\lambda I)$ compute karte hain, toh unknown $\lambda$ diagonal par appear hota hai, toh determinant $\lambda$ mein ek polynomial ke roop mein bahar aata hai — $\lambda$ ki powers ek saath add hoti hain. Chalte hain ek general $2\times 2$ ke liye dekhte hain taaki form earn ki jaaye, yaad na ki jaaye.

Degree generally $n$ kyun hoti hai? Kyunki $n$ diagonal entries mein se har ek ek factor $(\text{entry}-\lambda )$ contribute karta hai, aur $n$ aisi factors ko multiply karna highest power ${\lambda }^n$ deta hai. Full details: Characteristic polynomial.

11. Multiplicity — ek eigenvalue kitni baar "appear" karna chahiye

12. Span, subspace, basis, dimension — "poori line/plane of arrows" describe karna

Hum ne yeh words Section 2 mein promise ki theen; yahan hain, adding (Section 2) aur stretching (Section 3) se bane.

Kyun parent note care karta hai: ek eigenspace sirf ek arrow nahi balki non-turning arrows ka ek poora subspace hai. Uski dimension (geometric multiplicity, Section 11) woh hai jise parent Example 2 (ek full plane, $\dim =2$) aur Example 3 (sirf ek line, $\dim =1$) mein alag karta hai. Aur kyunki $E_{\lambda }=\mathrm{Null}(A-\lambda I)$ (Section 7), ek eigenspace sach mein ek subspace hai — ek null space hamesha (a), (b), (c) satisfy karta hai.

13. Rank, aur Rank–Nullity theorem — eigenspace ko size karna

Ab ki hum row-reduce kar sakte hain (Section 8), hum ek matrix measure kar sakte hain.

14. Greek letters aur shorthand jo aap milenge

Foundations topic ko kaise feed karte hain

Neeche ka map bottom-up padho: koi bhi arrow "$X\to Y$" ko "$Y$ se pehle $X$ chahiye" padho. Top boxes (raw ideas) se shuru karo aur neeche single goal box tak trace karo — har path jo aap walk kar sakte hain ek prerequisite hai jo ab aapke paas hai.

Single node sabse neeche, Finding eigenspaces, parent topic hai — aur ab yeh fully powered hai.

Equipment checklist

Khud test karo: right side cover karo aur reveal karne se pehle har ek answer do.