Foundations — Diagonalization — conditions, procedure

4.5.31 · D1 · Maths › Linear Algebra (Full) › Diagonalization — conditions, procedure

Is page pe assume kiya gaya hai ki parent note ke notation ke baare mein tumhe kuch bhi nahi pata. Hum har symbol ek ek karke banate hain, har ek pehle wale pe tikaa hua. Agar koi line aisa symbol use kare jo humne abhi tak earn nahi kiya, to wo ek bug hai — flag karo. Chalte hain ek arrow se.

1. Ek vector — the arrow

Picture: figure dekho. Origin se point $(3,1)$ tak ka arrow hi vector $\left(\begin{array}{c}3\\ 1\end{array}\right)$ hai. Number $x_1=3$ horizontal axis par uska shadow hai; $x_2=1$ vertical axis par uska shadow hai.

Topic ko iske kyu zarurat hai: matrix jo bhi karta hai, wo arrows ke saath karta hai. Diagonalization ek statement hai ki kaunse arrows simply behave karte hain, isliye arrows hamare pehle object hain. Dekho Change of Basis jab same arrow ko graph paper tilne par alag number-pairs se describe kiya ja sakta hai.

2. Ek matrix — the machine

Machine ek arrow ko kaise khaati hai (matrix–vector product): $A\mathbf{x}=\left(\begin{array}{cc}4& 1\\ 2& 3\end{array}\right)\left(\begin{array}{c}{x}_1\\ x_2\end{array}\right)=\left(\begin{array}{c}4x_1+1x_2\\ 2x_1+3x_2\end{array}\right).$

Picture: left pe grid normal graph paper hai. Right pe machine $A$ ne use warp kar diya hai — squares tilted parallelograms ban jaate hain. Zyaadaatar arrows ki dono length aur direction badal gayi.

Topic ko iske kyu zarurat hai: pura subject hai "samajhna ki ye machine kya karti hai." Ek general machine ek saath twist aur stretch karti hai — mushkil. Diagonalization ek aisa viewpoint dhundhta hai jahan wo sirf stretch kare.

3. Eigen-arrow — special direction

Picture: orange arrow ek ordinary arrow hai — $A$ ke baad ye ek nayi direction mein swing ho jaata hai (faded orange). Blue arrow ek eigenvector hai — $A$ ke baad ye apni hi line pe land karta hai, bas lamba. Ye "apni hi line pe rehna" hi $A\mathbf{x}=\lambda \mathbf{x}$ ka pura matlab hai.

• Agar $\lambda >1$: eigen-arrow badhta hai.
• Agar $0<\lambda <1$: ye chhhota hota hai.
• Agar $\lambda =1$: ye unchanged rehta hai (fixed).
• Agar $\lambda <0$: ye origin ke doosri side flip ho jaata hai, phir scale hota hai.
• Agar $\lambda =0$: arrow origin pe crush ho jaata hai (ye direction null direction hai).

Topic ko iske kyu zarurat hai: ye exactly wo "special arrows" hain jinke baare mein core idea baat karta hai. Diagonalization = unhe itna collect karna ki space rebuild ho sake. Poora detail Eigenvalues and Eigenvectors mein hai.

4. $\lambda I$, $A-\lambda I$, aur $\det$ — eigen-arrows kaise dhundhte hain

Hum $\lambda I$ kyun subtract karte hain. Eigen-equation ko rewrite karo, sab kuch ek side le jao: $A\mathbf{x}=\lambda \mathbf{x}⟹A\mathbf{x}-\lambda \mathbf{x}=0⟹(A-\lambda I)\mathbf{x}=0.$ Beech wale step ko $\lambda \mathbf{x}=\lambda I\mathbf{x}$ chahiye — isliye $I$ aata hai: tum ek plain number $\lambda$ ko matrix $A$ se subtract nahi kar sakte, lekin tum matrix $\lambda I$ ko kar sakte ho subtract. Ye number ko ek legal machine mein convert karta hai.

Topic ko iske kyu zarurat hai: ye puri procedure ka Step 1 hai. Left side $\lambda$ mein ek polynomial hai — Characteristic Polynomial. Uske roots stretch factors hain.

5. Null space, kernel, aur $\ker$ — jahan eigenvectors rehte hain

To ek given $\lambda$ ke liye eigenvectors exactly $\ker (A-\lambda I)$ ke nonzero arrows hain. Ye hai Step 2: $(A-\lambda I)\mathbf{x}=0$ solve karo.

Topic ko iske kyu zarurat hai: parent ki "geometric multiplicity" $g(\lambda )=\dim \ker (A-\lambda I)$ literally hai "ye eigenvalue kitne independent eigen-arrows supply karta hai." No kernel, no eigenvectors, no diagonalization.

6. Linear independence — "koi arrow redundant nahi"

Topic ko iske kyu zarurat hai: parent ka main theorem kehta hai $A$ diagonalizable hai iff uske paas $n$ linearly independent eigenvectors hain. Ab tum samajh gaye ki wo word itna load-bearing kyun hai.

7. Invertible, $P^{-1}$, aur $\text{diag}$ — answer assemble karna

Topic ko iske kyu zarurat hai: goal $A=PD{P}^{-1}$ padhta hai: "eigen-basis change undo karo ($P$), pure stretch karo ($D$), phir eigen-basis change apply karo ($P$)." Ruko — direction ke saath careful raho. $P$ ke columns mein eigenvectors hain, isliye $P$ ek arrow ko uske eigen-coordinates se build karta hai. $A\mathbf{x}=PD{P}^{-1}\mathbf{x}$ right-to-left padhne par: $P^{-1}$, $\mathbf{x}$ ko eigen-coordinates mein likhta hai, $D$ har ek ko stretch karta hai, $P$ ordinary arrow rebuild karta hai. Power payoff — $A^{100}=P{D}^{100}{P}^{-1}$ — isliye koi care karta hai; dekho Matrix Powers and Exponentials.

8. Similarity — relationship $A=PD{P}^{-1}$

Diagonalization similarity ka special case hai jahan doosra matrix $D$ diagonal hota hai. Jab koi diagonal $D$ kaam nahi karta, to humhara sabse close option Jordan Normal Form hai; jab $A$ symmetric hota hai, Spectral Theorem ek sundar diagonalization guarantee karta hai.

Foundations topic ko kaise feed karte hain

Equipment checklist

Har cloze padho, zor se jawab do, phir reveal karo.