4.5.23
Intuition Ek-sentence ideaKisi matrix ka determinant aapko us box (parallelepiped) ka signed volume batata hai jo uske column (ya row) vectors se banta hai — yeh batata hai ki woh linear map space ko kitna stretch karta hai , aur kya woh orientation flip karta hai .
Intuition Volume hi kyun?Ek matrix A linear map hai x ↦ A x det A Sign record karta hai ki map handedness (orientation) preserve karta hai ya reverse.
KYA chahiye hame: ek function V ( a , b ) a , b
KAISE pin karte hain: demand karo ki yeh geometric area jaisa behave kare.
Definition Axioms jo signed area ko follow karne chahiye
Multilinear : ek edge ko scale karne se area scale hota hai: V ( c a , b ) = c V ( a , b ) V ( a + a ′ , b ) = V ( a , b ) + V ( a ′ , b ) Alternating : do equal edges = flat box = zero area: V ( a , a ) = 0 Normalized : unit square ka area 1 V ( e 1 , e 2 ) = 1
Maano a = a 1 e 1 + a 2 e 2 b = b 1 e 1 + b 2 e 2
Multilinearity use karke expand karo:
V ( a , b ) = a 1 b 1 V ( e 1 , e 1 ) + a 1 b 2 V ( e 1 , e 2 ) + a 2 b 1 V ( e 2 , e 1 ) + a 2 b 2 V ( e 2 , e 2 )
Yeh step kyun? Har vector basis vectors ka combination hai; multilinearity scalars ko term by term bahar nikalne deti hai.
Ab alternating apply karo: V ( e 1 , e 1 ) = V ( e 2 , e 2 ) = 0 V ( a + b , a + b ) = 0 V ( e 2 , e 1 ) = − V ( e 1 , e 2 ) = − 1
V ( a , b ) = a 1 b 2 − a 2 b 1 = det ( a 1 a 2 b 1 b 2 )
det A > 0 counterclockwise ko counterclockwise rakhta hai → orientation preserved .det A < 0 reflect karta hai (jaise aaina) → orientation reversed .det A = 0 flatten ho gaya → vectors linearly dependent hain, map non-invertible hai.
Worked example 1 — Ek simple stretchA = ( 3 0 0 2 ) det A = 3 ⋅ 2 − 0 = 6 Yeh step kyun? Diagonal map x y 3 × 2 +
Worked example 2 — Ek reflection (negative determinant)A = ( 0 1 1 0 ) x y det = 0 ⋅ 0 − 1 ⋅ 1 = − 1 Yeh step kyun? Axes ko swap karna y = x ∣ − 1 ∣ = 1 −
Worked example 3 — Ek collapsed boxa = ( 2 , 4 ) b = ( 1 , 2 ) = 2 1 a det = 2 ⋅ 2 − 4 ⋅ 1 = 0 Yeh step kyun? b a det = 0 ⟺
Worked example 4 — 3D triple producta = ( 1 , 0 , 0 ) , b = ( 0 , 1 , 0 ) , c = ( 0 , 0 , 1 ) det = 1 c = ( 0 , 0 , − 1 ) det = − 1 Yeh step kyun? c
Common mistake "Negative determinant matlab negative volume — yeh impossible hai, toh maine galti ki."Kyun sahi lagta hai: Everyday life mein volumes positive hote hain. Fix: Magnitude ∣ det ∣ sign bonus information hai (orientation). Negative bilkul valid aur meaningful hai.
Common mistake "Non-square matrix ka determinant uska 'volume' deta hai."Kyun sahi lagta hai: Hum vector sets ke volume ki baat karte hain. Fix: det square matrix chahiye — same-dimensional box ke liye equal number of vectors aur dimensions. n k det ( A ⊤ A )
Common mistake "Agar main ek vector ko 2 se scale karoon, toh determinant sirf woh... ruko, kya poori cheez double hoti hai?"Kyun sahi lagta hai: Ek edge ko scale karna area ko linearly scale karta hai. Fix: det ( c A ) = c n det A n × n saare n columns ek saath scale karte ho.
Recall Compute karne se pehle predict karoQ: Bina compute kiye, kya det ( 1 2 2 4 ) Forecast: rows/columns proportional hain (( 2 , 4 ) = 2 ( 1 , 2 ) zero .
Verify: 1 ⋅ 4 − 2 ⋅ 2 = 0
Recall Quick self-test (answers cover karo)
∣ det A ∣ Sign kya batata hai? → kya orientation preserved (+) hai ya reversed (−).
det A = 0
∣ det A ∣ A
det A Orientation — positive = preserved, negative = reversed (reflection).
det A = 0 Spanned box ek lower dimension mein flatten ho jata hai → columns linearly dependent hain → map non-invertible hai.
Teen axioms jo determinant formula force karte hain Har column mein Multilinear, alternating (repeated columns par zero), normalized (det I = 1
3D signed volume as a triple product det [ a b c ] = a ⋅ ( b × c )
Jab poori matrix ko c det det ( c A ) = c n det A n × n
Axioms se 2D determinant formula det ( a 1 a 2 b 1 b 2 ) = a 1 b 2 − a 2 b 1
Recall Feynman: ek 12-saal ke bachche ko samjhaoEk rubber sheet imagine karo jisme ek chhota square bana hua hai. Matrix ek machine hai jo poori sheet ko stretch aur shear karti hai. Square ek tirche box mein badal jata hai. Determinant bas yeh hai: "naya box original square se kitne guna bada hai?" Agar machine sheet ko bhi ulta flip kar de (jaise page palat rahe ho), toh hum aage minus sign lagate hain. Aur agar machine square ko bilkul flat karke ek line mein squash kar de, toh jawab zero hai — kuch nahi bacha.
"SIZE aur SIDE." ∣ det ∣ SIZE (volume scaling); sign = kaun si SIDE (orientation, flipped hai ya nahi). Zero = squished .
multilinear alternating normalized
Linearly dependent columns
