Intuition The big pictureA matrix A A A machine that transforms vectors: x ⃗ ↦ A x ⃗ \vec{x} \mapsto A\vec{x} x ↦ A x inverse A − 1 A^{-1} A − 1 undoes that transformation, sending the output back to the input. If A A A A − 1 A^{-1} A − 1 cannot undo it: when A A A squashes area to zero (flattens 2D into a line). That squashing is measured by the determinant . No area ⟹ information lost ⟹ no inverse.
Definition Inverse matrixThe inverse of a square matrix A A A A − 1 A^{-1} A − 1 A A − 1 = A − 1 A = I A A^{-1} = A^{-1} A = I A A − 1 = A − 1 A = I I I I has an inverse is called invertible (or non-singular ). If no inverse exists, A A A singular .
For a 2 × 2 2\times 2 2 × 2 A = ( a b c d ) A = \begin{pmatrix} a & b \\ c & d \end{pmatrix} A = ( a c b d )
We want A − 1 = ( p q r s ) A^{-1} = \begin{pmatrix} p & q \\ r & s \end{pmatrix} A − 1 = ( p r q s ) A A − 1 = I A A^{-1} = I A A − 1 = I
Why start here? Because "inverse" is defined by A A − 1 = I AA^{-1}=I A A − 1 = I
( a b c d ) ( p q r s ) = ( 1 0 0 1 ) \begin{pmatrix} a & b \\ c & d \end{pmatrix}\begin{pmatrix} p & q \\ r & s \end{pmatrix} = \begin{pmatrix} 1 & 0 \\ 0 & 1 \end{pmatrix} ( a c b d ) ( p r q s ) = ( 1 0 0 1 )
Multiplying out gives four equations :
a p + b r = 1 ( 1 ) a q + b s = 0 ( 2 ) c p + d r = 0 ( 3 ) c q + d s = 1 ( 4 ) \begin{aligned}
ap + br &= 1 \quad(1) &\qquad aq + bs &= 0 \quad(2)\\
cp + dr &= 0 \quad(3) &\qquad cq + ds &= 1 \quad(4)
\end{aligned} a p + b r c p + d r = 1 ( 1 ) = 0 ( 3 ) a q + b s c q + d s = 0 ( 2 ) = 1 ( 4 ) Solve for p , r p, r p , r
Why these two? They only involve p , r p,r p , r
From (3): c p + d r = 0 ⇒ r = − c p d cp + dr = 0 \Rightarrow r = -\dfrac{cp}{d} c p + d r = 0 ⇒ r = − d c p d ≠ 0 d\neq0 d = 0 a p + b ( − c p d ) = 1 ⇒ p a d − b c d = 1 ⇒ p = d a d − b c ap + b\left(-\frac{cp}{d}\right) = 1 \Rightarrow p\,\frac{ad - bc}{d} = 1 \Rightarrow p = \frac{d}{ad-bc} a p + b ( − d c p ) = 1 ⇒ p d a d − b c = 1 ⇒ p = a d − b c d
Then r = − c a d − b c r = -\dfrac{c}{ad-bc} r = − a d − b c c
Solve for q , s q,s q , s By identical algebra:
q = − b a d − b c , s = a a d − b c q = \frac{-b}{ad-bc}, \qquad s = \frac{a}{ad-bc} q = a d − b c − b , s = a d − b c a
The quantity a d − b c ad-bc a d − b c — call it the determinant det A \det A det A
Intuition Why "no inverse when det = 0"det A = a d − b c \det A = ad-bc det A = a d − b c signed area of the parallelogram spanned by the columns of A A A det A = 0 \det A = 0 det A = 0 A A A
Worked example Example 1 — a clean invertible matrixA = ( 2 1 3 4 ) A = \begin{pmatrix} 2 & 1 \\ 3 & 4 \end{pmatrix} A = ( 2 3 1 4 )
Step 1: determinant. det A = ( 2 ) ( 4 ) − ( 1 ) ( 3 ) = 8 − 3 = 5 \det A = (2)(4)-(1)(3) = 8-3 = 5 det A = ( 2 ) ( 4 ) − ( 1 ) ( 3 ) = 8 − 3 = 5 Why first? If it were 0 we'd stop immediately — no inverse.
Step 2: adjugate. Swap diagonal, negate off-diagonal: ( 4 − 1 − 3 2 ) \begin{pmatrix} 4 & -1 \\ -3 & 2\end{pmatrix} ( 4 − 3 − 1 2 ) Why? This is exactly the pattern from the derivation.
Step 3: divide by det.
A − 1 = 1 5 ( 4 − 1 − 3 2 ) A^{-1} = \frac{1}{5}\begin{pmatrix} 4 & -1 \\ -3 & 2\end{pmatrix} A − 1 = 5 1 ( 4 − 3 − 1 2 )
Step 4: verify A A − 1 = I AA^{-1}=I A A − 1 = I 1 5 ( 2 1 3 4 ) ( 4 − 1 − 3 2 ) = 1 5 ( 5 0 0 5 ) = I . \frac{1}{5}\begin{pmatrix} 2 & 1 \\ 3 & 4\end{pmatrix}\begin{pmatrix} 4 & -1 \\ -3 & 2\end{pmatrix} = \frac15\begin{pmatrix} 5 & 0 \\ 0 & 5\end{pmatrix}=I. 5 1 ( 2 3 1 4 ) ( 4 − 3 − 1 2 ) = 5 1 ( 5 0 0 5 ) = I . Why verify? Forecast-then-verify — cheap insurance against sign slips.
Worked example Example 2 — solving a linear system with the inverseSolve { 2 x + y = 5 3 x + 4 y = 6 \begin{cases} 2x + y = 5 \\ 3x + 4y = 6\end{cases} { 2 x + y = 5 3 x + 4 y = 6 A x ⃗ = b ⃗ A\vec{x}=\vec{b} A x = b A A A b ⃗ = ( 5 6 ) \vec b = \binom{5}{6} b = ( 6 5 )
Step 1: x ⃗ = A − 1 b ⃗ \vec x = A^{-1}\vec b x = A − 1 b Why? Multiply A x ⃗ = b ⃗ A\vec x=\vec b A x = b A − 1 A^{-1} A − 1 A − 1 A x ⃗ = x ⃗ A^{-1}A\vec x = \vec x A − 1 A x = x
Step 2: x ⃗ = 1 5 ( 4 − 1 − 3 2 ) ( 5 6 ) = 1 5 ( 20 − 6 − 15 + 12 ) = 1 5 ( 14 − 3 ) = ( 2.8 − 0.6 ) . \vec x = \dfrac{1}{5}\begin{pmatrix} 4 & -1 \\ -3 & 2\end{pmatrix}\binom{5}{6} = \frac15\binom{20-6}{-15+12} = \frac15\binom{14}{-3}=\binom{2.8}{-0.6}. x = 5 1 ( 4 − 3 − 1 2 ) ( 6 5 ) = 5 1 ( − 15 + 12 20 − 6 ) = 5 1 ( − 3 14 ) = ( − 0.6 2.8 ) .
Check: 2 ( 2.8 ) + ( − 0.6 ) = 5.6 − 0.6 = 5 2(2.8)+(-0.6)=5.6-0.6=5 2 ( 2.8 ) + ( − 0.6 ) = 5.6 − 0.6 = 5 3 ( 2.8 ) + 4 ( − 0.6 ) = 8.4 − 2.4 = 6 3(2.8)+4(-0.6)=8.4-2.4=6 3 ( 2.8 ) + 4 ( − 0.6 ) = 8.4 − 2.4 = 6
Worked example Example 3 — a singular matrixA = ( 2 4 1 2 ) A = \begin{pmatrix} 2 & 4 \\ 1 & 2\end{pmatrix} A = ( 2 1 4 2 ) det A = ( 2 ) ( 2 ) − ( 4 ) ( 1 ) = 4 − 4 = 0. \det A = (2)(2)-(4)(1)=4-4=0. det A = ( 2 ) ( 2 ) − ( 4 ) ( 1 ) = 4 − 4 = 0. Conclusion: No inverse exists . Notice row 2 is half of row 1 — the rows/columns are parallel, area collapsed. The formula would demand dividing by 0 0 0
Common mistake "The inverse is just the reciprocal of each entry."Why it feels right: For a single number, x − 1 = 1 / x x^{-1}=1/x x − 1 = 1/ x Why it's wrong: Matrix multiplication mixes rows and columns; undoing it needs the swap/negate/divide structure, not entrywise flips. Test ( 2 1 3 4 ) \begin{pmatrix}2&1\\3&4\end{pmatrix} ( 2 3 1 4 ) ( 1 / 2 1 1 / 3 1 / 4 ) \begin{pmatrix}1/2&1\\1/3&1/4\end{pmatrix} ( 1/2 1/3 1 1/4 ) not multiply to I I I Fix: Always use 1 det A ( d − b − c a ) \frac{1}{\det A}\begin{pmatrix} d&-b\\-c&a\end{pmatrix} d e t A 1 ( d − c − b a )
Common mistake "Forgot to check the determinant before inverting."Why it feels right: You're eager to apply the formula.
Fix: Compute det A \det A det A first . If it's 0 0 0
Common mistake "Negated the diagonal and swapped the off-diagonal (backwards!)."Why it feels right: "Something gets negated, something gets swapped" — easy to mix up which.
Fix: ==Swap the main diagonal, negate the off-diagonal.== Mnemonic below locks it in.
Mnemonic "Swap the corners, flip the middle, share the det."
Swap the corners: a ↔ d a \leftrightarrow d a ↔ d Flip the middle: put minus signs on b b b c c c Share the det: divide everything by a d − b c ad-bc a d − b c
Recall Feynman: explain it to a 12-year-oldImagine a machine that takes a shape and stretches/tilts it. The inverse machine puts the shape back exactly as it was. To build the "undo" machine for a 2×2 grid of numbers: swap the two numbers on the main diagonal, put minus signs on the other two, and shrink everything by a special number called the determinant. If that special number is zero , the machine flattened the shape into a line and lost so much information that no "undo" machine can ever rebuild it — like un-mixing paint. So: zero determinant = no undo.
What is the defining property of A − 1 A^{-1} A − 1 A A − 1 = A − 1 A = I AA^{-1}=A^{-1}A=I A A − 1 = A − 1 A = I (identity matrix).
Formula for the inverse of ( a b c d ) \begin{pmatrix}a&b\\c&d\end{pmatrix} ( a c b d ) 1 a d − b c ( d − b − c a ) \frac{1}{ad-bc}\begin{pmatrix}d&-b\\-c&a\end{pmatrix} a d − b c 1 ( d − c − b a ) .
What is the determinant of a 2×2 matrix ( a b c d ) \begin{pmatrix}a&b\\c&d\end{pmatrix} ( a c b d ) When does a 2×2 matrix have NO inverse? When
det A = a d − b c = 0 \det A = ad-bc = 0 det A = a d − b c = 0 (matrix is singular).
Geometric meaning of det A = 0 \det A = 0 det A = 0 Columns are parallel; the map flattens area to zero, losing information.
In the inverse recipe, what happens to the main diagonal entries a , d a,d a , d They are swapped.
In the inverse recipe, what happens to b b b c c c They are negated (sign flipped).
How do you solve A x ⃗ = b ⃗ A\vec x=\vec b A x = b x ⃗ = A − 1 b ⃗ \vec x = A^{-1}\vec b x = A − 1 b .
Why is "reciprocal of each entry" wrong for a matrix inverse? Because matrix multiplication mixes rows and columns; it doesn't multiply back to
I I I .
What is the adjugate of ( a b c d ) \begin{pmatrix}a&b\\c&d\end{pmatrix} ( a c b d ) ( d − b − c a ) \begin{pmatrix}d&-b\\-c&a\end{pmatrix} ( d − c − b a ) , and
A − 1 = 1 det A adj A A^{-1}=\frac{1}{\det A}\text{adj}\,A A − 1 = d e t A 1 adj A .
swap diagonal negate off-diagonal
A times A inverse equals I
Intuition Hinglish mein samjhoDekho, matrix A A A A − 1 A^{-1} A − 1 A A − 1 = I AA^{-1}=I A A − 1 = I det A = a d − b c \det A = ad-bc det A = a d − b c a a a d d d b b b c c c det A \det A det A
Sabse important baat: agar det A = 0 \det A = 0 det A = 0
Practical use: agar tumhe equations solve karni hain, jaise 2 x + y = 5 2x+y=5 2 x + y = 5 3 x + 4 y = 6 3x+4y=6 3 x + 4 y = 6 A x ⃗ = b ⃗ A\vec x=\vec b A x = b x ⃗ = A − 1 b ⃗ \vec x = A^{-1}\vec b x = A − 1 b
