MATRICES AND DETERMINANTS

(1) MATRIX:
𝑎11 𝑎12 𝑎13
(i) (𝑎21 𝑎22 𝑎23 ) , 𝑤ℎ𝑒𝑟𝑒 𝑎𝑖𝑗 => 𝑖𝑡ℎ 𝑟𝑜𝑤 𝑎𝑛𝑑 𝑗 𝑡ℎ 𝑐𝑜𝑙𝑢𝑚𝑛. Order 𝑚 × 𝑛
𝑎31 𝑎32 𝑎33
(ii) Types: Square Matrix (m=n), Horizontal matrix (n>m), Vertical matrix (m>n),
row matrix (m=1), column matrix (n=1), diagonal matrix: (if 𝑎𝑖𝑗 = 0, 𝑖 ≠ 𝑗)
1 0 0
diagonal: 𝑑𝑖𝑎𝑔(1, 3, 5) = [0 3 0], scalar matrix: diagonal matrix with 𝑑1 = 𝑑2 = 𝑑3
0 0 5
If d=1, then identity matrix. Upper triangular matrix: 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑖 > 𝑗, 𝑎𝑖𝑗 = 0
Lower triangular matrix: for all 𝑖 < 𝑗, 𝑎𝑖𝑗 = 0
𝑛(𝑛−1)
(iii) Minimum number of zeroes in triangular matrix: 𝑤ℎ𝑒𝑟𝑒 𝑛 × 𝑛 𝑖𝑠 𝑜𝑟𝑑𝑒𝑟
2
𝑛(𝑛+1)
For max number of distinct elements: 0 is also a distinct element: + 1 = 𝑚𝑎𝑥
2

(2) OPERATIONS ON MATRICES:

(i) Equality: same order and same corresponding elements. Same with addition and
subtraction=> commutative, associative and cancellation holds.
(ii) Multiplying a scalar: multiply every element.
(iii) Matrix Multiplication: 𝐹𝑜𝑟 𝐴𝑚×𝑛 × 𝐵𝑝×𝑞 , ℎ𝑒𝑟𝑒 𝑛 = 𝑝, 𝑎𝑛𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑎𝑛𝑡 𝑚𝑎𝑡𝑟𝑖𝑥 𝐶𝑚×𝑞
Multiply each row with other matrix ka columns.
(iv) Properties of multiplication:
(a) 𝐴𝐵 ≠ 𝐵𝐴
(b) distributive and associative law holds.
(c) 𝑖𝑓 𝐴𝐵 = 0 𝑖𝑡 𝑑𝑜𝑒𝑠 𝑛𝑜𝑡 𝑚𝑒𝑎𝑛 𝐴 = 0, 𝐵 = 0
(d) 𝐴𝐼 = 𝐼𝐴 = 𝐴 𝑎𝑛𝑑 𝐼𝐼 = 𝐼 = 𝐼2
(3) SOME SPECIAL MATRICES:
(i) Transpose of a matrix: Matrix arranged by exchanging rows and columns.
5 2 𝑇 5 6
So ( ) =( )
6 −1 2 −1
Properties:
(a) (𝐴𝑇 )𝑇 = 𝐴
(b) (𝐾𝐴)𝑇 = 𝐾𝐴𝑇 where K is a scalar quantity.
(c) (𝐴 ± 𝐵)𝑇 = 𝐴𝑇 ± 𝐵𝑇
(d) Rearrangement law: (𝐴𝐵 )𝑇 = 𝐵𝑇 𝐴𝑇 , similarly (𝐴𝐵𝐶 )𝑇 = 𝐶 𝑇 𝐵𝑇 𝐴𝑇
(ii) Trace of a matrix: Sum of diagonal elements.
Properties:
(a) 𝑇𝑟(𝑝𝐴) = 𝑝 𝑇𝑟(𝐴) and 𝑇𝑟(𝐴 ± 𝐵) = 𝑇𝑟(𝐴) ± 𝑇𝑟(𝐵)
(b) 𝑇𝑟(𝐴𝐵) = 𝑇𝑟(𝐵𝐴)
(iii) Symmetric matrix: corresponding column and row values are equal. 𝐴 = 𝐴𝑇
(iv) Skew symmetric matrix: corresponding column & row values are -ve of each other.
and 𝐴 = −𝐴𝑇 . Diagonal elements are always negative in skew symmetric.
For any matrix: (𝐴 + 𝐴𝑇 ) = 𝑠𝑦𝑚𝑚𝑒𝑡𝑟𝑖𝑐, 𝑎𝑛𝑑 (𝐴 − 𝐴𝑇 ) = 𝑠𝑘𝑒𝑤 𝑠𝑦𝑚𝑚𝑒𝑡𝑟𝑖𝑐
Any matrix can be written as sum of a skew symmetric and symmetric matrix.
1 1
Eg: 𝐴 = (2𝐴) = ((𝐴 + 𝐴𝑇 ) + (𝐴 − 𝐴𝑇 ))
2 2
(v) Orthogonal matrix: 𝐴𝑇 𝐴 = 𝐴𝐴𝑇 = 𝐼
(vi) Idempotent matrix: 𝐴2 = 𝐴
(vii) Involutory matrix: 𝐴2 = 𝐼
(viii) Periodic matrix: 𝐴𝑘+1 = 𝐼, 𝑡ℎ𝑒𝑛 𝑝𝑒𝑟𝑖𝑜𝑑 = 𝑘
(ix) Nilpotent matrix: 𝐴𝑘 = 0, 𝑡ℎ𝑒𝑛 𝑖𝑛𝑑𝑒𝑥 = 𝑘
(4) DETERMINANTS:
(i) 𝑀𝑖𝑛𝑜𝑟: 𝑀𝑖𝑗 obtained by removing that row and column.
Cofactor: 𝐶𝑖𝑗 = (−1)𝑖+𝑗 𝑀𝑖𝑗
Determinant (D) = 𝑎11 𝐶11 + 𝑎12 𝐶12 + 𝑎13 𝐶13 and so on…, any row or column can be used.
The value of: 𝑎11 𝐶21 + 𝑎12 𝐶22 + 𝑎13 𝐶23 = 0
(ii) Multiplying a scalar: gets multiplied in any row/column. Same way to take column.
(iii) Properties:
(a) 𝐷 = 0 𝑖𝑓 𝑎𝑛𝑦 𝑟𝑜𝑤 𝑜𝑟 𝑐𝑜𝑙𝑢𝑚𝑛 𝑎𝑟𝑒 0 𝑜𝑟 2 𝑟𝑜𝑤𝑠 𝑜𝑟 𝑐𝑜𝑙𝑢𝑚𝑛𝑠 𝑎𝑟𝑒 𝑒𝑞𝑢𝑎𝑙. 𝑂𝑟 𝑖𝑛 𝑒𝑞𝑢𝑎𝑙 𝑟𝑎𝑡𝑖𝑜.
(b) For each exchange of a row/column, one negative sign comes.
(c) Elementary row/column transformations do not change value of determinant.
Like: 𝑅1 → 𝑅1 ± 𝛼𝑅2 ± 𝛽𝑅3 , same with column
(d) |𝐴𝐵| = |𝐴||𝐵|; |𝐴𝑇 | = |𝐴|; |𝑘𝐴| = 𝑘 𝑛 |𝐴|, 𝑤ℎ𝑒𝑟𝑒 𝑛 𝑖𝑠 𝑜𝑟𝑑𝑒𝑟 𝑜𝑓 𝐴
(e) det(𝑠𝑘𝑒𝑤 𝑠𝑦𝑚𝑚𝑒𝑡𝑟𝑖𝑐 𝑚𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑑𝑑 𝑜𝑟𝑑𝑒𝑟) = 0
(f) If det = 0, then that matrix is a singular matrix. Product of diagonal columns is det of a
diagonal matrix.
𝑎+𝑥 𝑏+𝑦 𝑐+𝑧 𝑎 𝑏 𝑐 𝑥 𝑦 𝑧
(iv) Sum of determinants: | 𝑝 𝑞 𝑟 | = |𝑝 𝑞 𝑟 | + | 𝑝 𝑞 𝑟 |
𝑓 𝑔 ℎ 𝑓 𝑔 ℎ 𝑓 𝑔 ℎ
𝑓 (𝑟 ) 𝑔 (𝑟 ) ℎ (𝑟 ) ∑ 𝑓 (𝑟 ) ∑ 𝑔 (𝑟 ) ∑ ℎ (𝑟 )
𝑛
(v) If 𝐷𝑟 = | 𝑎 𝑏 𝑐 | 𝑡ℎ𝑒𝑛 ∑𝑟=1 𝐷𝑟 = | 𝑎 𝑏 𝑐 |
𝑑 𝑒 𝑓 𝑑 𝑒 𝑓
(vi) Product of determinants => same as matrices but row by row, column by column allowed.
𝑓 (𝑥) 𝑓2 (𝑥) 𝑓 ′ (𝑥) 𝑓2′ (𝑥) 𝑓 (𝑥) 𝑓2 (𝑥)
(vii) Differentiation: ∆(𝑥) = | 1 | , ∆ ′ (𝑥 ) = | 1 | + | 1′ |
𝑔1 (𝑥) 𝑔2 (𝑥) 𝑔1 (𝑥) 𝑔2 (𝑥) 𝑔1 (𝑥) 𝑔2′ (𝑥)
(viii) If elements of a determinant are rational integral functions of x, and two rows or
columns become identical when x = a, then (𝑥 − 𝑎)𝑖𝑠 𝑎 𝑓𝑎𝑐𝑡𝑜𝑟 of determinant.
1 𝑥 𝑥2
(ix) General Expansions: |1 𝑦 𝑦 2 | = (𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥)
1 𝑧 𝑧2
1 𝑥 𝑥3
And : |1 𝑦 𝑦 3 | = (𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥)(𝑥 + 𝑦 + 𝑧)
1 𝑧 𝑧3
 1 𝑥2 𝑥3
And: : |1 𝑦 2 𝑦 3 | = (𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥)(𝑥𝑦 + 𝑦𝑧 + 𝑧𝑥)
1 𝑧2 𝑧3
𝑎 𝑏 𝑐
And: : |𝑏 𝑐 𝑎| = −(𝑎3 + 𝑏3 + 𝑐 3 − 3𝑎𝑏𝑐)
𝑐 𝑎 𝑏
(5) ADJOINT:
𝐶11 𝐶12 𝐶13 𝑇
(i) Adjoint matrix is the transpose of cofactor matrix: [𝐶21 𝐶22 𝐶23 ]
𝐶31 𝐶32 𝐶33
(ii) Properties:
(a) 𝐴(𝑎𝑑𝑗 𝐴) = (𝑎𝑑𝑗 𝐴)𝐴 = |𝐴|𝐼
(b) |𝑎𝑑𝑗 𝐴| = |𝐴|𝑛−1 where n is order of matrix.
(c) 𝑎𝑑𝑗 (𝑎𝑑𝑗 𝐴) = |𝐴|𝑛−2 𝐴
2
(d) |𝑎𝑑𝑗 (𝑎𝑑𝑗 𝐴)| = |𝐴|(𝑛−1)
(e) 𝑎𝑑𝑗 (𝐴𝐵) = (𝑎𝑑𝑗 𝐵)(𝑎𝑑𝑗 𝐴)
(f) 𝑎𝑑𝑗 𝐴𝑚 = (𝑎𝑑𝑗 𝐴)𝑚 and 𝑎𝑑𝑗 (𝑘𝐴) = 𝑘 𝑛−1 𝑎𝑑𝑗 𝐴
(6) INVERSE:
1
(i) If 𝐴𝐵 = 𝐵𝐴 = 𝐼 𝑡ℎ𝑒𝑛 𝐵 𝑖𝑠 𝑖𝑛𝑣𝑒𝑟𝑠𝑒 𝑜𝑓 𝐴. 𝐴−1 = |𝐴| 𝑎𝑑𝑗 𝐴
𝑎 𝑏 𝑑 −𝑏
(ii) If 𝐴 = [ ] 𝑡ℎ𝑒𝑛 𝑎𝑑𝑗 𝐴 = [ ], switch diagonal elements & give -ve sign (b & c)
𝑐 𝑑 −𝑐 𝑎
(iii) Properties of Inverse:
(a) (𝐴𝐵)−1 = 𝐵−1 𝐴−1 𝑎𝑛𝑑 𝑠𝑖𝑚𝑖𝑙𝑎𝑟𝑙𝑦 (𝐴𝑇 )−1 = (𝐴−1 )𝑇
1
(b) |𝐴−1 | = |𝐴|
(iv) If AB = null matrix then 𝑒𝑖𝑡ℎ𝑒𝑟 𝐴 = 𝑛𝑢𝑙𝑙, 𝐵 = 𝑛𝑢𝑙𝑙 𝑜𝑟 𝑏𝑜𝑡ℎ = 𝑛𝑢𝑙𝑙, 𝑜𝑟 𝑛𝑜𝑛𝑒 = 𝑛𝑢𝑙𝑙
Necessary condition, both A and B must be singular matrices.
(7) CAYLEY HAMILTON THEOREM:
(i) Every square matrix satisfies a polynomial equation called characteristic equation.
(ii) 𝑃 (𝜆) = |𝐴 − 𝜆𝐼| which is 0 when 𝜆 = 𝐴
(iii) Through this we can get a relation like 𝜆2 − 3𝜆 + 5 = 0 𝑤ℎ𝑖𝑐ℎ ℎ𝑎𝑠 𝑟𝑜𝑜𝑡 𝐴
So multiplying 𝐴−1 𝑏𝑜𝑡ℎ 𝑠𝑖𝑑𝑒𝑠 𝑤𝑒 𝑔𝑒𝑡 𝑡ℎ𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴−1 𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 𝐴
(iv) Sum of roots of characteristic equation = trace of matrix
Product of roots = determinant value
Degree = order of matrix
(8) GAUSS-JORDAN METHOD:
(i) For 𝑎1 𝑥 + 𝑏1 𝑦 + 𝑐1 = 0, 𝑎2 𝑥 + 𝑏2 𝑦 + 𝑐2 = 0
𝑎1 𝑏1 𝑎1 𝑏1 𝑐1 𝑎1 𝑏1 𝑐1
If ≠ 𝑡ℎ𝑒𝑛 𝑢𝑛𝑖𝑞𝑢𝑒 𝑠𝑜𝑙; 𝑖𝑓 = ≠ 𝑡ℎ𝑒𝑛 𝑛𝑜 𝑠𝑜𝑙; 𝑖𝑓 = = 𝑡ℎ𝑒𝑛 ∞ 𝑠𝑜𝑙
𝑎2 𝑏2 𝑎2 𝑏2 𝑐2 𝑎2 𝑏2 𝑐2

(ii) For three equations three variables:

𝑎1 𝑏1 𝑐1 𝑥 𝑑1
We define: 𝐴 = [𝑎2 𝑏2 𝑐2 ] 𝑎𝑛𝑑 𝑋 = [𝑦] 𝑎𝑛𝑑 𝐵 = [𝑑2 ]; relation: 𝐴𝑋 = 𝐵 or 𝑋 = 𝐴−1 𝐵
𝑎3 𝑏3 𝑐3 𝑧 𝑑3
(a) ∆≠ 0, 𝑡ℎ𝑒𝑛 𝐴−1 𝑒𝑥𝑖𝑠𝑡𝑠 𝑠𝑜 𝑢𝑛𝑖𝑞𝑢𝑒 𝑠𝑜𝑙
(b) ∆= 0 𝑡ℎ𝑒𝑛 𝐴−1 𝑑𝑜𝑒𝑠 𝑛𝑜𝑡 𝑒𝑥𝑖𝑠𝑡: 𝑖𝑓 (𝑎𝑑𝑗 𝐴)𝐵 = 𝑛𝑢𝑙𝑙 𝑡ℎ𝑒𝑛 ∞ 𝑠𝑜𝑙.
If (𝑎𝑑𝑗 𝐴)𝐵 ≠ 𝑛𝑢𝑙𝑙 𝑡ℎ𝑒𝑛 𝑛𝑜 𝑠𝑜𝑙
(9) CRAMER’S RULE:
𝑎1 𝑏1 𝑐1 𝑑1 𝑏1 𝑐1 𝑎1 𝑑1 𝑐1 𝑎1 𝑏1 𝑑1
(i) Let us define: ∆= |𝑎2 𝑏2 𝑐2 | , ∆𝑥 = |𝑑2 𝑏2 𝑐2 | , ∆𝑦 = |𝑎2 𝑑2 𝑐2 | , ∆𝑧 = |𝑎2 𝑏2 𝑑2 |,
𝑎3 𝑏3 𝑐3 𝑑3 𝑏3 𝑐3 𝑎3 𝑑3 𝑐3 𝑎3 𝑏3 𝑑3
Then 𝑥. ∆= ∆𝑥 ; 𝑦. ∆= ∆𝑦 ; 𝑧. ∆= ∆𝑧
(ii) Case 1: If ∆≠ 0 𝑡ℎ𝑒𝑛 𝑢𝑛𝑖𝑞𝑢𝑒 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
(iii) Case 2: If ∆= 0 𝑎𝑛𝑑 𝑎𝑛𝑦 𝑜𝑛𝑒 𝑜𝑟 𝑚𝑜𝑟𝑒 𝑜𝑓 (∆𝑥 , ∆𝑦 , ∆𝑧 ) ≠ 0, 𝑡ℎ𝑒𝑛 𝑛𝑜 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
(iv) Case 3: If ∆= 0 𝑎𝑛𝑑 𝑎𝑙𝑙 𝑜𝑓(∆𝑥 , ∆𝑦 , ∆𝑧 ) = 0, 𝑡ℎ𝑒𝑛 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛𝑠 𝑎𝑟𝑒 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡.
We take any 2 of the 3 equations and proceed. Take one of the variables as K and solve for
other variables. If the equations are solvable then ∞ 𝑠𝑜𝑙, 𝑖𝑓 𝑛𝑜𝑡: 𝑛𝑜 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
(v) For homogenous equations: 𝑎1 𝑥 + 𝑏1 𝑦 + 𝑐1 𝑧 = 0; 𝑎2 𝑥 + 𝑏2 𝑦 + 𝑐2 𝑧 = 0; 𝑎3 𝑥 + 𝑏3 𝑦 + 𝑐3 𝑧
(a) Case one if ∆≠ 0, 𝑡ℎ𝑒𝑛 (𝑥, 𝑦, 𝑧) = 0 (𝑡𝑟𝑖𝑣𝑖𝑎𝑙 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛)
(b) Case two if ∆= 0, 𝑡ℎ𝑒𝑛 ∞ 𝑣𝑎𝑙𝑢𝑒𝑠 𝑜𝑓 (𝑥, 𝑦, 𝑧) (𝑛𝑜𝑛 − 𝑡𝑟𝑖𝑣𝑖𝑎𝑙 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛)
(vi) Unique solution or ∞ 𝑠𝑜𝑙. 𝑎𝑟𝑒 𝑐𝑎𝑙𝑙𝑒𝑑 𝑐𝑜𝑛𝑠𝑖𝑠𝑡𝑒𝑛𝑡: 𝑛𝑜 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 => 𝑖𝑛𝑐𝑜𝑛𝑠𝑖𝑠𝑡𝑒𝑛𝑡
(10) 3-D VISUALISATION OF SYSTEM OF EQUATIONS:
(i) 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 = 𝑑 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑠 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑝𝑙𝑎𝑛𝑒 (𝑖𝑓 𝑜𝑛𝑙𝑦 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑠 𝑎𝑟𝑒 𝑑𝑖𝑓𝑓, || 𝑝𝑙𝑎𝑛𝑒𝑠)
(ii) Family of planes: 𝑃1 + 𝜆𝑃2 = 0

(iii) 𝐼𝑓 𝐷1 𝑖𝑠 𝑓𝑜𝑟𝑚𝑒𝑑 𝑏𝑦 𝑟𝑒𝑝𝑙𝑎𝑐𝑖𝑛𝑔 𝑒𝑙𝑒𝑚𝑒𝑛𝑡𝑠 𝑜𝑓 𝐷 𝑤𝑖𝑡ℎ 𝑡ℎ𝑒𝑖𝑟 𝑐𝑜𝑓𝑎𝑐𝑡𝑜𝑟𝑠, 𝐷1 = 𝐷 𝑛−1