Matrices & Determinants — Introduction
Level 3 — Production Paper
Time limit: 45 minutes
Total marks: 60
Instructions: Show all derivations from scratch. Where an "explain" is requested, write your reasoning in full sentences. Use / for mathematics.
Q1. (10 marks) From-scratch derivation of the inverse.
(a) Let . Starting from the requirement where , solve the resulting linear systems to derive the formula for . State clearly the condition for existence. (7)
(b) Explain out loud (in words) why a square matrix with cannot have an inverse, referencing your derivation. (3)
Q2. (12 marks) Symmetric / skew-symmetric decomposition.
(a) Prove that any square matrix can be written uniquely as where is symmetric and is skew-symmetric. Derive expressions for and . (6)
(b) Apply your result to , giving and explicitly. (4)
(c) State what all diagonal entries of any skew-symmetric matrix must equal, and justify from the definition. (2)
Q3. (12 marks) Determinant of a via cofactor expansion + properties.
Let
(a) Compute by cofactor expansion along the first column. Show every cofactor. (6)
(b) Recompute by expansion along a different row or column of your choice and confirm the values agree. (4)
(c) Explain out loud why expansion along any row or column gives the same answer. (2)
Q4. (10 marks) Non-commutativity and transpose properties.
(a) For , , compute and and state whether they are equal. (4)
(b) Verify the property for these matrices by computing both sides. (4)
(c) Explain out loud why matrix multiplication is generally not commutative, referring to the order/dimension conditions. (2)
Q5. (16 marks) Solving a system — two methods.
Consider the system
(a) Solve using Cramer's rule. Show , , and the final values. (6)
(b) Solve the same system using matrix inversion (), computing from scratch. (7)
(c) Explain out loud one condition under which neither method gives a unique solution, and what it means geometrically. (3)
Answer keyMark scheme & solutions
Q1 (10)
(a) gives two systems (one per column of ): Column 1: , . Column 2: , . (1 for setting up systems)
Solve column 1: from second eqn (or eliminate). Using elimination: , . (2) Column 2 similarly: , . (2)
Hence (1) Condition: exists iff . (1)
(b) The formula divides by ; if the division is undefined, so no matrix satisfies . Geometrically the rows/columns are linearly dependent, so collapses dimensions and cannot be undone. (3: any valid reasoning referencing derivation)
Q2 (12)
(a) Suppose with , . Take transpose: . (2) Add: . (1) Subtract: . (1) Check ✓, ✓. (1) Uniqueness: the two equations determine from uniquely. (1)
(b) . . (2) . (2)
(c) Diagonal entries are all . From , entry . (2)
Q3 (12)
(a) Expand along column 1 (entries ): . . (2) . (2) (middle term zero.) (2)
(b) Expand along row 3 (entries ): . Matches. (4)
(c) The cofactor expansion along any line is a way of grouping the same signed permutation terms of the determinant; the determinant is a single well-defined value independent of the chosen line. (2)
Q4 (10)
(a) . (2) . (1) . (1)
(b) . (1) , . . (2) Equal ✓. (1)
(c) Multiplication combines rows of the first with columns of the second; swapping order changes which rows meet which columns (and may not even be defined dimensionally), so results differ in general. (2)
Q5 (16)
(a) , . . (2) . (1) . (1) , . (2)
(b) , . (3) . (4) (matches (a))
(c) If , neither Cramer's rule nor inversion works (division by zero / no inverse). Geometrically the two lines are parallel — either coincident (infinitely many solutions) or distinct (no solution). (3)
[
{"claim":"det of Q3 matrix A = -1","code":"A=Matrix([[2,-1,3],[0,4,5],[1,0,2]]); result=(A.det()==-1)"},
{"claim":"Q4 AB != BA","code":"A=Matrix([[1,2],[0,1]]); B=Matrix([[1,0],[3,1]]); result=(A*B!=B*A)"},
{"claim":"Q4 (AB)^T = B^T A^T","code":"A=Matrix([[1,2],[0,1]]); B=Matrix([[1,0],[3,1]]); result=((A*B).T==B.T*A.T)"},
{"claim":"Q5 solution x=5/7 y=17/7","code":"A=Matrix([[3,2],[1,-4]]); b=Matrix([7,-9]); X=A.inv()*b; result=(X==Matrix([Rational(5,7),Rational(17,7)]))"},
{"claim":"Q2 S+K reconstructs M","code":"M=Matrix([[4,3],[7,2]]); S=(M+M.T)/2; K=(M-M.T)/2; result=(S+K==M and S.T==S and K.T==-K)"}
]