Further Mathematics
Matrix transformations and eigenvectors
- 1.
Column vectors are first reflected in y = x by S = [[0,1],[1,0]], then stretched parallel to x by D = [[2,0],[0,1]]. Find the combined matrix and image of (3,-1).
[3 marks] · no calculatorAnswer explanation
Draft walkthroughs are based on marking guidance, not independently verified derivations.
- Reflection swaps the coordinates to (-1,3); the stretch doubles only the first coordinate.
- Multiplying DS by the original column confirms the same image.
Marking points
- The product is DS, not SD.
- DS = [[0,2],[1,0]].
- The image is (-2,3).
Examiner tip: For column vectors the rightmost transformation acts first.
- 2.
A triangle of area 7 is transformed by [[2,1],[1,-1]]. Find its image area and state whether orientation is reversed.
[3 marks] · no calculatorAnswer explanation
Draft walkthroughs are based on marking guidance, not independently verified derivations.
- Areas scale by the absolute determinant, so a negative signed scale cannot produce negative area.
- The sign records orientation separately from the area scale: |-3| times 7.
Marking points
- The determinant is -3.
- Image area is 21 square units.
- Orientation is reversed because the determinant is negative.
Examiner tip: Use |det A| for area and det A's sign for orientation.
- 3.
Find the eigenvalues and corresponding eigenvector directions of A = [[3,1],[1,3]].
[3 marks] · no calculatorAnswer explanation
Draft walkthroughs are based on marking guidance, not independently verified derivations.
- A vector along y = x has both components multiplied by four.
- Along y = -x the components are multiplied by two; direct multiplication verifies both directions.
Marking points
- Characteristic equation: (3 - lambda)^2 - 1 = 0.
- Eigenvalues are 4 and 2.
- Directions are (1,1) for 4 and (1,-1) for 2.
Examiner tip: Associate each direction with its own eigenvalue; do not list unordered vectors.
- 4.
For A = [[k,2],[3,6]], find k for singularity. At that k, describe the image of the whole plane and the vectors sent to zero.
[3 marks] · no calculatorAnswer explanation
Draft walkthroughs are based on marking guidance, not independently verified derivations.
- At k = 1 the second row is three times the first, so two-dimensional inputs collapse onto one line.
- To find the kernel set the first output to zero; the second then vanishes automatically.
Marking points
- 6k - 6 = 0 gives k = 1.
- Outputs are (x + 2y, 3x + 6y), so the image is the line Y = 3X.
- The kernel is x + 2y = 0, or multiples of (-2,1).
Examiner tip: The image line and the kernel line are different objects.
- 5.
For A = [[3,1],[1,3]], derive A^n for positive integers n using its eigenvectors.
[3 marks] · no calculatorAnswer explanation
Draft walkthroughs are based on marking guidance, not independently verified derivations.
- The two eigenvector columns form a basis. Repeated action multiplies their coefficients by 4^n and 2^n respectively.
- Adjacent P^-1 P factors cancel in a repeated product; multiply the remaining three matrices to obtain the expression.
Marking points
- Use P = [[1,1],[1,-1]] and P^-1 = P/2.
- A = P diag(4,2) P^-1, hence A^n = P diag(4^n,2^n) P^-1.
- A^n = (1/2)[[4^n + 2^n,4^n - 2^n],[4^n - 2^n,4^n + 2^n]].
Examiner tip: A matrix power is not an entrywise power; diagonalise before raising entries.
- 6.
Let B = [[2,1],[0,2]]. Show that B is not diagonalisable and derive B^n for n >= 1.
[3 marks] · no calculatorAnswer explanation
Draft walkthroughs are based on marking guidance, not independently verified derivations.
- A repeated eigenvalue alone does not prevent diagonalisation; here the failure is the absence of a second independent eigenvector.
- All terms containing N^2 or a higher power vanish, leaving 2^n I + n*2^(n-1)N.
Marking points
- The only eigenvalue is 2 and its eigenspace is y = 0, of dimension one.
- Write B = 2I + N with N^2 = 0.
- The commuting binomial expansion gives B^n = [[2^n,n*2^(n-1)],[0,2^n]].
Examiner tip: Justify the binomial method by noting that I and N commute.
Marking points are indicative, not an official mark scheme. Accept equivalent valid methods and supported interpretations that address the task; award each mark once without requiring the model wording.