Mathematics: Applications & Interpretation HL
Matrices and Markov chains — Topic 1 HL
- 1.
Given A = [[2, 3], [−1, 4]] and B = [[5, −2], [0, 3]], find (a) A + B and (b) A − B.
[2 marks]Worked explanation
- Add corresponding entries: A+B=[[2+5,3-2],[-1+0,4+3]]=[[7,1],[-1,7]].
- Subtract entrywise: A-B=[[-3,5],[-1,1]]. Adding B back to this difference recovers A.
Common mistake: Matrix addition is entrywise, unlike matrix multiplication.
Marking points
- Adds corresponding entries to obtain A + B = [[7, 1], [−1, 7]].
- Subtracts corresponding entries to obtain A − B = [[−3, 5], [−1, 1]].
Examiner tip: Matrix addition and subtraction only combine entries in the same position — the matrices must have identical dimensions for this to be defined.
- 2.
Marking analysis: A learner attempts the following task: “Given A = [[2, 3], [−1, 4]] and B = [[5, −2], [0, 3]], find (a) A + B and (b) A − B.” Their response addresses only this point: “Adds corresponding entries to obtain A + B = [[7, 1], [−1, 7]].” Evaluate the response against the complete 2-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[2 marks]Marking points
- Recognises credit for the stated point: Adds corresponding entries to obtain A + B = [[7, 1], [−1, 7]].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Subtracts corresponding entries to obtain A − B = [[−3, 5], [−1, 1]].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 3.
Given A = [[1, 2], [3, 4]] and B = [[2, 0], [1, 3]], find the matrix product AB.
[3 marks]Worked explanation
- Each product entry is a row-column dot product: first row gives (1*2+2*1,1*0+2*3)=(4,6).
- Second row gives (3*2+4*1,3*0+4*3)=(10,12), so AB=[[4,6],[10,12]], a 2 by 2 matrix.
Common mistake: Multiplying corresponding entries is not the matrix product.
Marking points
- Multiplies row 1 of A by each column of B: (1×2 + 2×1, 1×0 + 2×3) = (4, 6).
- Multiplies row 2 of A by each column of B: (3×2 + 4×1, 3×0 + 4×3) = (10, 12).
- States AB = [[4, 6], [10, 12]].
Examiner tip: Remember that matrix multiplication is generally not commutative — AB and BA can give different results, so always keep track of the order.
- 4.
Marking analysis: A learner attempts the following task: “Given A = [[1, 2], [3, 4]] and B = [[2, 0], [1, 3]], find the matrix product AB.” Their response addresses only this point: “Multiplies row 1 of A by each column of B: (1×2 + 2×1, 1×0 + 2×3) = (4, 6).” Evaluate the response against the complete 3-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[3 marks]Marking points
- Recognises credit for the stated point: Multiplies row 1 of A by each column of B: (1×2 + 2×1, 1×0 + 2×3) = (4, 6).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Multiplies row 2 of A by each column of B: (3×2 + 4×1, 3×0 + 4×3) = (10, 12).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: States AB = [[4, 6], [10, 12]].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 5.
Find the determinant of the matrix A = [[4, 7], [2, 6]].
[2 marks]Worked explanation
- For [[a,b],[c,d]], determinant=ad-bc.
- det A=4(6)-7(2)=24-14=10, nonzero, so A is invertible.
Common mistake: The off-diagonal product is subtracted, not added.
Marking points
- Uses det(A) = ad − bc = (4)(6) − (7)(2).
- Obtains det(A) = 10.
Examiner tip: For a 2×2 matrix [[a, b], [c, d]], the determinant ad − bc is always computed as the product of the leading diagonal minus the product of the other diagonal.
- 6.
Marking analysis: A learner attempts the following task: “Find the determinant of the matrix A = [[4, 7], [2, 6]].” Their response addresses only this point: “Uses det(A) = ad − bc = (4)(6) − (7)(2).” Evaluate the response against the complete 2-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[2 marks]Marking points
- Recognises credit for the stated point: Uses det(A) = ad − bc = (4)(6) − (7)(2).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains det(A) = 10.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 7.
Find the inverse of the matrix A = [[3, 5], [1, 2]].
[3 marks]Worked explanation
- det A=3(2)-5(1)=1. Swap the diagonal entries and negate the off-diagonal entries, then divide by 1.
- A^-1=[[2,-5],[-1,3]]. Multiplication gives AA^-1=[[6-5,-15+15],[2-2,-5+6]]=I.
Common mistake: Taking reciprocals of entries does not invert a matrix.
Marking points
- Calculates det(A) = (3)(2) − (5)(1) = 1.
- Uses the inverse formula A⁻¹ = (1/det(A))[[d, −b], [−c, a]].
- Obtains A⁻¹ = [[2, −5], [−1, 3]].
Examiner tip: Swap the leading diagonal entries, negate the other two entries, then divide every entry by the determinant — a matrix only has an inverse when its determinant is non-zero.
- 8.
Marking analysis: A learner attempts the following task: “Find the inverse of the matrix A = [[3, 5], [1, 2]].” Their response addresses only this point: “Calculates det(A) = (3)(2) − (5)(1) = 1.” Evaluate the response against the complete 3-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[3 marks]Marking points
- Recognises credit for the stated point: Calculates det(A) = (3)(2) − (5)(1) = 1.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Uses the inverse formula A⁻¹ = (1/det(A))[[d, −b], [−c, a]].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains A⁻¹ = [[2, −5], [−1, 3]].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 9.
Use a matrix method to solve the simultaneous equations 2x + 3y = 12 and x − y = 1.
[5 marks]Worked explanation
- Write [[2,3],[1,-1]][x,y]^T=[12,1]^T. The coefficient determinant is -5 and its inverse is [[1/5,3/5],[1/5,-2/5]].
- Multiplying gives x=3,y=2. Check 2(3)+3(2)=12 and 3-2=1.
Common mistake: The right-hand-side vector must match the equation order.
Marking points
- Writes the system in matrix form [[2, 3], [1, −1]][x, y]ᵀ = [12, 1]ᵀ.
- Calculates the determinant of the coefficient matrix: (2)(−1) − (3)(1) = −5.
- Finds the inverse of the coefficient matrix.
- Multiplies the inverse by the constants vector [12, 1]ᵀ.
- Obtains x = 3, y = 2.
Examiner tip: The matrix method for solving simultaneous equations works regardless of how many variables there are, as long as the coefficient matrix has a non-zero determinant — a GDC can perform the inverse and multiplication directly.
- 10.
Marking analysis: A learner attempts the following task: “Use a matrix method to solve the simultaneous equations 2x + 3y = 12 and x − y = 1.” Their response addresses only this point: “Writes the system in matrix form [[2, 3], [1, −1]][x, y]ᵀ = [12, 1]ᵀ.” Evaluate the response against the complete 5-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[5 marks]Marking points
- Recognises credit for the stated point: Writes the system in matrix form [[2, 3], [1, −1]][x, y]ᵀ = [12, 1]ᵀ.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the determinant of the coefficient matrix: (2)(−1) − (3)(1) = −5.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Finds the inverse of the coefficient matrix.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Multiplies the inverse by the constants vector [12, 1]ᵀ.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains x = 3, y = 2.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 11.
Find the determinant of the 3×3 matrix A = [[1, 2, 3], [0, 1, 4], [5, 6, 0]].
[4 marks]Worked explanation
- Expand along the first row with alternating signs: det A=1(1*0-4*6)-2(0*0-4*5)+3(0*6-1*5).
- This is -24+40-15=1. As a check, row operations R3-5R1 then R3+4R2 produce diagonal entries 1,1,1 without changing the determinant.
Common mistake: The middle cofactor carries a minus sign.
Marking points
- Expands along the first row: det(A) = 1(1×0 − 4×6) − 2(0×0 − 4×5) + 3(0×6 − 1×5).
- Evaluates each 2×2 minor: 1(−24) − 2(−20) + 3(−5).
- Simplifies to −24 + 40 − 15.
- Obtains det(A) = 1.
Examiner tip: Cofactor expansion along a row or column with the most zeros minimises the arithmetic — here the first row has no zeros, but a GDC's determinant function avoids this choice entirely.
- 12.
Marking analysis: A learner attempts the following task: “Find the determinant of the 3×3 matrix A = [[1, 2, 3], [0, 1, 4], [5, 6, 0]].” Their response addresses only this point: “Expands along the first row: det(A) = 1(1×0 − 4×6) − 2(0×0 − 4×5) + 3(0×6 − 1×5).” Evaluate the response against the complete 4-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[4 marks]Marking points
- Recognises credit for the stated point: Expands along the first row: det(A) = 1(1×0 − 4×6) − 2(0×0 − 4×5) + 3(0×6 − 1×5).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Evaluates each 2×2 minor: 1(−24) − 2(−20) + 3(−5).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Simplifies to −24 + 40 − 15.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains det(A) = 1.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 13.
Given A = [[1, 2, 3], [4, 5, 6]] and B = [[7, 8], [9, 10], [11, 12]], find the matrix product AB, stating its dimensions.
[4 marks]Worked explanation
- A is 2 by 3 and B is 3 by 2, so AB is 2 by 2.
- Row-column products give [[7+18+33,8+20+36],[28+45+66,32+50+72]]=[[58,64],[139,154]].
Common mistake: The inner dimensions must match; the outer dimensions determine the result.
Marking points
- Confirms the product is defined since A is 2×3 and B is 3×2, so AB will be 2×2.
- Calculates the first row of AB: (1×7+2×9+3×11, 1×8+2×10+3×12) = (58, 64).
- Calculates the second row of AB: (4×7+5×9+6×11, 4×8+5×10+6×12) = (139, 154).
- States AB = [[58, 64], [139, 154]].
Examiner tip: Two matrices can only be multiplied if the number of columns in the first equals the number of rows in the second — always check this before attempting the multiplication.
- 14.
Marking analysis: A learner attempts the following task: “Given A = [[1, 2, 3], [4, 5, 6]] and B = [[7, 8], [9, 10], [11, 12]], find the matrix product AB, stating its dimensions.” Their response addresses only this point: “Confirms the product is defined since A is 2×3 and B is 3×2, so AB will be 2×2.” Evaluate the response against the complete 4-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[4 marks]Marking points
- Recognises credit for the stated point: Confirms the product is defined since A is 2×3 and B is 3×2, so AB will be 2×2.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the first row of AB: (1×7+2×9+3×11, 1×8+2×10+3×12) = (58, 64).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the second row of AB: (4×7+5×9+6×11, 4×8+5×10+6×12) = (139, 154).
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: States AB = [[58, 64], [139, 154]].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 15.
A machine is always in one of two states: Working (W) or Broken (B). If it is Working on a given day, the probability it is Working the next day is 0.8. If it is Broken, the probability it is Working the next day is 0.3. The machine is Working today. Use the transition matrix P = [[0.8, 0.2], [0.3, 0.7]] and the state vector [1, 0] to find the probability distribution for tomorrow.
[3 marks]Worked explanation
- The question uses row states in order (W,B), so update by multiplying on the right: [1,0]P.
- Tomorrow's distribution is [0.8,0.2], selecting P's first row. Both entries are nonnegative and sum to 1.
Common mistake: Using a column-state convention without transposing P reverses the update.
Marking points
- Multiplies the row state vector by the transition matrix: [1, 0] × P.
- Obtains the new state vector [0.8, 0.2].
- Interprets this as an 80% chance the machine is Working tomorrow and a 20% chance it is Broken.
Examiner tip: The row of the transition matrix corresponding to the current state gives next period's probabilities directly — multiplying by a state vector with a single 1 simply selects that row.
- 16.
Marking analysis: A learner attempts the following task: “A machine is always in one of two states: Working (W) or Broken (B). If it is Working on a given day, the probability it is Working the next day is 0.8. If it is Broken, the probability it is Working the next day is 0.3. The machine is Working today. Use the transition matrix P = [[0.8, 0.2], [0.3, 0.7]] and the state vector [1, 0] to find the probability distribution for tomorrow.” Their response addresses only this point: “Multiplies the row state vector by the transition matrix: [1, 0] × P.” Evaluate the response against the complete 3-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[3 marks]Marking points
- Recognises credit for the stated point: Multiplies the row state vector by the transition matrix: [1, 0] × P.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains the new state vector [0.8, 0.2].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Interprets this as an 80% chance the machine is Working tomorrow and a 20% chance it is Broken.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 17.
A machine is always in one of two states: Working (W) or Broken (B). Its transition matrix is P = [[0.8, 0.2], [0.3, 0.7]], where the first row gives next-day probabilities from Working and the second row from Broken. The machine is Working today, so the starting state vector is [1, 0]. Find the probability distribution for the day after tomorrow (two days ahead).
[4 marks]Worked explanation
- After one day the row state is [0.8,0.2]. Multiply again by P.
- Two-day state=[0.8(0.8)+0.2(0.3),0.8(0.2)+0.2(0.7)]=[0.70,0.30]. The entries sum to 1.
Common mistake: Two steps require P^2, not twice P.
Marking points
- Finds tomorrow's state vector: [1, 0] × P = [0.8, 0.2].
- Multiplies this state vector by P again: [0.8, 0.2] × P.
- Calculates the first entry: 0.8(0.8) + 0.2(0.3) = 0.70.
- Obtains the state vector for two days ahead as [0.70, 0.30].
Examiner tip: Applying the transition matrix repeatedly from the current state is equivalent to multiplying the original state vector by P raised to that power — both approaches give the same result.
- 18.
Marking analysis: A learner attempts the following task: “A machine is always in one of two states: Working (W) or Broken (B). Its transition matrix is P = [[0.8, 0.2], [0.3, 0.7]], where the first row gives next-day probabilities from Working and the second row from Broken. The machine is Working today, so the starting state vector is [1, 0]. Find the probability distribution for the day after tomorrow (two days ahead).” Their response addresses only this point: “Finds tomorrow's state vector: [1, 0] × P = [0.8, 0.2].” Evaluate the response against the complete 4-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[4 marks]Marking points
- Recognises credit for the stated point: Finds tomorrow's state vector: [1, 0] × P = [0.8, 0.2].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Multiplies this state vector by P again: [0.8, 0.2] × P.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the first entry: 0.8(0.8) + 0.2(0.3) = 0.70.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains the state vector for two days ahead as [0.70, 0.30].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 19.
For the transition matrix P = [[0.8, 0.2], [0.3, 0.7]], find the steady-state (long-run) probability distribution π = [π₁, π₂], using πP = π and π₁ + π₂ = 1.
[5 marks]Worked explanation
- Let pi=[w,b]. From pi P=pi, 0.8w+0.3b=w, hence 0.2w=0.3b; also w+b=1.
- Thus pi=[0.6,0.4]. Check [0.6,0.4]P=[0.48+0.12,0.12+0.28]=[0.6,0.4].
Common mistake: The stationary equation alone needs probability normalisation.
Marking points
- Sets up the steady-state equations from πP = π: 0.8π₁ + 0.3π₂ = π₁ and 0.2π₁ + 0.7π₂ = π₂.
- Simplifies the first equation to −0.2π₁ + 0.3π₂ = 0, giving π₂ = (2/3)π₁.
- Substitutes into π₁ + π₂ = 1 to obtain π₁(1 + 2/3) = 1.
- Solves to obtain π₁ = 0.6.
- Obtains π₂ = 0.4, giving the steady-state distribution [0.6, 0.4].
Examiner tip: The steady-state vector is the same regardless of the starting state, for a regular Markov chain — it depends only on the transition matrix itself.
- 20.
Marking analysis: A learner attempts the following task: “For the transition matrix P = [[0.8, 0.2], [0.3, 0.7]], find the steady-state (long-run) probability distribution π = [π₁, π₂], using πP = π and π₁ + π₂ = 1.” Their response addresses only this point: “Sets up the steady-state equations from πP = π: 0.8π₁ + 0.3π₂ = π₁ and 0.2π₁ + 0.7π₂ = π₂.” Evaluate the response against the complete 5-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[5 marks]Marking points
- Recognises credit for the stated point: Sets up the steady-state equations from πP = π: 0.8π₁ + 0.3π₂ = π₁ and 0.2π₁ + 0.7π₂ = π₂.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Simplifies the first equation to −0.2π₁ + 0.3π₂ = 0, giving π₂ = (2/3)π₁.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Substitutes into π₁ + π₂ = 1 to obtain π₁(1 + 2/3) = 1.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Solves to obtain π₁ = 0.6.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains π₂ = 0.4, giving the steady-state distribution [0.6, 0.4].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 21.
For the row-stochastic transition matrix P = [[0.8, 0.2], [0.3, 0.7]], calculate P^2 and explain its entries. Check that each row of P^2 sums to 1, and use its first row to state the two-step distribution starting from row vector [1, 0].
[4 marks]Worked explanation
- Multiply P by itself: P^2=[[0.64+0.06,0.16+0.14],[0.24+0.21,0.06+0.49]]=[[0.70,0.30],[0.45,0.55]].
- Entry (i,j) is the probability of ending in state j after two days starting in i. Row sums are 1; [1,0]P^2=[0.70,0.30].
Common mistake: Squaring each entry separately omits paths through intermediate states.
Marking points
- Multiplies P by itself to obtain P² = [[0.70, 0.30], [0.45, 0.55]].
- Explains that each entry (i, j) of P² gives the probability of moving from state i to state j in exactly two steps.
- Notes that the rows of P² still sum to 1, since the matrix still represents a valid probability distribution after two steps.
- Cross-checks that the top row of P² matches the two-step state vector found starting from [1, 0].
Examiner tip: Raising a transition matrix to the nth power always gives the n-step transition probabilities directly — this is often faster than repeatedly multiplying a state vector when you need several different starting states.
- 22.
Marking analysis: A learner attempts the following task: “For the row-stochastic transition matrix P = [[0.8, 0.2], [0.3, 0.7]], calculate P^2 and explain its entries. Check that each row of P^2 sums to 1, and use its first row to state the two-step distribution starting from row vector [1, 0].” Their response addresses only this point: “Multiplies P by itself to obtain P² = [[0.70, 0.30], [0.45, 0.55]].” Evaluate the response against the complete 4-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[4 marks]Marking points
- Recognises credit for the stated point: Multiplies P by itself to obtain P² = [[0.70, 0.30], [0.45, 0.55]].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Explains that each entry (i, j) of P² gives the probability of moving from state i to state j in exactly two steps.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Notes that the rows of P² still sum to 1, since the matrix still represents a valid probability distribution after two steps.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Cross-checks that the top row of P² matches the two-step state vector found starting from [1, 0].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 23.
Solve the matrix equation AX = B for the column vector X, where A = [[2, 1], [1, 1]] and B = [[8], [5]], by finding A⁻¹ and using X = A⁻¹B.
[5 marks]Worked explanation
- det A=2(1)-1(1)=1, so A^-1=[[1,-1],[-1,2]].
- X=A^-1B=[8-5,-8+10]^T=[3,2]^T. Check AX=[6+2,3+2]^T=[8,5]^T.
Common mistake: Multiply by the inverse on the left for AX=B.
Marking points
- Calculates det(A) = (2)(1) − (1)(1) = 1.
- Finds A⁻¹ = [[1, −1], [−1, 2]].
- Multiplies X = A⁻¹B.
- Calculates the first entry: 1(8) + (−1)(5) = 3.
- Obtains X = [[3], [2]], so x = 3 and y = 2.
Examiner tip: Solving AX = B using the inverse matrix is mathematically identical to the elimination method for simultaneous equations — it is simply a more systematic, scalable way to organise the same process.
- 24.
Marking analysis: A learner attempts the following task: “Solve the matrix equation AX = B for the column vector X, where A = [[2, 1], [1, 1]] and B = [[8], [5]], by finding A⁻¹ and using X = A⁻¹B.” Their response addresses only this point: “Calculates det(A) = (2)(1) − (1)(1) = 1.” Evaluate the response against the complete 5-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[5 marks]Marking points
- Recognises credit for the stated point: Calculates det(A) = (2)(1) − (1)(1) = 1.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Finds A⁻¹ = [[1, −1], [−1, 2]].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Multiplies X = A⁻¹B.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the first entry: 1(8) + (−1)(5) = 3.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains X = [[3], [2]], so x = 3 and y = 2.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 25.
A three-state weather model has transition matrix P = [[0.7, 0.2, 0.1], [0.1, 0.6, 0.3], [0.2, 0.2, 0.6]] for states (Sunny, Cloudy, Rainy) respectively. Today's distribution is [0.5, 0.3, 0.2]. Find tomorrow's probability distribution.
[5 marks]Worked explanation
- Use row state [0.5,0.3,0.2]P. Sunny probability=0.5(0.7)+0.3(0.1)+0.2(0.2)=0.42.
- Cloudy=0.10+0.18+0.04=0.32; Rainy=0.05+0.09+0.12=0.26. Tomorrow's distribution [0.42,0.32,0.26] sums to 1.
Common mistake: Combine contributions from every possible starting weather state.
Marking points
- Multiplies the row state vector [0.5, 0.3, 0.2] by the transition matrix P.
- Calculates the first entry: 0.5(0.7) + 0.3(0.1) + 0.2(0.2) = 0.42.
- Calculates the second entry: 0.5(0.2) + 0.3(0.6) + 0.2(0.2) = 0.32.
- Calculates the third entry: 0.5(0.1) + 0.3(0.3) + 0.2(0.6) = 0.26.
- States tomorrow's distribution as [0.42, 0.32, 0.26].
Examiner tip: With more than two states, a GDC's matrix multiplication function avoids tedious hand calculation — enter the state vector and transition matrix directly rather than computing each entry manually.
- 26.
Marking analysis: A learner attempts the following task: “A three-state weather model has transition matrix P = [[0.7, 0.2, 0.1], [0.1, 0.6, 0.3], [0.2, 0.2, 0.6]] for states (Sunny, Cloudy, Rainy) respectively. Today's distribution is [0.5, 0.3, 0.2]. Find tomorrow's probability distribution.” Their response addresses only this point: “Multiplies the row state vector [0.5, 0.3, 0.2] by the transition matrix P.” Evaluate the response against the complete 5-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[5 marks]Marking points
- Recognises credit for the stated point: Multiplies the row state vector [0.5, 0.3, 0.2] by the transition matrix P.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the first entry: 0.5(0.7) + 0.3(0.1) + 0.2(0.2) = 0.42.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the second entry: 0.5(0.2) + 0.3(0.6) + 0.2(0.2) = 0.32.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the third entry: 0.5(0.1) + 0.3(0.3) + 0.2(0.6) = 0.26.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: States tomorrow's distribution as [0.42, 0.32, 0.26].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 27.
A city's residents are classified as living Downtown (D) or in the Suburbs (S). Each year, 40% of Downtown residents move to the Suburbs, and 50% of Suburb residents move Downtown. The transition matrix is P = [[0.6, 0.4], [0.5, 0.5]]. If the population starts entirely Downtown, [1, 0], find the distribution after 2 years.
[4 marks]Worked explanation
- After one year [1,0]P=[0.6,0.4]. Apply the same annual transition again.
- After two years: [0.6(0.6)+0.4(0.5),0.6(0.4)+0.4(0.5)]=[0.56,0.44], summing to 1.
Common mistake: Some suburban residents return downtown in year two; do not simply square 0.6.
Marking points
- Multiplies [1, 0] by P to obtain the distribution after 1 year: [0.6, 0.4].
- Multiplies [0.6, 0.4] by P to obtain the distribution after 2 years.
- Calculates the first entry: 0.6(0.6) + 0.4(0.5) = 0.56.
- Obtains the distribution after 2 years as [0.56, 0.44].
Examiner tip: Each year's distribution depends only on the previous year's distribution, not on how the population reached that state — this 'memoryless' property is the defining feature of a Markov chain.
- 28.
Marking analysis: A learner attempts the following task: “A city's residents are classified as living Downtown (D) or in the Suburbs (S). Each year, 40% of Downtown residents move to the Suburbs, and 50% of Suburb residents move Downtown. The transition matrix is P = [[0.6, 0.4], [0.5, 0.5]]. If the population starts entirely Downtown, [1, 0], find the distribution after 2 years.” Their response addresses only this point: “Multiplies [1, 0] by P to obtain the distribution after 1 year: [0.6, 0.4].” Evaluate the response against the complete 4-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[4 marks]Marking points
- Recognises credit for the stated point: Multiplies [1, 0] by P to obtain the distribution after 1 year: [0.6, 0.4].
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Multiplies [0.6, 0.4] by P to obtain the distribution after 2 years.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the first entry: 0.6(0.6) + 0.4(0.5) = 0.56.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Obtains the distribution after 2 years as [0.56, 0.44].
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 29.
Show that the uniform distribution π = [1/3, 1/3, 1/3] is the steady-state distribution for the transition matrix P = [[0.7, 0.2, 0.1], [0.1, 0.6, 0.3], [0.2, 0.2, 0.6]], by verifying that πP = π.
[4 marks]Worked explanation
- Multiplying uniform pi by P gives one third of each column sum: (0.7+0.1+0.2)/3, (0.2+0.6+0.2)/3, (0.1+0.3+0.6)/3.
- Each sum is 1, hence pi P=[1/3,1/3,1/3]=pi. Here P is doubly stochastic, not merely row-stochastic.
Common mistake: Row sums of 1 alone do not guarantee a uniform stationary distribution.
Marking points
- Multiplies π = [1/3, 1/3, 1/3] by P.
- Calculates the first entry: (1/3)(0.7 + 0.1 + 0.2) = (1/3)(1) = 1/3.
- Notes that each column of P sums to 1 (the matrix is doubly stochastic), so every entry of πP equals (1/3)(1) = 1/3 by the same reasoning.
- Concludes that πP = [1/3, 1/3, 1/3] = π, confirming π is the steady-state distribution.
Examiner tip: Whenever every column of a transition matrix sums to 1 (a doubly stochastic matrix), the uniform distribution is always its steady state — this is a useful shortcut that avoids solving simultaneous equations.
- 30.
Marking analysis: A learner attempts the following task: “Show that the uniform distribution π = [1/3, 1/3, 1/3] is the steady-state distribution for the transition matrix P = [[0.7, 0.2, 0.1], [0.1, 0.6, 0.3], [0.2, 0.2, 0.6]], by verifying that πP = π.” Their response addresses only this point: “Multiplies π = [1/3, 1/3, 1/3] by P.” Evaluate the response against the complete 4-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[4 marks]Marking points
- Recognises credit for the stated point: Multiplies π = [1/3, 1/3, 1/3] by P.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Calculates the first entry: (1/3)(0.7 + 0.1 + 0.2) = (1/3)(1) = 1/3.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Notes that each column of P sums to 1 (the matrix is doubly stochastic), so every entry of πP equals (1/3)(1) = 1/3 by the same reasoning.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Concludes that πP = [1/3, 1/3, 1/3] = π, confirming π is the steady-state distribution.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.
- 31.
Explain what it means for a Markov chain to be 'regular', and explain why a regular Markov chain always converges to a unique steady-state distribution regardless of its starting state.
[3 marks] · no calculatorWorked explanation
- A finite Markov chain is regular if some positive integer power P^k has every entry strictly positive; every state can then reach every other in the same number of steps.
- For such a stochastic matrix, eigenvalue 1 has one normalised stationary probability vector and all other eigenvalues have modulus below 1. Repeated updates damp those other components, so every initial probability state converges to the same stationary distribution.
Common mistake: Regular does not require every entry of P itself to be positive.
Marking points
- For a finite Markov chain, regular means that some common positive integer power P^k has all entries strictly positive: every state can reach every other in exactly k steps. Mere reachability in some number of steps depending on the pair is not sufficient.
- Explains that this property guarantees the chain 'mixes' thoroughly over time, washing out the influence of the starting state.
- Concludes that as a result, the probability distribution converges to the same unique steady-state vector no matter which state the chain began in.
Examiner tip: Use the common-power positivity condition to distinguish a regular chain from a periodic irreducible chain. Then explain that repeated transitions converge to a unique stationary distribution independent of the initial distribution.
- 32.
Marking analysis: A learner attempts the following task: “Explain what it means for a Markov chain to be 'regular', and explain why a regular Markov chain always converges to a unique steady-state distribution regardless of its starting state.” Their response addresses only this point: “For a finite Markov chain, regular means that some common positive integer power P^k has all entries strictly positive: every state can reach every other in exactly k steps. Mere reachability in some number of steps depending on the pair is not sufficient.” Evaluate the response against the complete 3-mark task. Identify what earns credit and state every additional requirement needed for full marks.
[3 marks] · no calculatorMarking points
- Recognises credit for the stated point: For a finite Markov chain, regular means that some common positive integer power P^k has all entries strictly positive: every state can reach every other in exactly k steps. Mere reachability in some number of steps depending on the pair is not sufficient.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Explains that this property guarantees the chain 'mixes' thoroughly over time, washing out the influence of the starting state.
- Checks this criterion, crediting equivalent work already shown or identifying what is missing: Concludes that as a result, the probability distribution converges to the same unique steady-state vector no matter which state the chain began in.
Examiner tip: Check each criterion, including equivalent work already shown. Do not demand an extra step merely to rewrite an equivalent answer. Do not award the same idea twice or infer work the learner did not show.