GATE Instrumentation (IN): Engineering Mathematics (linear algebra, calculus, differential equations, complex variables, probability, numerical methods), previous-year questions solved
Every engineering mathematics question from the last five GATE Instrumentation Engineering papers, solved: matrix rank and eigenvalues, limits, maxima and integrals, differential equations, residues and contour integrals, probability, and Newton-Raphson and Euler steps. Each answer matches the official key and was recomputed by code with the numbers printed in the paper.
Questions from official GATE IN papers (IIT Kharagpur 2022, IIT Kanpur 2023, IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 19 Nov 2026 · 10 of 44 questions on this page
Which statements about the rank, rows and determinant of the 3 by 3 matrix M are correct?
Given , which of the following statement(s) is/are correct?
- (A)The rank of is 2
- (B)The rank of is 3
- (C)The rows of are linearly independent
- (D)The determinant of is 0
Answer: (A) and (D) Official key: A, D
- The rank is the number of linearly independent rows, and a square matrix is singular (determinant ) exactly when its rows are dependent. So all four statements hang on one question: are the three rows independent?
- Look at the first and third rows: . The rows are dependent, so (C) is false and , which makes (D) true.
- The first two rows are not multiples of each other (the ratio but ), so and are independent and the rank is exactly . (A) is true and (B) is false.
- Check by expansion along the first row: .
Trap: believing a non-zero-looking matrix has rank 3. We confirmed with exact fractions (Gaussian elimination gives rank , determinant ) and by solving with sympy, which returns .
Rank tells an engineer whether a set of measurements is really independent. When a calibration or sensor-fusion problem produces a matrix with a dependent row, two of the readings carry the same information and the system cannot be solved uniquely until one is dropped or a new sensor is added.
Which vectors are eigenvectors of A = [[4, 3], [9, -2]] with eigenvalues -5 and 7?
The matrix has eigenvalues and . The eigenvector(s) is/are ________
- (A)
- (B)
- (C)
- (D)
Answer: (A) and (C) Official key: A, C
- An eigenvector of is a direction that the matrix only stretches: . The eigenvalues are given, so test each option directly, or solve for each .
- For : , so and . The direction is , option (A): .
- For : , so and . The direction is , and is twice that, option (C): .
- Option (B): , not a multiple of . Option (D): , not a multiple of . Both fail. The answer is (A) and (C).
Trap: expecting one eigenvector only, or accepting a scaled vector as wrong: any non-zero multiple of an eigenvector is also an eigenvector. We tested every option with exact integer arithmetic and cross-checked the directions with numpy (eigenvalue 7 gives , eigenvalue gives ).
Eigenvalues and eigenvectors of a system matrix give the natural rates and shapes of a response: each eigenvector is a pattern that decays or grows at its own rate. Engineers read stability of a control loop or a state estimator from the sign of the eigenvalues, and the eigenvectors show which combination of states excites each mode.
Solution set of the two equations x - 2y + z = 0 and x - z = 0 in three unknowns
- (A)
- (B)
- (C)
- (D)
Answer: (A) Official key: A
- Two equations in three unknowns form a homogeneous system with . Its solutions form a subspace (the null space) of dimension , by the rank-nullity theorem.
- The two rows are not multiples of each other, so and the null space has dimension : a single line through the origin. That already rules out (B) and (C), which describe planes (two independent directions).
- Solve directly. The second equation gives . The first becomes , so . Every solution is .
- The solution set is , which is (A).
- Check (D): gives in the first equation, so it is not a solution.
Trap: picking the option with two free parameters because there are two equations in three unknowns. The number of free parameters is the nullity, three minus the rank, and the rank here is 2. We confirmed with an exact sympy null-space calculation (rank 2, nullity 1, basis [1, 1, 1]) and by substituting the generators of every option into both equations.
Null spaces tell an instrument engineer which combinations of inputs are invisible to a measurement. For example, if a set of redundant sensors is described by a matrix that has a non-trivial null space, there are directions in which the readings cannot distinguish one state from another.
Rank of the 3 x 2 matrix with entries 1, 4, -3, alpha, beta, 6 is one: find alpha/beta
Answer: Official key: -8 to -8
- A matrix has rank one when every row is a multiple of one single row, equivalently every minor is zero.
- The first row is . Row 2 must be times it: , so .
- Row 3 is and must equal , so and .
- The ratio is .
Trap: setting only one minor to zero and missing the second condition. Both unknowns need their own row to be a multiple of the first row, which has no unknowns in it. We solved all three minors with sympy: , , and the completed matrix has exact rank 1.
A rank-one matrix means all the channels carry the same information, scaled. Checking the rank of a measurement matrix is a standard way to find redundant or perfectly correlated sensors, and rank deficiency is also what makes a calibration problem impossible to solve uniquely.
Which set of three 3-element column vectors v1, v2, v3 gives a matrix M with rank 3?
- (A)
- (B)
- (C)
- (D)
Answer: (C) Official key: C
- The rank of a matrix is 3 exactly when its three columns are linearly independent, that is, when no column can be built from the others and the determinant is not zero. So the job is to find the set in which every column adds a new direction.
- (A): . The third column is the sum of the first two, so the rank is 2.
- (B): is the zero vector. A zero column never adds a direction, so the rank is at most 2.
- (D): . Again the columns are dependent, so the rank is 2.
- (C): expanding the determinant of along the second row gives , so the columns are independent and the rank is 3. Option (C).
Trap: checking only that no column is a multiple of another. In (A) and (D) no two columns are parallel, yet three of them still lie in one plane. We confirmed with an exact rank computation in sympy: ranks 2, 2, 3, 2 for (A) to (D), and for (C).
Rank is the test for whether a set of sensors or measurements carries independent information. If the rows of a calibration or observation matrix are dependent, the quantities being estimated cannot all be recovered, which is why redundant thermocouple placements need checking before a multi-point fit.
Determinant of P inverse for a real 3x3 matrix P with eigenvalues 1/4, 1 and -2
Answer: Official key: -2 to -2
- The determinant of a square matrix equals the product of its eigenvalues. Here .
- The determinant of an inverse is the reciprocal of the determinant, . So .
- A cross-check from the other side: the eigenvalues of are the reciprocals , , , and their product is .
Trap: stopping at , or forgetting the sign of the negative eigenvalue and answering . We tested it on 25 random real matrices with integer of determinant , using exact fractions: every one had these eigenvalues and exactly (and ).
Eigenvalues of a system matrix give the modes of a plant, and the sign and size of their product, the determinant, tell whether the matrix can be inverted at all. A near-zero determinant warns that a calibration or decoupling matrix will amplify sensor noise when inverted.
Rank of a 2n x 2n matrix whose entries are beta or -beta according to the parity of i + j
- (A)1
- (B)2
- (C)
- (D)
Answer: (A) Official key: A
- The rank of a matrix is the number of linearly independent rows. So the job is to compare the rows, not to compute a determinant.
- In an odd-numbered row , the entry in column has odd when is even and even when is odd. So that row reads .
- In an even-numbered row, the parities flip, so it reads . That is exactly times the odd row.
- Every row is therefore the same non-zero row, so there is only one independent row. The rank is . Option (A), whatever the value of and .
Trap: reading the size as a hint that the rank is (option D), or because there are rows of each kind. The two kinds of row are not independent. We built the matrix with sympy for to and two values of (1 and 7) and computed the rank exactly each time: it was 1 for every case.
Rank tells an engineer how many independent pieces of information a set of measurements really carries. When a sensor array or a set of calibration equations has rank 1, extra channels add no new information and the system cannot be solved for more than one unknown.
Value of x in the matrix [-1 -1; x -4] if an eigenvector is along (alpha, 2 alpha)
Answer: Official key: 2 to 2
- An eigenvector of satisfies : the matrix only stretches it, it does not turn it. The scale drops out, so use .
- First row: , and this must equal . So .
- Second row: , and this must equal . So and .
- Check: with the characteristic equation has roots and , and is the eigenvalue of .
Trap: finding the eigenvalues of the general matrix first. That is far longer, because is still unknown. Using the given eigenvector to read off from the row with no is the shortcut. We tested every integer from to with exact integer arithmetic: only makes an eigenvector, and sympy gives the eigenvalue for it.
Eigenvectors of a system matrix are its natural modes, the directions in which a state moves without changing shape. Knowing one mode and its direction is a quick way to back out an unknown coefficient when identifying a small two-state model from a measured response.
3x3 matrix P with P^3 - 4P^2 + 5P - 2I = 0: the set of all possible eigenvalues of P
The set of all possible eigenvalues of matrix is ___.
- (A)
- (B)
- (C)
- (D)
Answer: (C) Official key: C
- Principle: if a matrix satisfies a polynomial equation , every eigenvalue of is a root of . Take an eigenvector with . Then , and since , .
- Factorise . Try : , so is a factor. Dividing gives .
- So and the only values an eigenvalue of can take are and . The value is not a root, so it can never appear.
- Both roots really do occur for some : has eigenvalue only, has eigenvalue only, and has both. The set of all possible eigenvalues is therefore . Option (C).
Trap: picking {1} or {2} because one specific matrix has only that eigenvalue. The question asks for every value that could appear for some admissible , so both roots count. We checked this in code: every matrix built from diagonal or Jordan forms with entries from (conjugated by an integer matrix of determinant 1) was tested exactly against the equation, and the eigenvalues of the matrices that passed formed exactly the set {1, 2}; and both passed, and did not.
State-space models of sensors and plants are analysed through the eigenvalues of the system matrix, and a matrix that obeys a polynomial equation has eigenvalues locked to that polynomial's roots. The same idea sits behind the Cayley-Hamilton theorem used to compute matrix functions in controller and filter design.
Number of solutions of Ax = b for a given 3x3 matrix A and vector b
Let and
The number of solutions to the linear system of equations is _____.
- (A)0
- (B)1
- (C)6
- (D)infinitely many
Answer: (A) Official key: A
- Principle: is consistent only if . If the ranks differ there is no solution. If they are equal and equal to the number of unknowns, the solution is unique, and if they are equal and smaller, there are infinitely many.
- Determinant of : . So is singular and ; the rows are dependent.
- Find the dependence. The equations are , and . Adding the first two gives , but the third equation says . Row 3 equals row 1 plus row 2 on the left, yet on the right.
- Row reduction gives and . The ranks differ, so the system is inconsistent: 0 solutions. Option (A).
Trap: seeing a determinant of and answering 'infinitely many'. A zero determinant only rules out a unique solution. Whether there are none or infinitely many depends on whether lies in the column space, which is what the rank test checks. We confirmed this with exact fraction row reduction of and of the augmented matrix.
Calibration of a multi-sensor array solves a linear system for the unknown gains and offsets, and a singular or inconsistent system is a warning that two sensors measure the same thing or that a reading is wrong. Checking ranks tells you whether a sensor fusion problem has a solution at all.