GATE Mechanical: Linear Algebra (Engineering Mathematics), previous-year questions solved

Linear algebra is the most predictable part of the 13 Engineering Mathematics marks in GATE Mechanical: up to two questions a paper on rank, consistency of equations, eigenvalues or matrix properties. Here are all 7 with an official key from GATE ME 2022 (both sets), 2024 and 2025 (2026 had none), solved with the exam shortcut and checked against the key.

Questions from official GATE ME papers (IIT Kharagpur 2022, IISc Bengaluru 2024, IIT Roorkee 2025); solutions by Skill-Lync. · Updated 16 Nov 2026 · 7 questions

GATE 2022 · ME Set-1 · Q.15MCQ · 1 markMatrix algebra and symmetric matrices

2 × 2 matrix with entries in k is symmetric: what is k?

If is a symmetric matrix, the value of is ___________.
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A) Official key: A

  1. A matrix is symmetric when , that is . The diagonal entries can be anything; only the off-diagonal pair has to match.
  2. , so : option (A). Check: .

Every other option comes from a different (wrong) condition: setting the determinant to zero, , gives and , option (D); making the off-diagonal entries opposite (skew-symmetric style), , gives , option (C); making the diagonal entries equal, , gives , option (B).

Where this shows up in industry

The stiffness matrix of a linear elastic FEA model is symmetric (Maxwell-Betti reciprocity: the deflection at point i from a load at j equals the deflection at j from the same load at i). Solvers store only half of it and use symmetric methods such as Cholesky factorisation; frictional contact and follower loads make it unsymmetric, and ANSYS and Abaqus then need an unsymmetric solver option, which costs more memory and time.

GATE 2025 · ME · Q.11MCQ · 1 markMatrix algebra and symmetric matrices

A and B are real symmetric matrices of the same size: which statement is always correct?

Let A and B be real symmetric matrices of same size. Which one of the following options is correct?
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (C) Official key: C

  1. holds for any two matrices that can be multiplied, symmetric or not: entry of is , which is entry of . So option (C) is always correct.
  2. The others need extra conditions. defines an orthogonal matrix; for a symmetric it would mean , which also makes (D) true, and that is not true in general. Take : .
  3. fails too. With the same and : but .

For symmetric matrices the rule gives , so the product of two symmetric matrices is itself symmetric only when they commute. We checked all four statements on 297 random pairs of symmetric integer matrices: only (C) held every time.

Where this shows up in industry

The transpose rule is how FEA codes build element stiffness matrices: , where maps nodal displacements to strains and is the (symmetric) material matrix. Because , the element matrix comes out symmetric automatically, which is why the assembled stiffness matrix is symmetric too.

GATE 2022 · ME Set-1 · Q.46MSQ · 2 marksSystems of linear equations and rank

Homogeneous 2 × 2 system with parameter α: which (x, y) are non-trivial solutions?

The system of linear equations in real given by involves a real parameter and has infinitely many non-trivial solutions for special value(s) of . Which one or more among the following options is/are non-trivial solution(s) of for such special value(s) of ?
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A, B) Official key: A,B

  1. The row vector multiplies the matrix from the left, so the two equations are and .
  2. A homogeneous system has non-trivial solutions only when its determinant is zero: , so or .
  3. : both equations become , so every solution has . fits: (A).
  4. : and , both . fits: (B).
  5. needs and needs ; neither ratio occurs. Answer (A, B).

Trap: reading the system as matrix times column vector gives and . The determinant is the same, so the special values of do not change, but the solution directions do: for and for . You would then pick (D) alone.

Where this shows up in industry

This is the shape of every buckling and natural-frequency problem: has a non-zero mode shape only at the special values of that make the determinant zero. Just as here, the mode shape is known only as a direction (any multiple of it is also a solution), which is why FEA tools report normalised mode shapes, not absolute displacements.

GATE 2022 · ME Set-2 · Q.46MSQ · 2 marksSystems of linear equations and rank

A is a 3 × 5 real matrix of rank 2: what can you say about the solutions of Ax = 0?

is a real matrix of rank 2. For the set of homogeneous equations , where is a zero vector and is a vector of unknown variables, which of the following is/are true?
  1. (A)The given set of equations will have a unique solution.
  2. (B)The given set of equations will be satisfied by a zero vector of appropriate size.
  3. (C)The given set of equations will have infinitely many solutions.
  4. (D)The given set of equations will have many but a finite number of solutions.

Answer: (B, C) Official key: B,C

  1. has 5 unknowns (one per column). for any matrix, so the zero vector of size 5 always solves a homogeneous system: (B) is true.
  2. Rank-nullity theorem: , so . Three unknowns are free, and every choice of them gives a solution: (C) is true, and (A) is false.
  3. (D) can never happen over the real numbers: if is a non-zero solution, so is for every real . A linear system has no solution, exactly one, or infinitely many. Answer (B, C).

We built 200 random integer matrices of rank 2 and computed their null spaces exactly: each had nullity 3.

Where this shows up in industry

When an FEA model is not held down enough (a missing support, or a part connected to nothing), the stiffness matrix has a null space: rigid-body motions that need no force. The solver reports a singular matrix, a zero pivot or 'rigid body mode' warnings. A free-free modal analysis shows the same thing on purpose: six zero-frequency modes, one per rigid-body motion in 3D.

GATE 2024 · ME · Q.14MCQ · 1 markSystems of linear equations and rank

Three equations with unknown a and b: which values give infinitely many solutions?

Consider the system of linear equations The values of and such that there exists a non-trivial null space and the system admits infinite solutions are
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (D) Official key: D

  1. A non-trivial null space means the coefficient matrix is singular. Expanding along the first row, , which simplifies to . So .
  2. With , eliminate: gives , and gives .
  3. Both must hold together (otherwise the system has no solution): , so and . Then with free: infinitely many solutions. Option (D).

Trap: (option B) also makes the matrix singular, but then contradicts : the rank of the coefficient matrix is 2 while the augmented matrix has rank 3, so there is no solution at all. With (options A, C) the determinant is 16, not zero, and the solution is unique.

Where this shows up in industry

CAD sketchers run this rank test on their constraint equations: a sketch is fully defined when the constraint Jacobian has full rank, under-defined when it has free directions (a null space), and over-defined when a new dimension contradicts the others, which is the no-solution case here. SolidWorks and NX colour sketch geometry by exactly these three states.

GATE 2022 · ME Set-2 · Q.48NAT · 2 marksEigenvalues

Eigenvalues of [[3, p], [p, q]] add to 4 and multiply to −1: what is |p|?

If the sum and product of eigenvalues of a real matrix are 4 and respectively, then is ______ (in integer).

Answer: 2 Official key: 2 to 2

  1. For any square matrix, the sum of the eigenvalues equals the trace and the product equals the determinant. No need to find the eigenvalues themselves.
  2. Trace: , so .
  3. Determinant: , so , and .

Check: has eigenvalues , that is and . They add to 4 and multiply to .

Where this shows up in industry

The 2D stress tensor is a symmetric matrix of exactly this form. Its eigenvalues are the principal stresses, its trace and determinant are the stress invariants, and the centre and radius of Mohr's circle come straight from them. FEA post-processors compute principal stresses this way at every node.

GATE 2024 · ME · Q.36MCQ · 2 marksEigenvalues

Matrix [[1, a], [8, 3]] with a > 0: above what value of a does it have a negative eigenvalue?

The matrix (where ) has a negative eigenvalue if is greater than
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A) Official key: A

  1. The eigenvalues add to the trace, , and multiply to the determinant, .
  2. Their sum is positive, so they can't both be negative. Exactly one is negative when their product is negative: , that is : option (A).
  3. Directly: gives . The smaller root is negative when , again .

Check: at the eigenvalues are and ; at they are and , both positive. The roots are always real here because for .

Where this shows up in industry

A stiffness matrix with a negative eigenvalue means the structure has a direction in which it gives way instead of resisting: buckling, snap-through or loss of contact. Nonlinear solvers in Abaqus and ANSYS print 'negative eigenvalue' warnings for exactly this, and the sign of the eigenvalues of a Hessian tells an optimiser whether it has found a minimum or a saddle point.

From GATE prep to a CAE and CFD career

Talk to a counsellor.

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.

Talk to a career counsellor

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.