Home / Courses / GATE Physics / Practice / Numerical methods: roots, integration and ODEs

Numerical methods: roots, integration and ODEs

Mathematical methods · 79 questions and templates for GATE Physics, in order from warm-up to top rank. Check each answer, then read the full solution.

Loading questions… (they need JavaScript; nothing you type is sent anywhere)

Written questions with solutions

The same questions as a list, so you can print them or read without the practice box. Templates (fresh numbers) appear only in the practice box above.

  1. MCQWarm-up

    The Newton-Raphson iteration for solving f(x)=0 is
    1. xn+1=xn−f(xn)/f′(xn)
    2. xn+1=xn−f′(xn)/f(xn)
    3. xn+1=xn+f(xn)/f′(xn)
    4. xn+1=xn−f(xn)/f′′(xn)
    Answer and solution

    Answer: A

    Replace f by its tangent at xn: f(x)≈f(xn)+f′(xn)(x−xn). Setting this to zero gives xn+1=xn−f(xn)/f′(xn), where the tangent crosses the axis.
  2. MCQWarm-up

    The composite Simpson's 1/3 rule needs the number of sub-intervals to be
    1. odd
    2. even
    3. a multiple of 3
    4. any whole number
    Answer and solution

    Answer: B

    Simpson's 1/3 rule fits a parabola through each pair of adjacent sub-intervals (three points), so the sub-intervals must come in pairs: an even number, giving the weights 1,4,2,4,…,4,1. A multiple of 3 is needed by Simpson's 3/8 rule.
  3. NATWarm-up

    The bisection method starts with an interval of length 1 that brackets a root. The smallest number of bisections that makes the interval shorter than 10−3 is ______.
    Answer and solution

    Answer: 10

    Each bisection halves the interval: after n steps its length is 2−n. We need 2−n<10−3, i.e. n>3log2⁡10=9.97, so n=10 (2−10=1/1024).
  4. MCQWarm-up

    The trapezoidal rule gives the exact integral for all polynomials of degree up to
    1. 3
    2. 2
    3. 1
    4. 0
    Answer and solution

    Answer: C

    It replaces the function by straight lines between the points, so it is exact for straight lines (degree 1). For x2 it already errs: the error is −(b−a)h212f′′(ξ), which vanishes only when f′′=0.
  5. MCQWarm-up

    For dy/dx=f(x,y) with step h, the (explicit) Euler method is
    1. yn+1=yn+hf(xn+1,yn+1)
    2. yn+1=yn+h2[f(xn,yn)+f(xn+1,yn+1)]
    3. yn+1=yn−hf(xn,yn)
    4. yn+1=yn+hf(xn,yn)
    Answer and solution

    Answer: D

    Explicit Euler follows the slope at the start of the step: yn+1=yn+hf(xn,yn), the first two terms of the Taylor series. Using the slope at the end is implicit (backward) Euler; averaging the two slopes is the trapezoidal (Crank-Nicolson) method.
  6. NATWarm-up

    For dy/dx=y with y(0)=1, one Euler step of size h=0.1 gives y(0.1)≈ ______.
    Answer and solution

    Answer: 1.1 (±0.001)

    y1=y0+hf(x0,y0)=1+0.1×1=1.1. The exact value is e0.1=1.10517; Euler always undershoots this growing curve, because the slope keeps increasing during the step.
  7. MCQWarm-up

    Which root-finding method needs the derivative f′(x) at every step?
    1. Bisection
    2. Secant
    3. Newton-Raphson
    4. Regula falsi (false position)
    Answer and solution

    Answer: C

    Newton-Raphson uses xn+1=xn−f(xn)/f′(xn). The secant method replaces f′ by the slope through the last two points; bisection and false position use only values of f at the ends of a bracket.

    Why people get this wrong: The secant method looks like Newton's but needs no derivative.

  8. NATStandard

    Newton-Raphson is used on f(x)=x2−2 starting from x0=1. The second iterate x2 is ______.
    Answer and solution

    Answer: 1.4167 (±0.0005)

    The iteration is xn+1=xn−xn2−22xn=12(xn+2xn) (the Babylonian square root). x1=12(1+2)=1.5, x2=12(1.5+1.3333)=1.41667. 2=1.41421: already 3 correct figures.
  9. NATStandard

    The composite trapezoidal rule with step h=0.5 estimates ∫01x2dx as ______.
    Answer and solution

    Answer: 0.375 (±0.001)

    T=h2[f(0)+2f(0.5)+f(1)]=0.25(0+0.5+1)=0.375. The exact value is 1/3; the error 0.0417 matches (b−a)h212f′′=1×0.25×212. The chords lie above a convex curve, so the rule overestimates.
  10. NATStandard

    Simpson's 1/3 rule with h=1 (two sub-intervals) estimates ∫02x4dx as ______.
    Answer and solution

    Answer: 6.667 (±0.005)

    S=h3[f(0)+4f(1)+f(2)]=13(0+4+16)=20/3≈6.667. The exact value is 32/5=6.4. Simpson is exact up to cubics; a quartic shows its error, (b−a)h4180f(4)=2×24180=0.267.
  11. MCQStandard

    Near a simple root, the order of convergence of the Newton-Raphson method is
    1. 1
    2. 1.618
    3. 2
    4. 3
    Answer and solution

    Answer: C

    Expanding about the root r with error en=xn−r: en+1≈f′′(r)2f′(r)en2. The error is squared each step (quadratic convergence): the number of correct digits roughly doubles. 1.618 is the secant method; 1 is bisection.
  12. MCQStandard

    The order of convergence of the secant method (near a simple root) is
    1. 1
    2. (1+5)/2≈1.618
    3. 2
    4. 1/2
    Answer and solution

    Answer: B

    The secant error obeys en+1≈Cenen−1. Trying en+1∝enp gives p2=p+1, so p=(1+5)/2, the golden ratio. It is slower per step than Newton but needs only one new function value per step and no derivative.
  13. MCQStandard

    When an ODE is integrated over a fixed interval with the explicit Euler method, the global error scales with the step size h as
    1. O(h)
    2. O(h2)
    3. O(h3)
    4. O(h4)
    Answer and solution

    Answer: A

    Each step makes a local error O(h2) (the neglected Taylor term). Over a fixed interval there are N∝1/h steps, so the errors accumulate to O(h): Euler is a first-order method.

    Why people get this wrong: O(h2) is the local (per step) error, not the global one.

  14. MCQStandard

    The global error of the classical fourth-order Runge-Kutta method scales as
    1. O(h)
    2. O(h2)
    3. O(h5)
    4. O(h4)
    Answer and solution

    Answer: D

    RK4 matches the Taylor series up to h4, so its local error is O(h5) and the global error, after ∝1/h steps, is O(h4). Halving h cuts the error about 16 times, at the cost of 4 function evaluations per step.
  15. MCQStandard

    The iteration xn+1=g(xn) converges to a fixed point x∗=g(x∗), for starting values close enough, if
    1. g(x∗)=0
    2. |g′(x∗)|>1
    3. g′′(x∗)=0
    4. |g′(x∗)|<1
    Answer and solution

    Answer: D

    With en=xn−x∗: en+1=g(x∗+en)−x∗≈g′(x∗)en. The error shrinks by the factor |g′(x∗)| each step, so we need it below 1. If g′(x∗)=0 (as for Newton's method, viewed as a fixed-point iteration) the convergence is faster than linear.
  16. MSQStandard

    Which of these rules integrate every cubic polynomial exactly (on a single panel)?
    1. Simpson's 1/3 rule
    2. Simpson's 3/8 rule
    3. The trapezoidal rule
    4. Two-point Gauss-Legendre quadrature
    Answer and solution

    Answer: A, B, D

    Simpson's 1/3 rule is built for quadratics but by symmetry also integrates x3 exactly. The 3/8 rule fits a cubic. An n-point Gauss rule is exact up to degree 2n−1=3 for n=2. The trapezoidal rule is exact only for straight lines.
  17. NATStandard

    The central difference f(x+h)−f(x−h)2h with h=0.1 estimates the derivative of f(x)=x3 at x=1 as ______.
    Answer and solution

    Answer: 3.01 (±0.001)

    1.331−0.7290.2=0.6020.2=3.01. The exact value is 3, and the error h2f′′′/6=0.01×6/6=0.01 matches: the central difference is second-order accurate.
  18. MCQStandard

    Which root-finding method must start from an interval on which the function changes sign, and then can never fail to converge?
    1. Bisection
    2. Newton-Raphson
    3. Secant
    4. Fixed-point iteration
    Answer and solution

    Answer: A

    Bisection keeps a bracket [a,b] with f(a)f(b)<0 and halves it every step; for a continuous function a root always stays inside, so it converges (linearly, gaining one binary digit per step). The other methods are faster but can diverge or cycle.
  19. NATStandard

    The secant method is applied to f(x)=x2−3 with x0=1 and x1=2. The next iterate x2 is ______ (three decimal places).
    Answer and solution

    Answer: 1.667 (±0.002)

    f(1)=−2, f(2)=1. x2=x1−f(x1)x1−x0f(x1)−f(x0)=2−1⋅11−(−2)=2−13≈1.667. (The root is 3≈1.732.)

    Why people get this wrong: Using f(x0)−f(x1) in the denominator flips the correction and gives 2.333.

  20. MCQStandard

    The forward difference f(x+h)−f(x)h approximates f′(x) with truncation error of order
    1. h
    2. h2
    3. h3
    4. 1 (it does not improve as h→0)
    Answer and solution

    Answer: A

    Taylor: f(x+h)=f(x)+hf′(x)+h22f′′(x)+…, so the forward difference is f′(x)+h2f′′(x)+…: error O(h). The central difference cancels the h term and is O(h2).

    Why people get this wrong: Mixing up the forward difference with the central one, which is second order.

  21. NATExam level

    Newton-Raphson for the cube root of 10 uses f(x)=x3−10. Starting from x0=2, the first iterate x1 is ______.
    Answer and solution

    Answer: 2.1667 (±0.0005)

    x1=x0−x03−103x02=2−8−1012=2+16=2.1667. The true value is 101/3=2.1544.
  22. MCQExam level

    Newton-Raphson is applied to f(x)=x2, whose root x=0 is a double root. Starting from x0=1, the iterates
    1. converge quadratically
    2. converge linearly, the error halving each step
    3. diverge
    4. reach the root in one step
    Answer and solution

    Answer: B

    xn+1=xn−xn22xn=xn2: the error halves each step, which is linear convergence. At a root of multiplicity m Newton converges linearly with ratio (m−1)/m; the modified step x−mf/f′ restores quadratic convergence.

    Why people get this wrong: Quadratic convergence needs f′(r)≠0; at a multiple root f′(r)=0.

  23. NATExam level

    The explicit Euler method is used on dy/dt=−50y. The largest step size h for which the numerical solution does not grow is ______.
    Answer and solution

    Answer: 0.04 (±0.0005)

    Euler gives yn+1=(1−50h)yn. The solution stays bounded when |1−50h|≤1, i.e. 0<h≤2/50=0.04. For 0.02<h<0.04 it oscillates in sign while decaying; beyond 0.04 it blows up, although the true solution decays. Such "stiff" problems call for implicit methods.
  24. NATExam level

    Heun's method (improved Euler, a second-order Runge-Kutta method) is used for dy/dx=y, y(0)=1, with one step h=0.1. It gives y(0.1)≈ ______.
    Answer and solution

    Answer: 1.105 (±0.0005)

    Predictor: k1=f(0,1)=1, so y∗=1+0.1=1.1. Corrector slope k2=f(0.1,1.1)=1.1. y1=1+0.12(1+1.1)=1.105. Exact: 1.10517; the error dropped from 0.005 (Euler) to 0.00017.
  25. MCQExam level

    For a smooth integrand, halving the step size in the composite Simpson's 1/3 rule reduces the error by a factor of about
    1. 2
    2. 4
    3. 8
    4. 16
    Answer and solution

    Answer: D

    The composite Simpson error is −(b−a)h4180f(4)(ξ), so it scales as h4: halving h divides it by 24=16. (The trapezoidal rule, ∝h2, gains only a factor 4.)
  26. NATExam level

    A root of f lies in [1,2]. The number of bisection steps needed to shrink the bracketing interval below 10−6 is ______.
    Answer and solution

    Answer: 20

    After n steps the interval is 2−n. We need 2−n<10−6: n>6log2⁡10=19.93, so n=20 (2−20≈9.5×10−7). Bisection gains about 3.3 steps per decimal digit, regardless of the function.
  27. MCQExam level

    The interpolating polynomial through 4 data points with distinct x values
    1. is unique and has degree at most 3
    2. has degree exactly 4
    3. is unique only if the points are equally spaced
    4. can have any degree up to 4, depending on the method
    Answer and solution

    Answer: A

    A polynomial of degree ≤3 has 4 coefficients, fixed by 4 conditions through a Vandermonde system that is invertible for distinct x. Lagrange, Newton divided differences and the direct solve give the same polynomial. Its degree can be below 3 if the data happen to lie on a lower-degree curve.
  28. NATExam level

    The quadratic through the points (0,1), (1,3), (2,7) is used to extrapolate. Its value at x=3 is ______.
    Answer and solution

    Answer: 13

    Forward differences: Δ=2,4 and Δ2=2. Newton's form: p(x)=1+2x+22x(x−1)=x2+x+1. Then p(3)=13. (Continuing the difference table, the next first difference is 4+2=6, so 7+6=13.)
  29. MCQExam level

    For the Jacobi and Gauss-Seidel iterations for Ax→=b→, a simple sufficient condition for convergence from any starting vector is that A is
    1. symmetric
    2. strictly diagonally dominant
    3. upper triangular with zeros on the diagonal
    4. singular
    Answer and solution

    Answer: B

    If |aii|>∑j≠i|aij| for every row, the iteration matrix has norm below 1, so both methods converge. Symmetric positive-definite matrices also guarantee Gauss-Seidel convergence, but symmetry alone does not.
  30. NATExam level

    Simpson's 1/3 rule with two sub-intervals (h=π/2) estimates ∫0πsin⁡xdx as ______.
    Answer and solution

    Answer: 2.0944 (±0.002)

    S=π/23[sin⁡0+4sin⁡(π/2)+sin⁡π]=π6×4=2π3≈2.094. The exact value is 2, so the error is 4.7% with only three points; four sub-intervals bring it to 0.23%.

← All topics · Timed mock test →