LAISHA
Open Laisha

Indian Statistical Institute admission tests Optimization Techniques for Statisticians: practice questions

Indian Statistical Institute admission tests (BStat, MStat, JRF in Statistics) · India · Indian Statistical Institute

Original practice questions written to the published syllabus · not past papers
10worked here, free
41written to this exam in this subject
10syllabus units

About Indian Statistical Institute admission tests

Indian Statistical Institute admission tests (BStat, MStat, JRF in Statistics) is a postgraduate admission exam run by Indian Statistical Institute, for candidates in India. Written tests in two parts: an objective mathematics/statistics test and a descriptive subject test, followed by interview for some programmes (reported). JRF selection by written test and interview.

Official site: isical.ac.in.

We have not been able to confirm these details against the exam body's own specification, so fees, dates and pass marks are left out on purpose. Check the current format, eligibility, fees and dates on the official site.

Optimization Techniques for Statisticians in Indian Statistical Institute admission tests

The optimisation theory statisticians use: linear programming with simplex and duality, transportation and assignment, integer and goal programming for farm and crop planning, non-linear optimisation with Kuhn-Tucker conditions, and the numerical optimisers inside likelihood estimation. The Agribusiness track teaches operations research for managers; this goes to the theory.

Our 41 Indian Statistical Institute admission tests questions in this subject: 4 easy, 28 medium, 9 hard, across 10 units.

10 practice questions, worked

Choose an answer, then open the working. Each explanation says why the right answer is right and why the others are not.

These 10 were written to Indian Statistical Institute admission tests and are not shown on any other exam's page.

Question 1 · medium

In the entering column, no entry is positive. What does this show?

  1. The problem is infeasible
  2. The problem is unbounded
  3. The solution is degenerate
  4. The current solution is optimal
Answer and working

B. The problem is unbounded

The entering variable can increase without any basic variable becoming negative.

Question 2 · medium

When is a branch-and-bound node fathomed?

  1. Infeasible, integral, or no better bound
  2. When its LP value exceeds the incumbent
  3. Only when its LP is infeasible
  4. After a fixed number of branches
Answer and working

A. Infeasible, integral, or no better bound

Each of the three conditions shows that no better integer solution can lie in that branch.

Question 3 · hard

Why are KKT points not enough to find maximum-likelihood estimates of a normal mixture?

  1. Non-concavity: local optima
  2. KKT applies only to linear objectives
  3. Mixture likelihoods have no derivatives
  4. The constraints are always infeasible
Answer and working

A. Non-concavity: local optima

Non-concave likelihoods have several stationary points; multiple starts or global methods are needed.

Question 4 · hard

Which statistical algorithm is an application of dynamic programming?

  1. The Kaplan-Meier estimator
  2. Ordinary least squares
  3. Fisher's exact test
  4. The Viterbi algorithm
Answer and working

D. The Viterbi algorithm

Viterbi finds the most probable hidden state path by a DP recursion; sequence alignment is another example.

Question 5 · hard

What does the BFGS method approximate?

  1. The gradient from function values
  2. The constraints by penalties
  3. The objective by a linear function
  4. The Hessian, from gradients
Answer and working

D. The Hessian, from gradients

Quasi-Newton updates build curvature information without computing second derivatives.

Question 6 · easy

Let the feasible set of a linear program be non-empty and bounded. Which statement is always true?

  1. The optimum is unique
  2. An optimal solution is attained at an extreme point of the feasible set
  3. Every extreme point of the feasible set is optimal
  4. The optimal value is zero when the feasible set contains the origin
Answer and working

B. An optimal solution is attained at an extreme point of the feasible set

A linear function on a non-empty compact polyhedron attains its maximum, and since the polyhedron is the convex hull of its finitely many extreme points, the maximum is attained at one of them. The optimum need not be unique: an objective parallel to an edge gives a whole edge of optima. Extreme points generally have different objective values. Containing the origin says nothing about the optimal value.

Question 7 · medium

For the primal problem max cᵀx subject to Ax ≤ b, x ≥ 0, suppose the objective is unbounded above on the feasible set. Then the dual problem:

  1. Has an unbounded objective as well
  2. Has a unique optimal solution
  3. Is infeasible
  4. Has optimal value equal to zero
Answer and working

C. Is infeasible

By weak duality every dual feasible y gives bᵀy ≥ cᵀx for all primal feasible x; if the primal is unbounded no finite bᵀy can bound it, so the dual can have no feasible point. An unbounded dual would force the primal to be infeasible, not unbounded. A unique dual optimum or a zero value would give a finite bound on the primal, which is a contradiction.

Question 8 · medium

For max 3x₁ + 2x₂ subject to x₁ + x₂ ≤ 4, x₁ + 3x₂ ≤ 6, x₁, x₂ ≥ 0, the primal optimum is x = (4, 0). By complementary slackness the optimal dual solution (y₁, y₂) is:

  1. (0, 1)
  2. (2, 1/3)
  3. (7/2, −1/2)
  4. (3, 0)
Answer and working

D. (3, 0)

At x = (4, 0) the second constraint is slack (4 < 6), so y₂ = 0; x₁ > 0 makes the first dual constraint tight, y₁ + y₂ = 3, so y₁ = 3, and bᵀy = 12 = cᵀx confirms optimality. (0, 1) gives y₁ + y₂ = 1 < 3, infeasible. (2, 1/3) has y₂ > 0 on a slack constraint. (7/2, −1/2) solves both dual constraints as equations but has a negative y₂.

Question 9 · medium

An n × n assignment problem is written as a transportation problem with unit supplies and demands. For n ≥ 2, every basic feasible solution is:

  1. Degenerate, as only n of its 2n − 1 basic variables are positive
  2. Non-degenerate, as all 2n − 1 basic variables are positive
  3. Infeasible unless the cost matrix is symmetric
  4. Optimal whenever the costs are all non-negative
Answer and working

A. Degenerate, as only n of its 2n − 1 basic variables are positive

Each row and column receives exactly one unit, so exactly n cells are positive while a basis has 2n − 1 variables; the other n − 1 are basic at zero, which is why the Hungarian method is used in place of the transportation simplex. All 2n − 1 cannot be positive because only n assignments exist. Feasibility does not need symmetry. Non-negative costs do not make every basic solution optimal.

Question 10 · hard

For which values of k is f(x, y) = x² + kxy + y² a convex function on ℝ²?

  1. k ≥ 0 only
  2. All real k
  3. |k| ≤ 2
  4. |k| ≤ 1
Answer and working

C. |k| ≤ 2

The Hessian is [[2, k], [k, 2]], which is positive semidefinite when its diagonal is non-negative and its determinant 4 − k² ≥ 0, so |k| ≤ 2. Negative k down to −2 is allowed, so 'k ≥ 0 only' is wrong. For |k| > 2, for example k = 3, f(1, −1) = −1 < 0 = f(0, 0) shows non-convexity. |k| ≤ 1 is sufficient but leaves out 1 < |k| ≤ 2.

A study plan for Optimization Techniques for Statisticians in Indian Statistical Institute admission tests

This plan follows where our questions actually fall, so the time goes where the marks are likely to be. It is our reading of the syllabus, not the exam body's weighting.

  1. First, the heavy units (54% of the 41 questions): Integer programming: branch and bound (6, 15%); Numerical optimisation: gradient, Newton and quasi-Newton methods (6, 15%); Dynamic programming (5, 12%); Non-linear programming: Lagrange multipliers and Kuhn-Tucker conditions (5, 12%).

    Read the manual chapters for these first and do every worked question you can find on them.

  2. Then the middle band: Duality and sensitivity analysis (4, 10%); Goal programming for multi-objective farm planning (3, 7%); Linear programming: formulation and graphical solution (3, 7%); Quadratic programming and risk programming (MOTAD) (3, 7%).

    Aim for one sitting per unit, then mixed questions across them.

  3. Then the lighter units: Simplex method, Big-M and two-phase methods (3, 7%); Transportation and assignment problems (3, 7%).

    Learn the definitions, classifications and standard numbers; these units are more often recalled than worked.

  4. Leave time for the hard ones: Numerical optimisation: gradient, Newton and quasi-Newton methods (3 hard of 6), Non-linear programming: Lagrange multipliers and Kuhn-Tucker conditions (2 hard of 5), Quadratic programming and risk programming (MOTAD) (1 hard of 3).

    Come back to these after a first pass, untimed, and read every explanation, including for the options you rejected.

  5. Bank the easy marks: Linear programming: formulation and graphical solution (1 easy of 3); Simplex method, Big-M and two-phase methods (1 easy of 3).

    Do these early in revision so they are fast on the day.

  6. Finish timed. Sit the subject as a timed set without notes, mark it, and send each wrong answer back to its unit above.

Last updated · How Laisha writes and checks its pages · Report a mistake

Ask Laisha about this

Opens Laisha's chat with your question ready to send. Sign in or start free.

Unlock every question — Indian Statistical Institute admission tests Optimization Techniques for Statisticians

US$10 a month · one price worldwide · cancel any time

  • Every practice question in this subject, with full working
  • Timed exam papers and progress by unit
  • Ask Laisha to explain any answer — up to 1,000 questions a month
Free or Smarter
FreeSmarter
Grow guides, pest pages, articlesRead allRead all
Ask Laisha5 a dayUp to 1,000 a month
Exam practiceThe sample on each pageEvery question, timed papers
Crop calendar and pest alerts—For your own fields
Calculators and tools—All
Uploads and reports—100 a month, real files
10 days free for every new account, no card.