NATIONAL INSTITUTE OF TECHNOLOGY CALICUT

Department of Mathematics
MA2014E: Mathematics IV
Winter 2024-2025: Tutorial 5

A Derivative conditions of local maxima and minima, Hessian (Ref: Chong

6.1-6.2)
|f (x)|
1. Let f : Rn → R and g : Rn → R be functions such that lim = 0. (This is denoted as f (x) =
x→0 ∥|x∥
o(g(x)).) Show that for any given ϵ > 0, there exists δ > 0 such that if ∥x∥ < δ, then |f (x)| < ϵ|g(x)|.
2. Let f : Rn → R and g : Rn → R be functions that satisfy f (x) = −g(x) + r(x) where r(x) = o(g(x)) and
g(x) > 0 for all x ̸= 0. Then prove that there exist δ > 0 such that f (x) < 0 for all x with 0 < ∥x∥ < δ.
3. Show that if x∗ is a global minimizer of f over Ω, and x∗ ∈ Ω′ ⊂ Ω, then x∗ is a global minimizer of f
over Ω′ .
4. Find the maximum and minimum values of

f (x, y) = 2 + 2x + 2y − x2 − y 2

on the set (x, y) ∈ R2+ | x + y = 9 by representing the problem as an unconstrained optimization
problem in one variable.
5. Suppose that x∗ is a local minimizer of f over Ω, and Ω ⊂ Ω′ . Show that if x∗ is an interior point of Ω,
then x∗ is a local minimizer of f over Ω′ . Show that the same conclusion cannot be made if x∗ is not
an interior point of Ω.
6. Let c ∈ Rn , c ̸= 0 and consider the problem of minimizing the function f (x) = cT x over a constraint set
Ω ⊂ Rn . Show that we cannot have a solution in the interior of Ω.
7. Let S = {x ∈ Rn : Ax = b}. Show that d ∈ Rn is a feasible direction at x ∈ S if and only if Ad = 0.
2
8. Consider the univariate functions f (x) = (x2 − 1)3 ; g(x) = xe−2x and h(x) = xe−x . Find all the local
minima, maxima, and inflection points of f, g, and h.
9. For the given f defined over S, determine if the given point x∗ is definitely a local minimizer or definitely
not a local minimizer or possibly a local minimizer.
(a) f : R2 → R, S = {(x, y) : x ≥ 1}, x∗ = (1, 2), and gradient ∇f (x∗ ) = (1, 1).
(b) f : R2 → R, S = {(x, y) : x ≥ 1, y ≥ 2}, x∗ = (1, 2), and gradient ∇f (x∗ ) = (1, 0).
(c) f : R2 → R, S = {(x, y) : x ≥ 0, y ≥ 0}, x∗ = (1, 2), and gradient ∇f (x∗ ) = (0, 0), and Hessian
Hf (x∗ ) = I(identity matrix).
(d) f : R2 → R, S = {(x, y) : x ≥ 1, y ≥ 2}, x∗ = (1, 2), and gradient ∇f (x∗ ) = (1, 0), and Hessian
1 0
Hf (x∗ ) = (identity matrix).
0 −1

10. Consider the problem of minimizing −y 2 subject to |y| ≤ x2 and x ≥ 0 where x, y ∈ R. Is the point
(0, 0) a local minimizer, a strict local minimizer, a local maximizer, a strict local maximizer, or none of
the above?
11. Consider the following bivariate functions that are supposed to be minimized over R2 . Find all the points
satisfying the first-order necessary condition. Do these points satisfy second-order necessary conditions?

Please go on to the next page. . .

MA2014E: Tutorial 5 Page 2 of 5

a) f (x1 , x2 ) = x31 + x32

b) f (x1 , x2 ) = (x1 − x2 )4 + x21 − x22 − 2x1 + 2x2 + 1
   
2 5 3
12. Consider the function f : R2 → R defined as f (x) = xT x + xT x + 7.
−1 1 4
(a) Find the directional derivative of f at [0, 1]T in the direction [1, 0]T .
(b) Find all the points that satisfy the first-order necessary condition for f .
(c) Does f have a minimizer? If it does, then find all minimizer(s); otherwise, explain why it does not.

13. Consider the problem of minimizing f (x) over the set S ⊂ R2 where f (x, y) = 5y, and S = {(x, y) :
x2 + y ≥ 1}.
(a) Does the point x∗ = (0, 1)T satisfy the first order necessary condition?
(b) Does the point x∗ = (0, 1)T satisfy the second order necessary condition?
(c) Is the point x∗ = (0, 1)T , a local minimizer?

14. Consider the function f : R2 → R is given by f (x, y) = 4x2 − y 2 , and S = {(x, y) : x2 + 2x − y ≥ 0, x ≥

0, y ≥ 0}. Is the point x∗ = (0, 0) a local minimizer of f over S? If not, why not?
T
15. Consider the quadratic function f : Rn → R given by f (x) = x 2Qx − xT b, where b ∈ Rn and Q is a real
symmetric positive definite n × n matrix. Show that x∗ minimizes f if and only if x∗ satisfies the First
order necessary condition.

16. Consider the problem of minimizing f (x) over the set S ⊂ R2 where S = x ∈ R2 : x21 + x22 ≥ 1 and
f (x) = x2 .
(a) Find all point(s) satisfying the first-order necessary condition.
(b) Which of these points satisfies the second-order necessary condition?
(c) Which of these points are local minimizers?

17. Let φ : R → R be a strictly increasing C 2 function. Let D be an open set in Rn and let f : D → R also
be a C 2 function. Finally, suppose that f has a strict local maximum at x∗ ∈ D, and that Hf (x∗ ) is
negative definite. Show that the composition φ ◦ f also has a strict local maximum at x∗ .

18. Let f : Rn → R be a C 1 function, and let g = −f . Fix any x, and show that the Hessian Hg (x) is
positive definite if and only if Hf (x) is negative definite.

19. Let f (x, y) = x2 + y 2 and g(x, y) = x2 − y 2 . Verify that ∇f (0, 0) = ∇g(0, 0) = (0, 0) and both Hf (0, 0)
and Hg (0, 0 are positive semi-definite. Prove that 0, 0) is a point of local minima for f but not for
g. Does this contradict the second-order necessary or sufficient conditions for the existence of local
minima?

B Line Search Method, Method of Steepest Descent(Ref: Chong 7.3, 8.1-8.2)

20. Using the Newton’s method, find the minimizer of f (x) with the initial value x(0) and the required
accuracy ϵ, in the sense that stop when |x(k+1) − x(k) | < ϵ.

a) f (x) = 6e−2x + 2x2 , x(0) = 0.5, ϵ = 0.05

b) f (x) = x2 + 4 cos(x), x(0) = 1, ϵ = 0.05
c) f (x) = (x − 1)(x − 2)(x − 3), x(0) = 2.2, ϵ = 0.01

Please go on to the next page. . .

MA2014E: Tutorial 5 Page 3 of 5
3 4

4x − 3x x ≥ 0,
21. Consider the function f defined as f (x) = Apply Newton’s method starting from
4x3 + 3x4 x < 0.
two different points, first x0 = 0.40 and then x0 = 0.60. Does the method converge to a stationary
point? Observe six iterations in both cases and then conclude.
22. Use the method steepest decent to find the minimizer of f (x) with the initial value x(0) .
√
(a) f (x) = (x − 5)2 + (y − π)2 + 10, x(0) = (6.597, 5.891)
(b) f (x) = x2 + xy + 2y 2 , x(0) = (2, −2)
(c) f (x) = (x − y)2 + (y − z)4 , x(0) = (−2.6, 2., 2)
(d) f (x) = (x + y)2 + (y + z)2 , x(0) = (−4, 1, 1)
(e) f (x) = (x1 − x32 )2 + 3(x1 − x2 )4
T
23. Let f : Rn → R defined as f (x) = x 2Qx − xT b, where b ∈ Rn and Q is a real symmetric positive definite
n × n matrix. Suppose that we apply the steepest descent method to f (x) with x(0) ̸= Q−1 b. Show that
x(1) = Q−1 b if and only if x(0) is chosen such that Qx(0) − b is an eigenvector of Q.

C Linear Programming Problem: Formulation, Standard form, Basic solutions,

Fundamental Theorem of LPP (Ref: Chong 15.2-15.7)
24. Convert the following linear programming problem to standard form:
a) b) c)

max x1 + x2 max x1 − x2 max 6x1 − 2x2 + 9x3

subject to x1 + 2x2 ≤ 4, subject to 3x2 = x1 − 5 subject to 2x1 − 6x2 − x3 ≤ 100
3x1 + x2 ≥ 1, |x1 | ≤ 1, x1 + x2 + 9x − 3 ≤ 200
x1 , x2 ≥ 0. x2 ≤ 0. 0 ≤ x1 ≤ 50, x2 ≥ −60, x3 ≥ 5

25. Show that the same variable cannot be used for different inequalities when introducing slack variables.
26. (a) Does every linear programming problem in standard form have a non-empty feasible set? Justify
your answer.
(b) Does every linear programming problem in standard form with non-empty feasible set have an
optimal solution? Justify your answer.
27. Suppose x is a feasible solution to the linear program min cT x subject to Ax = b, x ≥ 0. Given that
A is an m × n matrix of rank m, show that there is a feasible solution y having the same value (i.e,
cT y = cT x) and having at most m + 1 positive components.
28. Show that the following LP problems are equivalent
Problem A: Problem B:

M in x1 + 2x2 − 3x3 + 4x4 M in x1 + 2x′2 − 3x′3 + 4x′4 − 3x5

s.t 3x1 − 2x2 + 5x3 − 6x4 = 20, s.t 3x1 − 2x′2 + 5x′3 − 6x′4 + 3x5 = 20,
x1 + 7x2 − 6x3 + 9x4 = 30, x1 + 7x′2 − 6x′3 + 9x′4 − 10x5 = 30,
x1 ≥ 0, x2 , x3 ,x4 - unrestricted. x1 , x′2 , x′3 , x′4 , x5 ≥ 0
29. Suppose a boat manufacturer produces two types of boats for the sports and camping trade, a family
rowboat, and a sports canoe. The boats are molded from aluminum by a large pressing machine and
finished by hand labor. A rowboat requires 50 KG of aluminum, 6 minutes of machine time, and 3

Please go on to the next page. . .

MA2014E: Tutorial 5 Page 4 of 5

hours of finishing labor; a canoe requires 30 KG of aluminum, 5 minutes of machine time, and 5 hours
of finishing labor. For the next 3 months, the company can commit up to 2 tons of aluminum, 5 hours
of machine time, and 200 hours of labor to manufacture the small boats. The company realizes a Rs.50
K profit on the sale of a rowboat and a Rs.60 K profit on the sale of a canoe. Write an LPP formulation
of the problem to maximize the profit.

30. A ghee producer must feed her cattle stock daily at least 550 units of nutritional element A, 500 units
of nutritional element B, and 820 units of nutritional element C. She has available two feeds. 1 KG of
Feed X costs Rs.80 and contains 2 units of A, 5 units of B, and 7 units of C. 1 KG of Feed Y costs
Rs.30 and contains 3 units of A, 1 unit of B, and 2 units of C. The ghee producer wants to determine
what combination of the two feeds will meet the dietary requirements of her cattle and keep costs at a
minimum. Write an LPP formulation to find the least expensive adequate feeding diet.

31. Using the graphical method, find a solution for the problems given in the questions 29 and 30.

32. Find the basic solutions to the following LPP problems, classify them as basic feasible solutions, and
degenerate solutions.

a) b)

M in x1 − 2x2 + 2x3 − 5x4 M in x1 − 2x2 + 2x3 − 5x4

s.t x1 + x2 + 2x3 + x4 = 6, s.t 2x1 + 3x2 − x3 − x4 = −1,
3x2 + x3 + 8x4 = 3, 4x1 + x2 + x3 − 2x4 = 9,
x1 , x2 , x3 , x4 ≥ 0 x1 , x2 , x3 , x4 ≥ 0

33. For the LP problem of

M in 5x1 + 2x2 + 3x3 + x4

s.t x1 + x2 − 2x3 + 3x4 = 2,
−2x1 + x3 = 2,
x1 , x2 , x3 , x4 ≥ 0

(a) Show that there can be only two basic feasible solutions to the problem.
(b) Compute these two basic feasible solutions.

D The Dual of an LPP, duality theorems (Ref: Chong 17.1 and 17.2)

34. Find the dual LPP of the following problems

a) b) c)

M ax 6x1 + x2 + 4x3 M in x1 − 2x2 + 3x3 M in 12x1 + 2x2

s.t 3x1 + 7x2 + x3 ≤ 15, s.t 4x1 + 5x2 − 6x3 = 7, s.t 8x1 − x2 ≤ 21,
x1 − 2x2 + 3x3 = 20, 8x1 − 9x2 + 10x3 ≤ 11, x1 − 6x2 ≥ 13,
x1 , x2 , x3 ≥ 0 x1 , x2 ≥ 0,x3 unrestricted 3x1 − 4x2 = −11,
x1 unrestricted, x2 ≥ 0.

Please go on to the next page. . .

MA2014E: Tutorial 5 Page 5 of 5

35. Using duality or weak duality theorem, show that the following LPP and its dual do not have any
feasible solutions.

M ax x1
s.t x1 − x2 ≤ 1,
−x1 + x2 ≤ −2,
x1 , x2 ≥ 0.

36. Which of the following are possible scenarios? Justify your answers.
a) Primal: Optimal and Dual: Unbounded b) Primal: Optimal and Dual: Infeasible
c) Primal: Unbounded and Dual: Unbounded d) Primal: Unbounded and Dual: infeasible
e) Primal: Infeasible and Dual: Infeasible

37. Consider the LP problem of minimizing cT x subject to Ax ≤ b.

(a) Find the dual to the above problem.
(b) Suppose b = 0, and there exists a vector y ≥ 0 such that y T A + cT = 0T . Does the given LPP have
an optimal feasible solution? Justify your answer.

38. Consider the LP problem of minimizing cT x subject to x ≥ 0. Using duality theorem, prove that solution
to this LPP exists if and only if c ≥ 0.
39. Let A be a given matrix, and b be a given vector. Show that there exists a vector x such that Ax ≥ b
and x ≥ 0 if and only if for any given vector y satisfying AT y ≤ 0 and y ≥ 0, we have bT y ≤ 0.

40. Let A be a given matrix, and b be a given vector. Show that there exists a vector x such that Ax = b
and x ≥ 0 if and only if for any given vector y satisfying AT y ≤ 0, we have bT y ≤ 0.
41. Let A be a given matrix, and b be a given vector. Show that there exists a vector x such that Ax ≤ b
if and only if for any given vector y satisfying AT y = 0 and y ≥ 0, we have bT y ≥ 0.
42. Let A be a given matrix, and b be a given vector. Show that there exists a vector x such that Ax < 0
if and only if for any given vector y satisfying AT y = 0 and y ≥ 0, we have y = 0.

End of the page