CHEN20051

One hour

THE UNIVERSITY OF MANCHESTER

CHEMICAL ENGINEERING OPTIMISATION

[Date of Examination]

[Time of Examination]

Answer both questions in this exam paper

Total 50 marks

Electronic calculators may be used in accordance with the University regulations.

Tables of mathematical formulae may be used without restriction.

You may use your notes and other material related to the course.

© The University of Manchester, 2020

Page 1 of 5
 CHEN20051

Question 1: Nonlinear Programming [25 marks in total].

1. Consider the following unconstrained optimisation problem:

min 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) = 2𝑥1 + 2𝑥22 + 𝑎2 ∙ 𝑥33 − 𝑥2

where 𝑎 is a real number and 𝑥1 , 𝑥2 , 𝑥3 are continuous variables to be optimised.

a. The objective function 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) is (multiple correct choices are possible):

A. A strictly convex function;
B. A non-strictly convex function;
C. A strictly concave function;
D. Neither a convex nor a concave function;
E. A non-strictly concave function;
F. Both a convex and a concave function;

[2.5 marks]

b. The eigenvalues of 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) are (multiple correct choices are possible):

A. 0 B. 2 C. 4 D. 6 E. 6𝑎 F. 6𝑎2
[2.5 marks]

c. The stationary points (𝑥1∗ , 𝑥2∗ , 𝑥3∗ ) of 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) are (multiple correct choices
are possible):

A. (0.5, 0.25, 0) B. (2, 0.25, 0) C. (0.5, 0, 3𝑎2 ) D. (0.5, 0, 6𝑎2 )

E. (0.5, 0.25, 6𝑎2 ) F. Do not exist
[2.5 marks]

d. Regarding the above optimisation problem, the point (0.5, 0.25, 0) could be
(multiple correct answers are possible):

A. A local, but not global, minimum solution;

B. A global minimum solution;
C. A global maximum solution;
D. A local, but not global, maximum solution;
E. A saddle point;
F. Neither a minimum solution nor a maximum solution nor a saddle point.
[2.5 marks]

2. Consider the following constrained optimisation problem:

min 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) = 𝑥12 + 𝑥2 + 𝑥3

s. t. 𝑥12 + 𝑥22 + 𝑥32 = 1

Page 2 of 5
 CHEN20051

where 𝑥1 , 𝑥2 , 𝑥3 are continuous variables to be optimised.

a. This optimisation problem has (multiple correct answers are possible):

A. A convex objective function;

B. A concave objective function;
C. A non-convex and non-concave objective function;
D. A non-convex region;
E. A convex region;
F. A convex programming problem.
[2.5 marks]

b. Assume 𝑤 is the Lagrange multiplier, the necessary conditions of optimality

of this optimisation problem include (multiple correct answers are possible):

A. 𝑤 ∙ (𝑥12 + 𝑥22 + 𝑥32 ) = 0;

B. 1 + 2𝑤 ∙ 𝑥2 = 0;
C. 1 + 2𝑤 ∙ 𝑥3 = 0;
D. 𝑤 ≥ 0;
E. 𝑤 ≤ 0;
F. 𝑤 ∙ (𝑥12 + 𝑥22 + 𝑥32 ) ≤ 0.
[2.5 marks]

c. Write down all the stationary points (𝑥1∗ , 𝑥2∗ , 𝑥3∗ ) of this optimisation problem:

[5.0 marks]

d. For a general constrained one-dimensional optimisation problem:

min 𝑓(𝑥)

s. t. 𝑎<𝑥<𝑏

if 𝑓(𝑥) is a strictly convex function and 𝑎 < 𝑏 (𝑎, 𝑏 are real numbers), choose
all the correct statements (multiple correct answers are possible), and
provide a brief explanation:

A. This problem must have only one global minimum solution;

B. This problem does not have a global maximum solution;
C. This problem must have at least one global minimum solution;
D. This problem must have at least one local minimum solution;
E. This problem can only have at most one global minimum solution;
F. This problem must have one local but not global minimum solution.
[5.0 marks]

End of Question 1.

Paper continues/…

Page 3 of 5
 CHEN20051

Question 2: Linear and Mixed-integer Programming [25 marks in total].

1. Consider the following linear programming problem:

max 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) = 𝑥1 + 𝑥3

s. t. 2𝑥1 − 𝑥2 + 𝑥3 ≤ 2
𝑥2 + 𝑥3 ≤ 2
𝑥1 + 𝑥2 + 𝑥3 ≥ 0

a. This optimisation problem has (multiple correct choices are possible):

A. A convex objective function;
B. A concave objective function;
C. A convex region;
D. A non-convex region;
E. Several concave constraints;
F. A non-convex and non-concave objective function.

[2.5 marks]

b. The values of Lagrange multipliers of this optimisation problem are (multiple

correct choices are possible):

A. 1.0 B. 0 C. -1.0 D. 0.5 E. -0.5 F. 0.25

[2.5 marks]

c. Write down all the active constraints of the above optimisation problem:

[2.5 marks]

d. Choose all the correct statements (multiple correct choices are possible),
and provide a brief explanation:

A. This problem only has one global maximum solution, and this is (2, 2, 0);
B. This problem only has one global maximum solution, but (2, 2, 0) is only a
local not global maximum solution;
C. This problem only has one global maximum solution, and this is (1, 1, 1);
D. This problem only has one global maximum solution, but (1, 1, 1) is only a
local not global maximum solution;
E. This problem only has two global maximum solutions, and they are (2, 2,
0) and (1, 1, 1);
F. This problem has more than two global maximum solutions.
[5.0 marks]

e. List two optimisation algorithms that can solve this optimisation problem:

[2.5 marks]

Page 4 of 5
 CHEN20051

2. Consider the following mixed-integer linear programming problem:

max 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) = 𝑥1 + 𝑥3

s. t. 2𝑥1 − 𝑥2 + 𝑥3 ≤ 2
𝑥2 + 𝑥3 ≤ 2
𝑥1 + 𝑥2 + 𝑥3 ≥ 0
𝑥1 ≥ 0, 𝑥3 ≥ 0
where 𝑥1 and 𝑥2 are continuous variables, and 𝑥3 is an integer variable:

a. This optimisation problem (multiple correct answers are possible):

A. has a convex region;

B. has a non-convex region, and cannot be reformulated into a convex
region;
C. has a non-convex region, but can be reformulated into a convex region;
D. is a convex programming problem;
E. is not a convex programming problem, and cannot be reformulated into a
convex programming problem;
F. is not a convex programming problem, but can be reformulated into a
convex programming problem;
[2.5 marks]

b. Choose all the correct statements (multiple correct choices are possible),
and provide a brief explanation:

A. This problem only has one global maximum solution;

B. This problem only has two global maximum solutions;
C. This problem has more than two global maximum solutions;
D. (2, 2, 0) is a global maximum solution of this problem;
E. (2, 2, 0) is one of the local maximum solutions but not a global maximum
solution of this problem.
F. (2, 2, 0) is the only local maximum solution of this problem, but it is not a
global maximum solution of this problem.
[5.0 marks]

c. List an optimisation algorithm that can solve this optimisation problem:

[2.5 marks]

End of Question 2.

END OF EXAMINATION PAPER

Page 5 of 5