0% found this document useful (0 votes)
24 views4 pages

2021 11 Quiz Solution

Uploaded by

Juan Ponitra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
24 views4 pages

2021 11 Quiz Solution

Uploaded by

Juan Ponitra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Optimization and Algorithms

November 10, 2021

Write your name:

Write your student number:

Quiz
1. A basic question. (3 points) Task 1 of the project is about solving the problem
PT PT −1 2
minimize
x,u t=1 kEx(t) − q(t)k + λ t=1 ku(t)k (1)
subject to x(1) = xinitial
x(t + 1) = Ax(t) + Bu(t), for 1 ≤ t ≤ T − 1.

What does Ex(t) represent?

(A) The velocity of the target at time t


(B) The position of our vehicle at time t
(C) The control signal at time t
(D) The velocity of our vehicle at time t
(E) The position of the target at time t
(F) The state of our vehicle at time t

Write your answer (A, B, C, D, E, or F) here: B

2. Levenberg-Marquardt. (4 points) Suppose we use the Levenberg-Marquardt (LM)


method to address the optimization problem

minimize
n
f (x), (2)
x∈R
PP
where f : Rn → R is given by f (x) = 2
p=1 (fp (x)) . Assume that the function
n
fp : R → R is differentiable for 1 ≤ p ≤ P .
Suppose the current iterate is xk . To get the next iterate, the LM method starts by
solving one of the following problems:

(A) PP
minimize
n p=1 fp (xk ) + ∇fp (xk )T (x − xk ) + λk kx − xk k2
x∈R

(B)
minimize
n
f (xk ) + ∇f (xk )T (x − xk ) + λk kx − xk k2
x∈R

(C)
PP 2
minimize
n p=1 fp (xk ) + ∇fp (xk )T (x − xk ) + λk kx − xk k2
x∈R

1
(D)
PP 2
minimize
n p=1 fp (xk ) + ∇fp (xk )T (x − xk ) + λk kx − xk k
x∈R

(E)
2
minimize
n
f (xk ) + ∇f (xk )T (x − xk ) + λk kx − xk k2
x∈R

(F)
P 2
P
minimize
n p=1 fp (xk ) + ∇fp (xk )T (x − xk ) + λk kx − xk k2
x∈R

Which problem?
Write your answer (A, B, C, D, E, or F) here: C

3. Gradient. (3 points) Consider the function f : R3 → R,


   2
x2
f (x1 , x2 , x3 ) =  x3  + 1 .
x1

The gradient ∇f (4, 0, 3) is one of the following vectors:

(A)  
0
36/5
48/5
(B)  
0
9
8
(C)  
48
0
36
(D)  
0
36
48
(E)  
9
0
8
(F)  
48/5
 0 
36/5

2
Which one?
Write your answer (A, B, C, D, E, or F) here: F

4. Random target. (3 points) Task 5 of the project is about analyzing an optimization


problem of the form
K T T −1
!
X X X 2
minimize pk kExk (t) − qk (t)k + λ kuk (t)k (3)
x1 ,u1 ,x2 ,u2
t=1 t=1
|k=1 {z }
f (x1 ,u1 ,x2 ,u2 )
subject to x1 (1) = xinitial
x1 (t + 1) = Ax1 (t) + Bu1 (t) for 1 ≤ t ≤ T − 1
x2 (1) = xinitial
x2 (t + 1) = Ax2 (t) + Bu2 (t) for 1 ≤ t ≤ T − 1.

Recall that A, B, and xinitial are given constants. Also, λ > 0, p1 > 0, p2 > 0, and
p1 + p2 = 1.
One of the following statements about problem (3) is true:

(A) The function f is strongly convex


(B) The function f is a quadratic
(C) The solution of the optimization problem (3) when p1 = 0.7, p2 = 0.3, and
λ = 1 is the same as the solution of the optimization problem (3) when p1 = 0.7,
p2 = 0.3, and λ = 10
(D) The solution of the optimization problem (3) when p1 = 0.7, p2 = 0.3, and
λ = 1 is the same as the solution of the optimization problem (3) when p1 = 0.2,
p2 = 0.8, and λ = 1
(E) We have x?1 (t) = x?2 (t) and u?1 (t) = u?2 (t) for t = 1, 2, . . . , 24, where the symbol
(x?1 , u?1 , x?2 , u?2 ) denotes the solution of optimization problem (3)
(F) The function f is not convex

Which one?
Write your answer (A, B, C, D, E, or F) here: D

5. Convexity. (4 points) Consider the optimization problem

minimize
n
kAx − bk + f (x), (4)
x∈R

where A ∈ Rm×n and b ∈ Rn . Consider the following choices for the form of
function f :

(A) f (x) = kx − ck
(B) f (x) = (cT x)2
Tx
(C) f (x) = ec + kck2
(D) f (x) = cT x
(E) f (x) = (x − c)T x

3
(F) f (x) = kxk + cT x.

For one of the six forms of f above, the optimization problem (4) is guaranteed
to have a unique global minimizer, regardless of how the constants A, b, and c are
chosen.
For which form?
Write your answer (A, B, C, D, E, or F) here: E

6. Trade-off. (3 points) Consider the optimization problem

minimize
n
f (x) + ρ g(x), (5)
x∈R

where f and g are nonnegative functions Rn → R (that is, f (x) ≥ 0 and g(x) ≥ 0
for all x ∈ Rn ), and ρ is a positive number.
Suppose that x1 is a global minimizer for (5) when ρ = ρ1 , and suppose that x2 is a
global minimizer for (5) when ρ = ρ2 . Consider that ρ2 > ρ1 > 0.
One of the following inequalities is guaranteed to be true:

(A) f (x1 ) + ρ1 g(x1 ) > f (x2 ) + ρ2 g(x2 )


(B) f (x2 ) > f (x1 )
(C) g(x2 ) ≥ g(x1 )
(D) g(x2 ) > g(x1 )
(E) f (x1 ) + ρ1 g(x1 ) < f (x2 ) + ρ2 g(x2 )
(F) f (x2 ) ≥ f (x1 )

Which inequality?
Write your answer (A, B, C, D, E, or F) here: F

You might also like