MCQ _ Convex Lecture_7

Linear Programming (LP):

1. Linear programming is an optimization problem of the form:

a) max
cTx
b) min
cTx
c) max
Ax
d) min
Ax

2. The simplex algorithm, introduced by Dantzig, is a direct solver for:

a) Quadratic programs
b) Linear programs
c) Semidefinite programs
d) Conic programs

3. The diet problem is an example of:

a) Linear programming
b) Second-order cone programming
c) Semidefinite programming
d) Quadratic programming

4. In the transportation problem,cij represents:

a) Cost from source i to destination j
b) Demand at destination j
c) Supply at source i
d) Units of commodity shipped from i to j

Quadratic Programming (QP):

1. Quadratic programming becomes convex when:

a)

MCQ _ Convex Lecture_7 1

 Q is positive definite
b)
Q is positive semidefinite
c)
Q is negative definite
d)
Q is negative semidefinite

2. Portfolio optimization involves trading off:

a) Performance and risk
b) Supply and demand
c) Cost and benefit
d) Nutritional requirements

3. Support vector machines (SVM) optimization is a:

a) Linear program
b) Quadratic program
c) Second-order cone program
d) Semidefinite program

4. The lasso problem involves minimizing the l1-norm subject to:

a)
Ax=b
b)
Xβ=y
c)
Dx≤d
d)
kβk1≤s

Semidefinite Programming (SDP):

1. The Lovasz theta function is used for:

a) Portfolio optimization
b) Solving linear systems
c) Graph problems
d) Convex hull computation

MCQ _ Convex Lecture_7 2

 2. Trace norm minimization in SDPs is associated with:
a) Finding the highest-rank solution
b) Finding the lowest-rank solution
c) Minimizing the
l1-norm
d) Maximizing the
l2-norm

Conic Programming:

1. Conic programming involves optimization over:

a) Real numbers
b) Vectors
c) Cones
d) Matrices

2. Every linear program is a special case of conic programming with the cone:
a)
Sn
b)
Sn+
c) R
n
d)
Sn++

Second-Order Cone Programming (SOCP):

1. A second-order cone program has constraints in the form:

a)
kDix+dik2 ≤ eiTx+fi
b)
Dix+di ≤ eiTx+fi
c)
kDix+dik2 ≥ eiTx+fi
d)
Dix+di ≥ eiTx+fi

MCQ _ Convex Lecture_7 3

 2. The hierarchy LPs ⊆ QPs ⊆SOCPs ⊆ SDPs ⊆ Conic programs indicates that:
a) QPs are equivalent to SOCPs
b) QPs are a special case of SOCPs
c) SOCPs are a special case of QPs
d) SOCPs and QPs are unrelated

Miscellaneous:

1. The Schur complement theorem is used to:

a) Transform linear programs into quadratic programs
b) Write SOCP constraints as SDP constraints
c) Solve semidefinite programs efficiently
d) Convert convex optimization problems to nonconvex problems

2. The Lovasz sandwich theorem relates the Lovasz theta function to:
a) Clique number and chromatic number
b) Portfolio optimization
c) Semidefinite programming
d) Quadratic programming

3. A quadratic program is in standard form when:

a)
Ax=b
b)
x≥0
c)
Q≻0
d)
cTx + 0.5xTQx

4. The statement "Any quadratic program can be rewritten in standard form" means:
a) All quadratic programs are equivalent
b) Standard form simplifies the problem structure
c) Standard form is more complex than the original problem
d) Standard form is applicable only to specific cases

5. The motivation for semidefinite programming involves generalizing linear

programming by changing:
a) Constraints from equality to inequality

MCQ _ Convex Lecture_7 4

 b) Constraints from inequality to equality
c) Constraints from ≤ to different (partial) order
d) Objectives from minimization to maximization

6. In semidefinite programming, the space of positive semidefinite matrices is denoted

by:
a)
Sn
b)
Sn+
c)
Sn++
d) R
n

7. A conic program involves optimization over:

a) Vectors
b) Real numbers
c) Cones
d) Matrices

8. The hierarchy LPs ⊆ QPs ⊆ SOCPs ⊆ SDPs ⊆ Conic programs implies:

a) QPs are more general than SDPs
b) SDPs are more general than LPs
c) SOCPs are a subset of QPs
d) Conic programs are more general than SOCPs

9. The Dantzig selector allows for an approximation (Xβ≈y) without requiring exact
equality, and it is a:
a) Linear program
b) Quadratic program
c) Second-order cone program
d) Semidefinite program

10. The second-order cone Q is defined as:

a) {(
x,t):kxk2≤t}
b) {(
x,t):kxk2≥t}

MCQ _ Convex Lecture_7 5

 c) {(
x,t):kxk2=t}
d) {(
x,t):kxk2<t}

11. The Schur complement theorem relates a matrix A to its:

a) Determinant
b) Inverse
c) Eigenvalues
d) Transpose

12. Quadratic programs "sneak" into the hierarchy by being equivalent to:
a) SDPs
b) SOCPs
c) LPs
d) Conic programs

13. The Lovasz theta function is associated with:

a) Graph problems
b) Portfolio optimization
c) Linear systems
d) Quadratic programs

14. Trace norm minimization in SDPs is associated with:

a) Finding the highest-rank solution
b) Finding the lowest-rank solution
c) Minimizing the
l1-norm
d) Maximizing the
l2-norm

15. A conic program involves optimization over:

a) Real numbers
b) Vectors
c) Cones
d) Matrices

16. Every linear program is a special case of conic programming with the cone:
a)

MCQ _ Convex Lecture_7 6

 Sn
b)
Sn+
c) R
n
d)
Sn++

Linear Programming (LP):

1. Which algorithm provides a noniterative solver for LPs?

a) Interior point methods
b) Dantzig's simplex algorithm
c) Gradient descent
d) Newton's method

2. In the diet problem, what does Dij represent?

a) Cost of food
j
b) Minimum required intake of nutrient
i
c) Content of nutrient
i per unit of food j
d) Units of food
j in the diet

3. What is the interpretation of xij in the transportation problem?

a) Per-unit shipping cost from
i to j
b) Units shipped from
i to j
c) Supply at source
i
d) Demand at destination
j

Quadratic Programming (QP):

MCQ _ Convex Lecture_7 7

 1. What does the term 0.5xTQx represent in a quadratic program?
a) Quadratic approximation
b) Linear approximation
c) Objective function
d) Regularization term

2. In portfolio optimization, what does μ represent?

a) Covariance matrix
b) Expected assets' returns
c) Risk aversion
d) Portfolio holdings

3. What type of problem is the support vector machine (SVM)?

a) Linear program
b) Quadratic program
c) Second-order cone program
d) Semidefinite program

Semidefinite Programming (SDP):

1. What is the Lovasz theta function used for in graph theory?

a) Clustering
b) Clique number computation
c) PageRank calculation
d) Random walk analysis

2. In trace norm minimization, what does kXktr represent?

a) Frobenius norm
b) Spectral norm
c) Trace norm
d) Operator norm

Conic Programming:

1. What is D(x)+d in a conic program?

a) Linear map
b) Objective function
c) Decision variable
d) Constraint

MCQ _ Convex Lecture_7 8

 Second-Order Cone Programming (SOCP):

1. What is the second-order cone Q defined as?

a) {(
x,t):kxk2≤t}
b) {(
x,t):kxk2≥t}
c) {(
x,t):kxk2=t}
d) {(
x,t):kxk2<t}

2. Why are QPs considered SOCPs?

a) They share a common objective function
b) Every QP can be rewritten as an SOCP
c) SOCPs are a subset of QPs
d) QPs are equivalent to SOCPs

Miscellaneous:

1. What does the Lovasz sandwich theorem relate to in graph theory?

a) Minimum spanning trees
b) Chromatic number and clique number
c) Network flows
d) Hamiltonian cycles

2. In SDPs, what space does positive definite matrices (Sn++) belong to?
a) R
n
b)
Sn
c)
Sn+
d) R
n×n

3. What does the hierarchy LPs ⊆ QPs ⊆ SOCPs ⊆ SDPs ⊆ Conic programs imply?
a) SDPs are equivalent to QPs

MCQ _ Convex Lecture_7 9

 b) SOCPs are a subset of QPs
c) QPs are more general than SDPs
d) Conic programs are equivalent to LPs

4. What does the Schur complement theorem relate to in the context of matrices?
a) Matrix inversion
b) Diagonalization
c) Matrix factorization
d) Matrix positivity

5. What is the motivation for semidefinite programming?

a) Achieving a unique solution
b) Generalizing linear programming
c) Solving non-convex problems
d) Enhancing numerical stability

6. What is the Lovasz theta function used to approximate?

a) Chromatic number
b) Clique number
c) Network connectivity
d) Eigenvector centrality

7. In the context of conic programming, what does the cone K represent?

a) Linear maps
b) Closed convex cone
c) Objective function
d) Decision variable

8. What does the term "basis pursuit" refer to in optimization?

a) Finding a basis for a vector space
b) Sparse approximation
c) Basis set expansion
d) Linear independence

9. What is the primary purpose of the Dantzig selector?

a) Convex relaxation
b) Sparse approximation
c) Matrix factorization
d) Eigenvalue computation

MCQ _ Convex Lecture_7 10

 Linear Programming (LP):

1. Who introduced linear programming in the late 1930s?

a) Richard Bellman
b) George Dantzig
c) Leonid Kantorovich
d) John von Neumann

2. What is the significance of Dantzig's simplex algorithm?

a) It provides an iterative solver for LPs
b) It solves non-convex optimization problems
c) It is only applicable to quadratic programs
d) It handles only linear constraints

3. LPs are considered convex optimization problems. Why?

a) The objective function is quadratic
b) The feasible region is a convex set
c) LPs always have a unique solution
d) LPs involve non-linear constraints

Quadratic Programming (QP):

1. What does the term kβk0 represent in the basis pursuit problem?
a) ℓ0ℓ0 norm of
β
b) ℓ1ℓ1 norm of
β
c) ℓ2ℓ2 norm of
β
d) Frobenius norm of
β

2. In the context of the lasso problem, what does s represent?

a) Tuning parameter
b) Regularization term
c) Solution vector
d) Scalar constant

3. Why is the Dantzig selector considered a modification of basis pursuit?

a) It has a different objective function

MCQ _ Convex Lecture_7 11

 b) It allows for approximate equality
c) It uses a different norm for approximation
d) It involves a different set of constraints

Semidefinite Programming (SDP):

1. Why is SDP considered a generalization of linear programming?

a) It involves quadratic objectives
b) It allows for non-linear constraints
c) It generalizes the concept of positive semidefinite matrices
d) It is applicable only to non-convex problems

2. In the context of SDP, what is kXktr?

a) Trace norm of
X
b) Frobenius norm of
X
c) Spectral norm of
X
d) Rank of
X

Conic Programming:

1. In a conic program, what does the closed convex cone K represent?

a) Feasible region
b) Decision variable
c) Objective function
d) Constraint set

Second-Order Cone Programming (SOCP):

1. How does the second-order cone Q differ from a standard Euclidean ball?
a) It has a non-convex shape
b) It is not a closed set
c) It allows for quadratic constraints
d) It involves a combination of linear and quadratic constraints

MCQ _ Convex Lecture_7 12

 Miscellaneous:

1. What is the primary purpose of the trace norm minimization problem?

a) Clustering
b) Matrix factorization
c) Sparse approximation
d) Eigenvalue computation

2. What does the Lovasz theta function aim to maximize?

a) Chromatic number
b) Clique number
c) Random walk efficiency
d) Spectral radius

3. How are second-order cone programs related to quadratic programs?

a) SOCPs are equivalent to QPs
b) QPs are a special case of SOCPs
c) SOCPs and QPs are mutually exclusive
d) QPs cannot be reformulated as SOCPs

4. What does the Schur complement theorem provide a condition for?

a) Matrix orthogonality
b) Matrix positivity
c) Matrix diagonalization
d) Matrix sparsity

5. In portfolio optimization, what does the covariance matrix Q represent?

a) Expected assets' returns
b) Risk aversion
c) Portfolio holdings
d) Market volatility

Linear Programming:

1. Which of the following is NOT a characteristic of a linear program?

a) It is a convex optimization problem.
b) It involves a linear objective function.
c) It can have nonlinear constraints.
d) It can be solved using the simplex algorithm.

MCQ _ Convex Lecture_7 13

 2. The diet problem is an example of a linear program that aims to:
a) Maximize profit.
b) Minimize cost.
c) Find the shortest path.
d) Schedule tasks.

3. The standard form of a linear program is written as:

a) min c^T x subject to Ax = b, x ≥ 0
b) max c^T x subject to Ax ≤ b
c) min c^T x subject to Dx ≤ d, Ax = b
d) max c^T x subject to Ax = b, x ≥ 0

4. The Dantzig selector is a modification of the basis pursuit problem that:

a) Allows for approximate solutions.
b) Introduces a regularization term.
c) Involves a nonlinear constraint.
d) Uses a different optimization algorithm.

Quadratic Programming:

1. A quadratic program is convex if and only if the matrix Q is:

a) Positive definite.
b) Positive semidefinite.
c) Negative definite.
d) Negative semidefinite.

2. Portfolio optimization is an example of a quadratic program that aims to:

a) Maximize return while minimizing risk.
b) Find the optimal portfolio weights.
c) Predict future stock prices.
d) Hedge against market volatility.

3. The lasso problem is a quadratic program that is used for:

a) Linear regression with regularization.
b) Feature selection.
c) Sparse signal recovery.
d) All of the above.

Semidefinite Programming:

MCQ _ Convex Lecture_7 14

 1. A semidefinite program is an optimization problem with constraints involving:
a) Linear inequalities.
b) Quadratic inequalities.
c) Positive semidefinite matrices.
d) Non-convex functions.

2. The standard form of a semidefinite program is written as:

a) min C • X subject to Ai • X = bi, i = 1, ..., m, X ≥ 0
b) min C • X subject to Ai • X = bi, i = 1, ..., m, X ⪰ 0
c) max C • X subject to Ai • X = bi, i = 1, ..., m, X ⪰ 0
d) max C • X subject to Ai • X ≤ bi, i = 1, ..., m, X ⪰ 0

3. The Lovasz theta function is an example of a semidefinite program used in:

a) Graph theory.
b) Combinatorial optimization.
c) Machine learning.
d) All of the above.

Conic Programming and SOCPs:

1. A conic program is a generalization of linear and semidefinite programs that

involves:
a) A cone constraint.
b) A convex objective function.
c) A linear constraint.
d) All of the above.

2. A second-order cone program (SOCP) is a conic program with constraints involving:

a) Second-order cones.
b) Linear inequalities.
c) Quadratic inequalities.
d) Positive semidefinite matrices.

3. Every linear program is an SOCP, and every SOCP is a semidefinite program.

a) True
b) False

Linear Programming:

MCQ _ Convex Lecture_7 15

 1. The transportation problem minimizes the total cost of shipping goods:
a) True
b) False

2. The basis pursuit problem finds the sparsest solution to an underdetermined linear
system by minimizing the:
a) L0 norm
b) L1 norm
c) L2 norm
d) None of the above

3. Duality theory in linear programming provides an alternative viewpoint to optimize

and analyze the problem. For a minimization problem, what does the dual problem
minimize?
a) The primal objective function
b) The negative of the primal objective function
c) The sum of the primal objective function and the dual objective function
d) The Lagrange multipliers of the primal problem

Quadratic Programming:

1. In portfolio optimization, what does the risk aversion parameter γ control?

a) The expected return of the portfolio
b) The variance of the portfolio
c) The trade-off between expected return and risk
d) The number of assets in the portfolio

2. The lasso problem can be equivalently written as a quadratic program by

introducing a slack variable for the L1 constraint. What is the role of the tuning
parameter s in the lasso formulation?
a) It controls the sparsity of the solution
b) It balances the data fitting term and the regularization term
c) It determines the maximum allowable error
d) It has no effect on the solution

3. What is the advantage of using interior-point methods for solving quadratic

programs compared to the simplex algorithm?
a) They can handle non-convex problems
b) They converge faster in general

MCQ _ Convex Lecture_7 16

 c) They work directly with the quadratic objective function
d) They are guaranteed to find the global optimum

Semidefinite Programming:

1. The trace norm minimization problem finds the lowest-rank solution to an

underdetermined system by minimizing the:
a) Singular value sum
b) Frobenius norm
c) Trace norm
d) Nuclear norm

2. The Lovasz theta function provides an upper bound on the chromatic number of a
graph. What is the relationship between the theta function and the clique number of
the graph?
a) The theta function is always greater than the clique number
b) The theta function is always less than or equal to the clique number
c) The theta function is equal to the clique number for all graphs
d) There is no known relationship between the theta function and the clique number

3. What is the main advantage of using semidefinite programs compared to linear

programs for certain optimization problems?
a) They can handle non-convex problems
b) They can optimize over matrices directly
c) They are always easier to solve
d) They guarantee to find the global optimum

Conic Programming and SOCPs:

1. Second-order cone programs are a special type of conic program because their
cone constraints involve:
a) Convex quadratic functions
b) Ellipsoids
c) Hyperplanes
d) Linear combinations of hyperplanes

2. What is the relationship between the feasible region of a linear program and the
feasible region of its corresponding SOCP reformulation?
a) The SOCP feasible region is always smaller than the linear program feasible
region

MCQ _ Convex Lecture_7 17

 b) The SOCP feasible region is always larger than the linear program feasible
region
c) The SOCP feasible region is the same as the linear program feasible region
d) The relationship depends on the specific linear program

3. How can you check if a quadratic program is equivalent to an SOCP?

a) By rewriting the quadratic objective function as a linear objective function in terms
of new variables
b) By diagonalizing the quadratic matrix Q
c) By verifying if all constraints can be written as second-order cone constraints
d) There is no efficient way to check this in general

Answers:

Linear Programming (LP):

Linear programming is an optimization problem of the form:

b) min cTx

The simplex algorithm, introduced by Dantzig, is a direct solver for:

b) Linear programs

The diet problem is an example of:

a) Linear programming

In the transportation problem, cij represents:

a) Cost from source i to destination j

Quadratic Programming (QP):

Quadratic programming becomes convex when:

b) Q is positive semidefinite

Portfolio optimization involves trading off:

a) Performance and risk

MCQ _ Convex Lecture_7 18

 Support vector machines (SVM) optimization is a:

b) Quadratic program

The lasso problem involves minimizing the l1-norm subject to:

d) kβk1≤s

Semidefinite Programming (SDP):

The Lovasz theta function is used for:

c) Graph problems

Trace norm minimization in SDPs is associated with:

b) Finding the lowest-rank solution

Conic Programming:

Conic programming involves optimization over:

c) Cones

Every linear program is a special case of conic programming with the cone:

b) Sn+

Second-Order Cone Programming (SOCP):

A second-order cone program has constraints in the form:

a) kDix+dik2 ≤ eiTx+fi

The hierarchy LPs ⊆ QPs ⊆ SOCPs ⊆ SDPs ⊆ Conic programs indicates

that:

b) QPs are a special case of SOCPs

Miscellaneous:

The Schur complement theorem is used to:

b) Write SOCP constraints as SDP constraints

The Lovasz sandwich theorem relates the Lovasz theta function to:

a) Clique number and chromatic number

A quadratic program is in standard form when:

MCQ _ Convex Lecture_7 19

 d) cTx + 0.5xTQx

4 ⇒b
The motivation for semidefinite programming involves generalizing linear
programming by changing:

c) Constraints from ≤ to different (partial) order

In semidefinite programming, the space of positive semidefinite matrices is

denoted by:

b) Sn+

A conic program involves optimization over:

c) Cones

The hierarchy LPs ⊆ QPs ⊆ SOCPs ⊆ SDPs ⊆ Conic programs implies:

d) Conic programs are more general than SOCPs

The Dantzig selector allows for an approximation (Xβ≈y) without requiring

exact equality, and it is a:

a) Linear program

The second-order cone Q is defined as:

a) {(x,t):kxk2≤t}

The Schur complement theorem relates a matrix A to its:

b) Inverse

Quadratic programs "sneak" into the hierarchy by being equivalent to:

b) SOCPs

The Lovasz theta function is associated with:

a) Graph problems

Trace norm minimization in SDPs is associated with:

b) Finding the lowest-rank solution

A conic program involves optimization over:

MCQ _ Convex Lecture_7 20

 c) Cones

Every linear program is a special case of conic programming with the cone:

b) Sn+
Linear Programming (LP):

Which algorithm provides a noniterative solver for LPs?

b) Dantzig's simplex algorithm

In the diet problem, what does Dij represent?

c) Content of nutrient i per unit of food j

What is the interpretation of xij in the transportation problem?

b) Units shipped from i to j

Quadratic Programming (QP):

What does the term 0.5xTQx represent in a quadratic program?

c) Objective function

In portfolio optimization, what does μ represent?

b) Expected assets' returns

What type of problem is the support vector machine (SVM)?

b) Quadratic program

Semidefinite Programming (SDP):

What is the Lovasz theta function used for in graph theory?

b) Clique number computation

In trace norm minimization, what does kXktr represent?

c) Trace norm

Conic Programming:

What is D(x)+d in a conic program?

d) Constraint

Second-Order Cone Programming (SOCP):

MCQ _ Convex Lecture_7 21

 What is the second-order cone Q defined as?

a) {(x,t):kxk2≤t}

Why are QPs considered SOCPs?

b) Every QP can be rewritten as an SOCP

Miscellaneous:

What does the Lovasz sandwich theorem relate to in graph theory?

b) Chromatic number and clique number

In SDPs, what space does positive definite matrices (Sn++) belong to?

c) Sn+

What does the hierarchy LPs ⊆ QPs ⊆ SOCPs ⊆ SDPs ⊆ Conic programs
imply?

c) QPs are more general than SDPs

What does the Schur complement theorem relate to in the context of

matrices?

a) Matrix inversion

What is the motivation for semidefinite programming?

b) Generalizing linear programming

What is the Lovasz theta function used to approximate?

a) Chromatic number.

a&b

In the context of conic programming, what does the cone K represent?

b) Closed convex cone

What does the term "basis pursuit" refer to in optimization?

b) Sparse approximation

What is the primary purpose of the Dantzig selector?

b) Sparse approximation

MCQ _ Convex Lecture_7 22

 Linear Programming (LP):

Who introduced linear programming in the late 1930s?

b) George Dantzig

What is the significance of Dantzig's simplex algorithm?

a) It provides an iterative solver for LPs

LPs are considered convex optimization problems. Why?

b) The feasible region is a convex set

What does kβk0 represent in the basis pursuit problem?

a) ℓ0ℓ0 norm of β

Quadratic Programming (QP):

In the context of the lasso problem, what does s represent?

a) Tuning parameter

Why is the Dantzig selector considered a modification of basis pursuit?

a) It allows for approximate equality

Semidefinite Programming (SDP):

Why is SDP considered a generalization of linear programming?

c) It generalizes the concept of positive semidefinite matrices

In the context of SDP, what is kXktr?

a) Trace norm of X

Conic Programming and SOCPs:

In a conic program, what does the closed convex cone K represent?

a) Feasible region

How does the second-order cone Q differ from a standard Euclidean ball?

a) It has a non-convex shape

Miscellaneous:

What is the primary purpose of the trace norm minimization problem?

MCQ _ Convex Lecture_7 23

 d) Eigenvalue computation

What does the Lovasz theta function aim to maximize?

b) Clique number

How are second-order cone programs related to quadratic programs?

b) QPs are a special case of SOCPs

What does the Schur complement theorem provide a condition for?

b) Matrix positivity

In portfolio optimization, what does the covariance matrix Q represent?

d) Market volatility

Linear Programming:

1. c) It can have nonlinear constraints. Linear programs, by definition, have linear

constraints.

2. b) Minimize cost. The diet problem seeks the least expensive diet that meets
nutritional requirements.

3. a) min c^T x subject to Ax = b, x ≥ 0 This is the standard form for linear programs.

4. a) Allows for approximate solutions. The Dantzig selector relaxes the exact
equality constraint in basis pursuit.

Quadratic Programming:

1. b) Positive semidefinite. This condition ensures convexity of the quadratic

program.

2. a) Maximize return while minimizing risk. Portfolio optimization balances reward

and risk.

3. d) All of the above. The lasso problem is versatile in linear regression, feature
selection, and sparse signal recovery.

Semidefinite Programming:

1. c) Positive semidefinite matrices. Constraints involve matrices that are positive

semidefinite.

MCQ _ Convex Lecture_7 24

 2. b) min C • X subject to Ai • X = bi, i = 1, ..., m, X ⪰ 0 This is the standard form for
semidefinite programs.

3. a) Graph theory. The Lovasz theta function has applications in graph theory.

Conic Programming and SOCPs:

1. d) All of the above. Conic programs encompass linear, semidefinite, and second-
order cone programs.

2. a) Second-order cones. SOCPs have constraints involving second-order cones.

3. b) False While all LPs are SOCPs, not all SOCPs are SDPs.

Linear Programming (continued):

1. a) True The transportation problem aims to minimize shipping costs.

2. b) L1 norm The basis pursuit problem minimizes the L1 norm for sparsity.

3. b) The negative of the primal objective function The dual problem offers an
alternative perspective.

Quadratic Programming (continued):

1. c) The trade-off between expected return and risk The risk aversion parameter
balances reward and risk.

2. a) It controls the sparsity of the solution The tuning parameter s governs

sparsity in the lasso problem.

3. b) They converge faster in general Interior-point methods often outperform the

simplex algorithm in speed.

Semidefinite Programming (continued):

1. c) Trace norm Trace norm minimization finds the lowest-rank solution.

2. b) The theta function is always less than or equal to the clique number The
Lovasz theta function provides an upper bound on the chromatic number.

3. b) They can optimize over matrices directly SDPs allow for direct optimization
over matrices, unlike LPs.

Conic Programming and SOCPs (continued):

1. b) Ellipsoids SOCPs involve constraints defined by ellipsoids.

MCQ _ Convex Lecture_7 25

 2. c) The SOCP feasible region is the same as the linear program feasible
region LPs and their SOCP reformulations have equivalent feasible regions.

3. c) By verifying if all constraints can be written as second-order cone

constraints This is the key check for SOCP equivalence.

MCQ _ Convex Lecture_7 26