ISYE 6501 MIDTERM 2 — ORIGINAL
PRACTICE QUESTION BANK
QUESTIONS 1–100 WITH ANSWERS AND DETAILED
RATIONALES
INTRODUCTION
This practice bank is designed to provide rigorous preparation for ISYE 6501
Midterm 2–style assessment. The questions emphasize conceptual understanding,
quantitative reasoning, interpretation of analytical results, and selection of appropriate
statistical and machine-learning methods. Scenarios are intentionally constructed to
test application rather than memorization. Topics include optimization, simulation,
regression, classification, decision trees, model assessment, clustering, dimensionality
reduction, probability, and practical analytics decisions. Each question contains four
answer choices, one best answer, and a detailed rationale explaining why the correct
choice is appropriate and why the alternatives are less appropriate. These are original
study questions and are not reproduced from, or represented as, official Georgia Tech
examination materials.
CORE DOMAINS
1.
Linear and Nonlinear Optimization
2.
3.
Simulation and Monte Carlo Methods
4.
5.
Probability and Statistical Reasoning
6.
7.
Linear Regression
8.
9.
Logistic Regression
1
, 10.
11.
Classification
12.
13.
Decision Trees and Ensemble Methods
14.
15.
Model Evaluation and Validation
16.
17.
Clustering and Unsupervised Learning
18.
19.
Practical Analytics and Model Selection
20.
QUESTION 1
A company wants to minimize transportation cost while satisfying warehouse
capacity, customer demand, and nonnegative shipment requirements. Which modeling
framework is most appropriate?
A) Principal component analysis
B) Linear programming
C) K-means clustering
D) Logistic regression
Rationale: Linear programming is appropriate when the objective and constraints
can be expressed as linear functions of decision variables. PCA is for dimensionality
reduction, k-means is for clustering, and logistic regression models binary outcomes
rather than optimization problems.
QUESTION 2
In a linear optimization model, the objective function represents:
2
, A) The set of feasible constraints
B) The random error in the model
C) The quantity being maximized or minimized
D) The probability distribution of the decision variables
Rationale: The objective function specifies what the optimization procedure seeks to
maximize or minimize, such as profit, cost, or utilization. Constraints instead define
allowable solutions.
QUESTION 3
A feasible solution to an optimization problem is one that:
A) Always produces the global optimum
B) Has the largest objective value
C) Satisfies all model constraints
D) Contains no decision variables equal to zero
Rationale: Feasibility means satisfying every constraint. A feasible point does not
necessarily optimize the objective.
QUESTION 4
A linear program has an unbounded objective. This means:
A) No feasible solution exists
B) The objective can improve indefinitely within the feasible region
C) Every feasible solution is optimal
D) The constraints are necessarily nonlinear
Rationale: An unbounded maximization problem permits the objective to increase
without a finite upper bound, or an analogous situation occurs for minimization. This
differs from infeasibility, where no solution satisfies all constraints.
QUESTION 5
A manager must decide how many units of products A and B to manufacture. Product
A earns $30 profit and requires 2 labor hours; product B earns $50 and requires 4
labor hours. The factory has 100 labor hours. Which objective is appropriate?
A) Minimize 30A + 50B
B) Minimize 2A + 4B
C) Maximize 30A + 50B
D) Maximize 2A + 4B
Rationale: Since the goal is to maximize profit, the objective is 30A + 50B. Labor
consumption belongs in the constraint: 2A + 4B ≤ 100.
QUESTION 6
3
PRACTICE QUESTION BANK
QUESTIONS 1–100 WITH ANSWERS AND DETAILED
RATIONALES
INTRODUCTION
This practice bank is designed to provide rigorous preparation for ISYE 6501
Midterm 2–style assessment. The questions emphasize conceptual understanding,
quantitative reasoning, interpretation of analytical results, and selection of appropriate
statistical and machine-learning methods. Scenarios are intentionally constructed to
test application rather than memorization. Topics include optimization, simulation,
regression, classification, decision trees, model assessment, clustering, dimensionality
reduction, probability, and practical analytics decisions. Each question contains four
answer choices, one best answer, and a detailed rationale explaining why the correct
choice is appropriate and why the alternatives are less appropriate. These are original
study questions and are not reproduced from, or represented as, official Georgia Tech
examination materials.
CORE DOMAINS
1.
Linear and Nonlinear Optimization
2.
3.
Simulation and Monte Carlo Methods
4.
5.
Probability and Statistical Reasoning
6.
7.
Linear Regression
8.
9.
Logistic Regression
1
, 10.
11.
Classification
12.
13.
Decision Trees and Ensemble Methods
14.
15.
Model Evaluation and Validation
16.
17.
Clustering and Unsupervised Learning
18.
19.
Practical Analytics and Model Selection
20.
QUESTION 1
A company wants to minimize transportation cost while satisfying warehouse
capacity, customer demand, and nonnegative shipment requirements. Which modeling
framework is most appropriate?
A) Principal component analysis
B) Linear programming
C) K-means clustering
D) Logistic regression
Rationale: Linear programming is appropriate when the objective and constraints
can be expressed as linear functions of decision variables. PCA is for dimensionality
reduction, k-means is for clustering, and logistic regression models binary outcomes
rather than optimization problems.
QUESTION 2
In a linear optimization model, the objective function represents:
2
, A) The set of feasible constraints
B) The random error in the model
C) The quantity being maximized or minimized
D) The probability distribution of the decision variables
Rationale: The objective function specifies what the optimization procedure seeks to
maximize or minimize, such as profit, cost, or utilization. Constraints instead define
allowable solutions.
QUESTION 3
A feasible solution to an optimization problem is one that:
A) Always produces the global optimum
B) Has the largest objective value
C) Satisfies all model constraints
D) Contains no decision variables equal to zero
Rationale: Feasibility means satisfying every constraint. A feasible point does not
necessarily optimize the objective.
QUESTION 4
A linear program has an unbounded objective. This means:
A) No feasible solution exists
B) The objective can improve indefinitely within the feasible region
C) Every feasible solution is optimal
D) The constraints are necessarily nonlinear
Rationale: An unbounded maximization problem permits the objective to increase
without a finite upper bound, or an analogous situation occurs for minimization. This
differs from infeasibility, where no solution satisfies all constraints.
QUESTION 5
A manager must decide how many units of products A and B to manufacture. Product
A earns $30 profit and requires 2 labor hours; product B earns $50 and requires 4
labor hours. The factory has 100 labor hours. Which objective is appropriate?
A) Minimize 30A + 50B
B) Minimize 2A + 4B
C) Maximize 30A + 50B
D) Maximize 2A + 4B
Rationale: Since the goal is to maximize profit, the objective is 30A + 50B. Labor
consumption belongs in the constraint: 2A + 4B ≤ 100.
QUESTION 6
3
