Georgia Tech ISYE 6644 Simulation and Modeling for
Engineering and Science: Midterm Exam (Units 1-7) -
Questions with Correct Answers
100 questions
1. In a single-server queue, traffic intensity rho equals:
A. Arrival rate divided by service rate (lambda/mu)
B. Service rate divided by arrival rate
C. Arrival rate times service rate
D. Number in queue
Correct Answer: A. Arrival rate divided by service rate (lambda/mu)
Rationale: Must be below 1 for stability.
2. If rho is 1 or greater in a single-server queue:
A. The queue empties quickly
B. The queue grows without bound (unstable)
C. Waiting time is zero
D. The system is overstaffed
Correct Answer: B. The queue grows without bound (unstable)
Rationale: Need rho below 1.
3. A stochastic model is one that:
A. Includes random components such as random interarrival times
B. Has no randomness
C. Is always continuous
D. Is solved only with calculus
Correct Answer: A. Includes random components such as random interarrival times
Rationale: Opposite of deterministic.
4. The 'output' of a simulation should be treated as:
A. An exact constant
B. A deterministic function only
C. A random variable whose distribution must be estimated
D. Irrelevant
Correct Answer: C. A random variable whose distribution must be estimated
Rationale: Statistical analysis.
5. In simulation terminology, an entity is:
A. An object that moves through the system, such as a customer or part
B. A server that never moves
C. A random number
D. A statistical test
Correct Answer: A. An object that moves through the system, such as a customer or part
Rationale: Entities have attributes.
6. Because stochastic simulation output is random:
A. A single run gives the exact true mean
B. Results must be analyzed statistically, often with multiple replications
C. No confidence intervals are needed
D. Output never varies with the random number seed
Correct Answer: B. Results must be analyzed statistically, often with multiple replications
Rationale: Output analysis.
,7. A discrete-event simulation model changes state:
A. Continuously according to differential equations only
B. Never
C. At discrete points in time when events occur
D. At every nanosecond regardless of events
Correct Answer: C. At discrete points in time when events occur
Rationale: e.g., arrivals and departures.
8. Monte Carlo simulation primarily refers to:
A. Event-scheduling of queues
B. Using random sampling to estimate quantities like integrals or probabilities
C. Building physical prototypes
D. Deterministic optimization
Correct Answer: B. Using random sampling to estimate quantities like integrals or
probabilities
Rationale: Static simulation.
9. The 'GIGO' principle reminds analysts that:
A. Great input gives great output always
B. Poor input data lead to poor output (garbage in, garbage out)
C. Output does not depend on input
D. Arena removes all errors
Correct Answer: B. Poor input data lead to poor output (garbage in, garbage out)
Rationale: Input analysis matters.
10. A Turing test approach to validation involves:
A. Testing random numbers for independence
B. Having experts try to distinguish model output from real-system output
C. Fitting distributions
D. Computing batch means
Correct Answer: B. Having experts try to distinguish model output from real-system output
Rationale: Face validity.
11. A model in which the analyst steps through events by hand is called:
A. A continuous model
B. A hand simulation
C. An analytical queue
D. A Turing test
Correct Answer: B. A hand simulation
Rationale: Builds intuition.
12. Simulation optimization refers to:
A. Searching for the best input settings using simulation output
B. Making the code run faster only
C. Fitting a distribution
D. Testing the RNG
Correct Answer: A. Searching for the best input settings using simulation output
Rationale: e.g., OptQuest.
, 13. Why do we use different random number seeds across replications?
A. To make all runs identical
B. To speed up the computer
C. To obtain independent output observations
D. To remove the need for statistics
Correct Answer: C. To obtain independent output observations
Rationale: Independence enables classical CIs.
14. A key advantage of simulation is that it allows:
A. Guaranteed optimal solutions
B. Elimination of all modeling assumptions
C. Results with zero variability
D. Testing 'what if' scenarios without disrupting the real system
Correct Answer: D. Testing 'what if' scenarios without disrupting the real system
Rationale: Safe experimentation.
15. Simulation 'run length' refers to:
A. The CPU time of the computer
B. The number of lines of code
C. The period of the RNG
D. The amount of simulated time (or number of events) in a single run
Correct Answer: D. The amount of simulated time (or number of events) in a single run
Rationale: Run control.
16. Integration by parts states that the integral of u dv equals:
A. uv minus the integral of v du
B. uv plus the integral of v du
C. u'v'
D. The integral of u times the integral of v
Correct Answer: A. uv minus the integral of v du
Rationale: Used for x e^(-x) integrals.
17. The integral of x^2 from 0 to 1 is:
A. 1/2
B. 1/3
C. 1
D. 2/3
Correct Answer: B. 1/3
Rationale: x^3/3 evaluated 0 to 1.
18. The limit of (1 + 1/n)^n as n goes to infinity is:
A. 1
B. Infinity
C. 0
D. e
Correct Answer: D. e
Rationale: Definition of e.
Engineering and Science: Midterm Exam (Units 1-7) -
Questions with Correct Answers
100 questions
1. In a single-server queue, traffic intensity rho equals:
A. Arrival rate divided by service rate (lambda/mu)
B. Service rate divided by arrival rate
C. Arrival rate times service rate
D. Number in queue
Correct Answer: A. Arrival rate divided by service rate (lambda/mu)
Rationale: Must be below 1 for stability.
2. If rho is 1 or greater in a single-server queue:
A. The queue empties quickly
B. The queue grows without bound (unstable)
C. Waiting time is zero
D. The system is overstaffed
Correct Answer: B. The queue grows without bound (unstable)
Rationale: Need rho below 1.
3. A stochastic model is one that:
A. Includes random components such as random interarrival times
B. Has no randomness
C. Is always continuous
D. Is solved only with calculus
Correct Answer: A. Includes random components such as random interarrival times
Rationale: Opposite of deterministic.
4. The 'output' of a simulation should be treated as:
A. An exact constant
B. A deterministic function only
C. A random variable whose distribution must be estimated
D. Irrelevant
Correct Answer: C. A random variable whose distribution must be estimated
Rationale: Statistical analysis.
5. In simulation terminology, an entity is:
A. An object that moves through the system, such as a customer or part
B. A server that never moves
C. A random number
D. A statistical test
Correct Answer: A. An object that moves through the system, such as a customer or part
Rationale: Entities have attributes.
6. Because stochastic simulation output is random:
A. A single run gives the exact true mean
B. Results must be analyzed statistically, often with multiple replications
C. No confidence intervals are needed
D. Output never varies with the random number seed
Correct Answer: B. Results must be analyzed statistically, often with multiple replications
Rationale: Output analysis.
,7. A discrete-event simulation model changes state:
A. Continuously according to differential equations only
B. Never
C. At discrete points in time when events occur
D. At every nanosecond regardless of events
Correct Answer: C. At discrete points in time when events occur
Rationale: e.g., arrivals and departures.
8. Monte Carlo simulation primarily refers to:
A. Event-scheduling of queues
B. Using random sampling to estimate quantities like integrals or probabilities
C. Building physical prototypes
D. Deterministic optimization
Correct Answer: B. Using random sampling to estimate quantities like integrals or
probabilities
Rationale: Static simulation.
9. The 'GIGO' principle reminds analysts that:
A. Great input gives great output always
B. Poor input data lead to poor output (garbage in, garbage out)
C. Output does not depend on input
D. Arena removes all errors
Correct Answer: B. Poor input data lead to poor output (garbage in, garbage out)
Rationale: Input analysis matters.
10. A Turing test approach to validation involves:
A. Testing random numbers for independence
B. Having experts try to distinguish model output from real-system output
C. Fitting distributions
D. Computing batch means
Correct Answer: B. Having experts try to distinguish model output from real-system output
Rationale: Face validity.
11. A model in which the analyst steps through events by hand is called:
A. A continuous model
B. A hand simulation
C. An analytical queue
D. A Turing test
Correct Answer: B. A hand simulation
Rationale: Builds intuition.
12. Simulation optimization refers to:
A. Searching for the best input settings using simulation output
B. Making the code run faster only
C. Fitting a distribution
D. Testing the RNG
Correct Answer: A. Searching for the best input settings using simulation output
Rationale: e.g., OptQuest.
, 13. Why do we use different random number seeds across replications?
A. To make all runs identical
B. To speed up the computer
C. To obtain independent output observations
D. To remove the need for statistics
Correct Answer: C. To obtain independent output observations
Rationale: Independence enables classical CIs.
14. A key advantage of simulation is that it allows:
A. Guaranteed optimal solutions
B. Elimination of all modeling assumptions
C. Results with zero variability
D. Testing 'what if' scenarios without disrupting the real system
Correct Answer: D. Testing 'what if' scenarios without disrupting the real system
Rationale: Safe experimentation.
15. Simulation 'run length' refers to:
A. The CPU time of the computer
B. The number of lines of code
C. The period of the RNG
D. The amount of simulated time (or number of events) in a single run
Correct Answer: D. The amount of simulated time (or number of events) in a single run
Rationale: Run control.
16. Integration by parts states that the integral of u dv equals:
A. uv minus the integral of v du
B. uv plus the integral of v du
C. u'v'
D. The integral of u times the integral of v
Correct Answer: A. uv minus the integral of v du
Rationale: Used for x e^(-x) integrals.
17. The integral of x^2 from 0 to 1 is:
A. 1/2
B. 1/3
C. 1
D. 2/3
Correct Answer: B. 1/3
Rationale: x^3/3 evaluated 0 to 1.
18. The limit of (1 + 1/n)^n as n goes to infinity is:
A. 1
B. Infinity
C. 0
D. e
Correct Answer: D. e
Rationale: Definition of e.
