Simulation and Modeling

Assignment – Short Answer Questions

Q1. Define simulation. What are its advantages and disadvantages?

Simulation is the process of designing a model of a real system and conducting experiments on that
model, in order to understand how the real system behaves or to evaluate different strategies for its
operation. Instead of experimenting on the actual system (which may be costly, slow, or risky), we build
a computer model that imitates the system's behaviour over time and study its outputs.

Example: A bank builds a computer model of its teller counters to see how customer waiting time
changes if one more teller is added, without disturbing the real bank operations.

Advantages of Simulation
• Allows study of complex systems that are difficult to analyse mathematically.
• Time can be compressed or expanded — years of operation can be studied in minutes.
• Safer and cheaper than experimenting with a real system (e.g., testing a nuclear plant).
• Different policies/designs can be compared before implementation.
• Helps in understanding why certain results occur (diagnostic capability).

Disadvantages of Simulation
• Model building requires special training/skills and can be time-consuming.
• Simulation results are usually only estimates; large runs may be needed for accuracy.
• Can be expensive in terms of time and computing resources for complex systems.
• There is a risk of using simulation when an analytical (exact) solution already exists and would be
simpler.
• Poorly validated models can give misleading results if not checked carefully.

Q2. What is a model? Classify different types of models with examples.

A model is a simplified representation of a real system, object, or process that captures the features
relevant to the study while ignoring unnecessary details. Models help us analyse, predict, or understand
the behaviour of a system without dealing directly with the real system.

Classification of Models
• Physical (Iconic) Models — scaled physical replicas, e.g., a toy aeroplane model, a globe.
• Mathematical (Symbolic) Models — use mathematical/logical relationships, e.g., an equation for
projectile motion.
• Static Models — represent a system at one particular point in time (time does not play a role), e.g.,
a Monte Carlo model of a dice game.
• Dynamic Models — represent a system as it changes/evolves over time, e.g., a queueing model of
a bank over an 8-hour day.
• Deterministic Models — contain no random components; output is fully predictable, e.g., a simple
interest calculation.
• Stochastic (Probabilistic) Models — contain one or more random inputs, so output is itself
random, e.g., customer arrival times at a shop.
• Discrete Models — state variables change only at separate points in time, e.g., number of
customers in a queue.
• Continuous Models — state variables change continuously over time, e.g., water level in a dam.

Q3. Explain discrete-event simulation and continuous simulation.

Discrete-Event Simulation (DES)
In discrete-event simulation, the state of the system changes only at specific points in time, called
events (e.g., a customer arrival, a service completion). Between two consecutive events, nothing
changes, so the simulation clock can jump directly from one event to the next. DES is widely used for
queueing systems, manufacturing lines, and communication networks.

Continuous Simulation
In continuous simulation, the state of the system changes continuously over time, usually described
using differential equations. The simulation clock advances in small, fixed time steps, and the state
variables are updated at every step. It is commonly used for physical/engineering systems such as fluid
flow, population growth, or chemical reactions.

Fig. 1: State change pattern — discrete-event (step changes) vs continuous simulation (smooth curve)

Q4. What is the Monte Carlo method? Where is it used?

The Monte Carlo method is a simulation technique that uses repeated random sampling to obtain
numerical results for problems that may be deterministic in principle but are too complex to solve
analytically. It relies on generating a large number of random numbers and using them, along with
probability distributions, to estimate the value of an unknown quantity.

Example: Estimating the value of π by randomly throwing points into a square containing a circle and
checking the fraction of points that fall inside the circle.

Applications of Monte Carlo Method

• Estimating areas, volumes, and integrals that are hard to solve directly.
• Risk analysis in finance and project management (e.g., estimating project completion time).
• Reliability and inventory analysis.
• Physics simulations, such as particle transport and nuclear reactor design.
• Queueing systems and other stochastic operations research problems.
 Q5. Define random number generation and its importance in simulation.
Random number generation is the process of producing a sequence of numbers that appear to be
statistically independent and uniformly distributed between 0 and 1. In practice, computers use
algorithms called pseudo-random number generators (PRNGs) (e.g., the Linear Congruential
Method) that produce numbers which behave like true random numbers but are actually generated by a
deterministic formula, so the sequence can be reproduced if needed for testing.

Importance in Simulation
• Random numbers are the basic building blocks used to generate random variates (e.g., customer
arrival times, service times) from any required probability distribution.
• They introduce the randomness needed to realistically represent stochastic systems.
• Good quality (statistically sound) random numbers ensure the simulation results are valid and
unbiased.
• They allow experiments to be repeated under identical random conditions for comparing different
system designs fairly.

Q6. What are input models in simulation? Give examples.

An input model defines the probability distributions and parameters that describe the random inputs
driving a simulation. Since a simulation is only as good as its inputs, input modelling involves collecting
real data, choosing a suitable theoretical distribution that fits the data, and estimating its parameters.

Examples of Input Models

• Customer inter-arrival times at a shop — often modelled using an Exponential distribution.
• Service times at a counter or machine — often modelled using Normal, Exponential, or Gamma
distributions.
• Time between machine breakdowns — modelled using Exponential or Weibull distribution.
• Number of defective items in a batch — modelled using Binomial or Poisson distribution.

Q7. What is warm-up period in simulation?

The warm-up period (also called the transient or initialisation period) is the initial phase of a simulation
run during which the system has not yet reached its normal, representative operating condition — for
example, a simulated shop that starts completely empty does not reflect typical busy-hour behaviour.
Statistics collected during this period are biased and are usually discarded, and data collection for
performance measures begins only after the system has warmed up, so that the results reflect steady,
representative behaviour rather than the unrealistic starting condition.

Q8. Explain the concept of steady state in simulations.

A system is said to be in steady state when its statistical behaviour (such as average waiting time,
average queue length, or utilisation) becomes stable and no longer depends on the starting conditions or
on time — i.e., the probability distribution of the system's state stops changing as the simulation
continues to run. Before reaching steady state, a system passes through a transient (warm-up) phase
where its behaviour is still influenced by the initial conditions.
Steady-state analysis is important when we want long-run performance measures of a system (e.g., the
long-run average number of customers in a queue), rather than measures that are only valid for a
specific short time window.

Q9. Differentiate between deterministic and stochastic models.

Basis Deterministic Model Stochastic Model

Contains one or more random

Randomness Contains no random variables
variables

Same output every time for the Output can differ each run, even with
Output
same input the same input

Predictability Fully predictable Involves uncertainty/probability

Simple interest calculation, a fixed Customer arrivals at a bank,

Example
production schedule machine breakdown times

Q10. Describe verification and validation in simulation studies.

Verification
Verification asks: “Have we built the model right?” It is the process of checking that the simulation
model has been correctly translated into a computer program — i.e., that the code is free of logical and
programming errors and behaves the way the model design intended. Techniques include structured
walkthroughs, debugging, tracing program execution, and checking outputs against simple
hand-calculated cases.

Validation
Validation asks: “Have we built the right model?” It is the process of checking that the simulation model
is an accurate representation of the real system for the purposes of the study. This is done by comparing
the model's output data with real system data (historical records), consulting domain experts, and
performing sensitivity analysis to see if the model reacts to changes the way the real system would.

In short, verification focuses on the correctness of the model's implementation, while validation focuses
on the correctness/realism of the model's representation of reality. Both steps are essential before the
simulation results can be trusted for decision-making.

Q11. Define GPSS.

GPSS (General Purpose Simulation System) is a high-level, block-structured simulation language,
originally developed by Geoffrey Gordon at IBM in 1961, that is used for building discrete-event
simulation models — especially of queueing and transaction-flow systems such as job-shops, computer
networks, and communication systems. In GPSS, a model is built by connecting predefined blocks
(such as GENERATE, QUEUE, SEIZE, ADVANCE, RELEASE, TERMINATE) that represent the logical
flow of transactions (entities, such as customers or jobs) moving through the system.

Q12. Describe any two GPSS blocks with diagram.

1. GENERATE Block
The GENERATE block creates transactions (entities) and introduces them into the model at defined time
intervals. It is always the first block of a transaction's journey through the model and typically specifies
the mean inter-arrival time, spread, and other timing parameters. For example, GENERATE 5,2 creates
a new transaction on average every 5 time units, with a spread of ±2 units.

2. TERMINATE Block
The TERMINATE block removes a transaction from the model once it has completed its journey through
the system, and optionally reduces a termination counter that is used to control how long the simulation
run continues. For example, TERMINATE 1 removes the transaction and decreases the termination
count by 1.

Fig. 2: Flow of a transaction through GENERATE and TERMINATE blocks in a GPSS model