Unit 1: Introduction to Simulation
1. Difference between static physical and dynamic physical models.
Solutoin
2. Describe different phases of simulation study with help of flowchart.
Solution
Phases of Simulation Study
A simulation study is carried out in four main phases consisting of several steps.
1. Phase I: Problem Formulation (Discovery Phase)
Define the problem clearly
Set objectives and project plan
Decide whether simulation is appropriate
2. Phase II: Model Building and Data Collection
Model conceptualization (abstract system)
Data collection
Model translation (coding)
Verification (check correctness of program)
Validation (check model with real system)
3. Phase III: Running the Model
Experimental design
Production runs
Output analysis
More runs if required
4. Phase IV: Implementation
Documentation and reporting
Implementation of results
3. What are the application areas of simulation?
Solution
Simulation is the process of designing a model of a real system and conducting experiments with it for the purpose of analysis, design, optimization, and training.
Application Areas of Simulation
Industrial and Business: Used for manufacturing system design, inventory control, supply chain management, and business decision-making.
Healthcare: Used for hospital management, patient flow analysis, medical facility planning, and disease/epidemic modeling.
Transportation: Used for traffic control, road network design, railway/airline scheduling, and airport logistics planning.
Engineering: Used in civil engineering for bridge, building, and earthquake simulation, as well as system design and performance testing.
Computer and Communication Systems: Used for network design, performance evaluation, operating system testing, and internet communication systems.
Military and Defense: Used for war strategy simulation, mission planning, operator training, and weapon system testing.
Environmental: Used for pollution control analysis, waste management systems, and environmental impact studies.
Training and Education: Used in flight simulators, operator training systems, and virtual learning environments to train personnel safely.
Scientific Research: Used for biological system modeling, population studies, physics simulations, and queuing system analysis.
4. What is a system modeling? What are the steps involved in system modeling.
Solution
System Modeling
System modeling is the process of creating a simplified representation of a real-world system to understand its behavior and predict its performance under different conditions.
Steps Involved in System Modeling
1. Phase I: Problem Formulation (Discovery Phase)
Define the problem clearly
Set objectives and project plan
Decide whether simulation is appropriate
2. Phase II: Model Building and Data Collection
Model conceptualization (abstract system)
Data collection
Model translation (coding)
Verification (check correctness of program)
Validation (check model with real system)
3. Phase III: Running the Model
Experimental design
Production runs
Output analysis
More runs if required
4. Phase IV: Implementation
Documentation and reporting
Implementation of results
5. Why model of a system is built? What is static model? Differentiate between static and dynamic mathematical models in simulation.
solution
A model of a system is built to study and analyze a system without using the real system.
Reasons:
To analyze systems before implementation and reduce design errors
To save cost and time compared to real experimentation
To study systems that are complex or not directly observable
To predict the effect of changes without disturbing the real system
To provide training environments (e.g., simulators)
To understand relationships between system components
Static Model
A static model is a model that represents a system at a particular point in time and does not consider changes over time.
Features:
Time-independent
Shows system in equilibrium/state
Simple and less flexible
Does not show system behavior over time
Example:
Architectural model of a building
Snapshot of a market at a fixed price
6. Differentiate between discrete and continuous system.
Solution
7. Discuss the merits and demerits of system simulation.
Solution
Merits (Advantages)
Allows safe experimentation without affecting real system
Saves cost and time in system design and testing
Helps in understanding complex systems
Useful for “what-if” analysis
Can identify bottlenecks and inefficiencies
Time can be compressed or expanded
Suitable for dangerous or real-time systems
Helps in decision making and planning
Demerits (Disadvantages)
Requires skilled personnel and training
Model building is time-consuming and costly
Results may be difficult to interpret
Output may be affected by randomness
Models may not be accurate representations
Not suitable when analytical solution is easier
Requires large amount of data
8. What do you understand by static mathematical model? Explain with example. Differentiate between stochastic and deterministic activities.
Solution
A static mathematical model is a model that represents a system using mathematical equations at a specific point in time, without considering any changes over time.
It is time-independent (no past or future behavior).
Describes the system in a state of equilibrium.
Does not show how variables change, only their final relationship.
Focuses on structure rather than behavior.
9. Describe the phases in simulation.
Solution
Phases in Simulation
The simulation process is divided into four main phases:
1. Phase I: Problem Formulation (Discovery Phase)
Define the problem clearly
Set objectives and project plan
Decide whether simulation is appropriate
2. Phase II: Model Building and Data Collection
Model conceptualization (abstract system)
Data collection
Model translation (coding)
Verification (check correctness of program)
Validation (check model with real system)
3. Phase III: Running the Model
Experimental design
Production runs
Output analysis
More runs if required
4. Phase IV: Implementation
Documentation and reporting
Implementation of results
Conclusion
Simulation is an iterative process, meaning steps may be repeated until satisfactory results are obtained.
10. Write short notes on:
a. System and its environment
System:
A system is a collection of interrelated components that work together to achieve a common objective.
Components of a System:
Entities: Objects of interest in the system
Attributes: Properties of entities
Activities: Processes that cause changes in the system
State of System:
The state is the condition of the system at a specific time, described by its variables.
b. System Environment
The system environment consists of all external factors that affect the system but are not part of it.
A boundary separates the system from its environment
The environment influences system behavior
Types of Activities:
Endogenous: Occur inside the system
Exogenous: Occur outside but affect the system
Types of Systems:
Open System: Interacts with environment
Closed System: No interaction (rare in real life)
Unit 2: Simulation of Continuous and Discrete System
1. What is analog computer? Explain with suitable example.
Solution
Analog Computer
An analog computer is a computing device that uses continuous physical quantities (such as voltage, current, or mechanical motion) to represent and solve mathematical problems.
-Works with continuous data
-Used to solve differential equations
-Provides approximate results
-Faster for real-time simulation of physical systems
Example: Electrical Circuit Analogy
An electrical circuit can be used to simulate a mechanical system.
Components like:
Resistance (R) → represents damping
Inductance (L) → represents mass
Capacitance (C) → represents elasticity
By adjusting these components, the behavior of a real system can be studied.
2. Differentiate between analog and digital computer.
Solution
3. Why accuracy of analog computer is low? Explain analog computer with suitable example.
Solution
Analog Computer:-
An analog computer is a computing device that uses continuous physical quantities (such as voltage, current, or mechanical motion) to represent and solve mathematical problems.
Accuracy of Analog Computer is Low:
Limited Precision in Measurement:
Physical quantities like voltage cannot be measured exactly, leading to errors.
Noise and Disturbances:
Electrical noise affects signals and reduces accuracy.
Component Imperfections:
Devices like resistors and capacitors have tolerances and are not ideal.
Approximate Nature:
Analog computers give approximate results, not exact values.
Example: Electrical Circuit Analogy
An electrical circuit can be used to simulate a mechanical system.
Components like:
Resistance (R) → represents damping
Inductance (L) → represents mass
Capacitance (C) → represents elasticity
By adjusting these components, the behavior of a real system can be studied.
4. Explain Monte Carlo simulation.
Solution
Monte Carlo Simulation:
Monte Carlo simulation is a technique that uses random numbers and repeated sampling to simulate and analyze the behavior of a system with uncertainty.
It is based on probability and randomness
Used when systems are too complex for analytical solutions
Produces approximate results through multiple trials
Steps in Monte Carlo Simulation:
Define the problem and model
Identify probability distributions of input variables
Generate random numbers
Perform simulation trials
Analyze the results
Example:
Suppose we want to estimate the probability of getting heads in coin tosses.
We simulate many random coin tosses using random numbers.
By repeating this many times, we estimate the probability.
5. Explain the concept of discrete event simulation. Explain poisson’s arrival pattern.
Solution
Discrete Event Simulation:
Discrete Event Simulation (DES) is a simulation technique in which the state of a system changes only at specific discrete points in time due to the occurrence of events.
State changes occur only at event times
Between events, the system state remains constant
Events include arrival, departure, service completion, etc.
Widely used in systems like banks, queues, networks
Example:
In a bank system, the number of customers changes only when:
A customer arrives
A customer leaves after service
Poisson Arrival Pattern:
A Poisson arrival pattern describes a process where arrivals occur randomly and independently over time.
Key Characteristics:
Arrivals occur one at a time
Arrivals are independent
Average arrival rate (λ) is constant
Used to model random events like customer arrivals
6. Write short notes on:
a. Feedback system
b. Non-Stationary Poisson process
c. Digital analog simulator
a. Feedback System
A feedback system is a system in which the output is fed back into the input to control or regulate the system.
Types:
Positive Feedback: Enhances or increases output
Negative Feedback: Reduces or stabilizes output
Example:
Thermostat in AC: maintains room temperature by using feedback
b. Non-Stationary Poisson Process
A Non-Stationary Poisson Process is a process in which the arrival rate (λ) changes over time.
Key Points:
Arrival rate is not constant
Depends on time (rush hours, peak periods, etc.)
Violates stationary increment property
Used in systems with varying traffic like networks or customer flow
c. Digital Analog Simulator
A digital-analog simulator is a digital system or software that simulates the behavior of an analog computer to solve continuous mathematical problems.
Key Features:
Uses digital computers to simulate continuous systems
Performs operations like integration, addition, and differentiation
More accurate and flexible than physical analog computers
Example:
Simulation of a mechanical or electrical system using software like simulation tools (e.g., MATLAB/Simulink)
Unit 3: Queuing System
1.Write short notes:
Queuing discipline
Random variate
Queuing Discipline:
Queuing discipline is the rule used to decide the order of serving customers/jobs in a queue.
It determines which task is processed first in a waiting line.
It is important for system performance and fairness.
FCFS (First Come First Served): Jobs are served in the order of arrival.
LCFS (Last Come First Served): The most recent job is served first.
Priority Scheduling: Jobs are served based on priority level.
Round Robin: Each job gets a fixed time slice in rotation.
Shortest Job First (SJF): Job with the smallest execution time is served first.
Example: In a bank, customers are usually served using FCFS discipline.
Conclusion: Queuing discipline ensures efficient and fair processing of tasks in a system.
Random Variate
A random variate is a numerical value generated from a probability distribution.
It represents the outcome of a random process.
It is widely used in simulation and modeling.
It can be discrete (e.g., number of customers).
It can be continuous (e.g., time between arrivals).
Generated using random number generators and transformation methods.
Common distributions: Uniform, Normal, Exponential, Poisson.
Example: Generating inter-arrival time of customers using an exponential distribution.
Conclusion: Random variates help in simulating real-world randomness in systems.
2.Define traffic intensity and server utilization. Write down the Kendall’s notation for a queuing system with an example.
Solution
Traffic Intensity:
Traffic intensity is the measure of load on a queuing system.
It is the ratio of arrival rate to service capacity.
It indicates how busy the system is.
Represented as ρ (rho).
Formula: ρ = λ / (μ × s)
, where λ = arrival rate, μ = service rate, s = number of servers.
If ρ < 1, system is stable.
If ρ ≥ 1, system becomes overloaded and queues grow.
Example: If λ = 4 customers/min, μ = 5 customers/min, s = 1 → ρ = 0.8.
Conclusion: Traffic intensity shows system load and stability condition.
Server Utilization:
Server utilization is the fraction of time a server is busy serving customers.
It reflects the efficiency of server usage.
It is closely related to traffic intensity.
Represented as ρ (rho) in single-server systems.
Value ranges from 0 to 1.
High utilization means server is busy most of the time.
Low utilization means server is idle often.
Example: If server is busy 80% of the time, utilization = 0.8.
Conclusion: Server utilization indicates how effectively resources are used.
Kendall’s Notation:
Kendall’s notation is a standard format to describe a queuing system.
It represents arrival, service, and system characteristics.
General form: A / B / s : (N / D)
A: Arrival distribution (e.g., M = Markovian/Poisson).
B: Service time distribution (e.g., M = Markovian,exponential).
s: Number of servers.
N: Population size .
D: Queue discipline.
Common symbols:
M: Markovian (Poisson/exponential)
D: Deterministic
G: General distribution
Example: M/M/1
Poisson arrivals, exponential service time, 1 server.
Conclusion: Kendall’s notation provides a compact way to describe queuing models.
3.What is a queuing system? Explain the elements of queuing system with example.
Solution
Queuing System
A queuing system is a model used to represent waiting lines in real-life systems.
It consists of customers, queue, and service mechanism.
It helps to analyze delays and system performance.
Used in banks, hospitals, networks, call centers.
Example: Customers waiting in a bank queue for service.
Conclusion: A queuing system models waiting situations to improve efficiency.
Elements of Queuing System are
Calling Population (Source):
Set of customers that arrive in the system.
Can be finite or infinite.
Example: People visiting a bank.
Arrival Process:
Describes how customers arrive over time.
Can follow Poisson distribution or other patterns.
Example: Customers arriving randomly at a counter.
Queue (Waiting Line):
Place where customers wait before service.
Can be single queue or multiple queues.
Example: Line in front of a ticket counter.
Service Mechanism:
Describes how service is provided.
Includes number of servers and service rate.
Example: One or more clerks serving customers.
Queuing Discipline:
Rule for serving customers.
Types: FCFS, Priority, SJF, Round Robin.
Example: Bank follows First Come First Served.
System Capacity:
Maximum number of customers the system can hold.
Can be limited or unlimited.
Example: Limited seats in a waiting room.
Conclusion: Elements of a queuing system define how customers arrive, wait, and get served, affecting overall performance.
3. What is the calling population? Explain the arrival and service process in a queue.
Solution
Calling Population:
Calling population is the source of customers that enter a queuing system.
It represents the total number of potential customers.
It affects the arrival pattern and system behavior.
Can be finite (limited customers).
Can be infinite (very large, practically unlimited).
Finite population affects arrival rate over time.
Infinite population assumes constant arrival rate.
Example:
Finite: Machines in a repair system.
Infinite: Customers arriving at a bank.
Conclusion: Calling population defines the source and size of incoming customers.
Arrival Process:
Arrival process describes how customers enter the queue.
It defines the arrival rate and pattern over time.
It is important for queue formation and waiting time.
Represented by arrival rate (λ).
Can follow Poisson distribution (random arrivals).
Inter-arrival time may be exponential.
Arrivals can be single or in batches.
Example: Customers arriving randomly at a service counter.
Service Process:
Service process describes how customers are served.
It defines the service rate and service time.
It impacts the speed of queue clearance.
Represented by service rate (μ).
Service time may follow exponential or deterministic distribution.
Can have single or multiple servers.
Service can be uniform or variable.
Example: A cashier serving customers at a fixed or variable rate.
Conclusion: Arrival and service processes determine queue length, waiting time, and system efficiency.
4. Explain basic characteristics of a Queueing System
Solution
A queuing system is defined by features that describe how customers arrive, wait, and are served.
These characteristics determine system behavior and performance.
They are essential for analyzing waiting time and efficiency.
The basic characteristics of a Queueing System:
Calling Population — The source of customers entering the system, which can be finite or infinite.
Arrival Process — The pattern of arrivals is defined by the arrival rate (λ) and a distribution such as Poisson.
Service Process — The manner in which service is provided, defined by service rate (μ) and service time distribution.
Number of Servers (Channels) — The number of service points available, either single-server or multi-server.
Queue Capacity — The maximum number of customers allowed in the system, either limited or unlimited.
Queuing Discipline — The rule used for serving customers, such as FCFS, Priority, SJF, or Round Robin.
System Structure — The arrangement of queues and servers, such as single queue-single server or single queue-multiple server.
5. Explain the Kendall’s notation for a queuing system? What are the various performance measures in a single-server queuing System? Explain which of them determines system stability and how?
Soluiton
Kendall’s Notation:
Kendall’s notation is a standard format to describe a queuing system.
It represents arrival, service, and system characteristics.
General form: A / B / s : (N / D)
A: Arrival distribution (e.g., M = Markovian/Poisson).
B: Service time distribution (e.g., M = Markovian,exponential).
s: Number of servers.
N: Population size .
D: Queue discipline.
Common symbols:
M: Markovian (Poisson/exponential)
D: Deterministic
G: General distribution
Example: M/M/1
Poisson arrivals, exponential service time, 1 server.
Conclusion: Kendall’s notation provides a compact way to describe queuing models.
Performance Measures in a Single-Server Queuing System:
Performance measures evaluate the efficiency and behavior of a queue.
They help in analyzing waiting time, queue length, and utilization.
Average Number of Customers in the System (L) — Total customers present in the system, including both those waiting and the one being served.
Average Number of Customers in the Queue (Lq) — Number of customers waiting in line, excluding the one currently being served.
Average Time Spent in the System (W) — Total time a customer spends from arrival to departure, including both waiting and service time.
Average Waiting Time in the Queue (Wq) — Time a customer spends only in the queue before service begins.
Server Utilization (ρ) — Proportion of time the server remains busy, calculated as ρ = λ/μ.
Throughput — Rate at which customers are successfully served and discharged from the system.
Stability Measure — System Utilization (ρ):
System Utilization (ρ) is the primary measure that determines whether a single-server queuing system is stable or not.
How it Determines Stability:
Stable System (ρ < 1) — When service rate (μ) exceeds arrival rate (λ), the server clears tasks faster than they arrive, the system reaches a steady state, and all measures like L and Lq remain finite.
Unstable System (ρ ≥ 1) — When arrival rate equals or exceeds service rate, the server cannot keep up, causing queue length (Lq) and waiting time (Wq) to grow infinitely.
Critical Point (ρ → 1) — As utilization approaches 100%, even minor fluctuations in arrivals create a permanent backlog that the server cannot eliminate, causing wait times to rise exponentially.
IN simple term
Stable System (ρ < 1) : Imagine a cashier at a shop who serves customers faster than they arrive. The line never gets too long because the cashier always has time to clear it. Everything runs smoothly.
Unstable System (ρ ≥ 1) : Now imagine more customers are arriving than the cashier can handle. The line keeps growing longer and longer with no end. The cashier can never catch up, so the queue becomes endless.
Critical Point (ρ → 1) : This is like the cashier being just barely keeping up. Even one extra customer arriving at the wrong moment creates a small pile-up, and since the cashier has zero free time, that pile-up never gets cleared — it just keeps building slowly
6. List down the applications of the queuing system?
Solution
Applications of Queuing System:
Banking Systems:- Used to manage customer waiting lines at counters and ATMs.
Hospital Management:- Applied to handle patient flow in OPD, emergency wards, and appointments.
Telecommunication Systems:- Used to manage call traffic and reduce network congestion.
Computer Systems:- Applied in CPU scheduling, process queues, and job management.
Transportation Systems:- Used for traffic control at toll booths, airports, and railway stations.
Customer Service Centers:- Applied to manage incoming calls and reduce waiting time in help desks.
Manufacturing Systems:- Used to handle production lines and schedule machine servicing.
Retail and Supermarkets:- Applied to manage billing counters and checkout queues efficiently.
7. Explain traffic intensity and server utilization.
Solution
Traffic Intensity:
Traffic intensity (ρ) is the measure of load on a queuing system.
It is the ratio of arrival rate to service rate.
It shows how busy the system is.
Formula: ρ = λ / μ (for single server)
If ρ < 1, system is stable.
If ρ ≥ 1, system is unstable.
Higher ρ means longer queues and delays.
Example: λ = 3, μ = 5 → ρ = 0.6 (stable system)
Conclusion: Traffic intensity determines system load and stability.
Server Utilization:
Server utilization is the fraction of time a server is busy.
It indicates efficiency of server usage.
In single-server systems, it is equal to traffic intensity (ρ).
Value ranges from 0 to 1.
High utilization → server is mostly busy.
Low utilization → server is often idle.
Helps in capacity planning.
Example: If server is busy 70% of the time → utilization = 0.7.
Conclusion: Server utilization shows how effectively the server is used.
8. Explain the arrival pattern and the service process.
Solution
Arrival Pattern:
Arrival pattern describes how customers arrive in the system.
It defines the timing and distribution of arrivals.
It affects queue formation.
Represented by arrival rate (λ).
Often follows Poisson distribution.
Inter-arrival time may be exponential.
Can be single or batch arrivals.
Example: Customers arriving randomly at a bank.
Service Process:
The service process describes how customers are served.
It defines the service time and service rate.
It affects waiting time and system performance.
Represented by service rate (μ).
Service time may follow an exponential or deterministic distribution.
Can have single or multiple servers.
Service may be uniform or variable.
Example: A cashier serving customers at a certain rate.
Conclusion: Arrival pattern and service process together determine queue behavior and efficiency.
Unit 4: Markov Chains
1. Explain a Markov chain with a suitable example. What are the different application areas of Markov chains?
Solution
Markov Chain:
A Markov Chain is a mathematical model describing a sequence of events where the probability of the next event depends only on the current state, not on how it got there.
This is called the "memoryless property."
Example — Weather Prediction:
Assume the weather has three states: Sunny, Cloudy, Rainy.
If Sunny today → 70% chance Sunny, 20% Cloudy, 10% Rainy tomorrow.
If Rainy today → 50% chance Rainy, 50% Cloudy tomorrow.
These probabilities are stored in a Transition Matrix, which allows us to predict weather several days ahead based only on today's condition — not last week's.
Application Areas of Markov Chain:
Search Engines (Google PageRank):— Models a random surfer clicking links; the probability of landing on a page determines its rank and importance.
Text Prediction & NLP:— Autocomplete on phones predicts the next word based only on the current word typed using frequency patterns.
Finance & Economics:— Models stock market trends (Bull/Bear/Stagnant) and estimates loan default risk based on a customer's current credit status.
Genetics & Biology:— Models DNA sequences where the next base pair depends on the current one, helping identify genes and predict protein structures.
Queueing Theory:— Calculates waiting times and service efficiency by modeling how customers arrive and move through a service system.
Example Numerical
Unit 5: Random Numbers
1. Explain the independence and uniformity property of a random number. For the following sample of random numbers, perform a test for independence using the K-S test. (D0.05,10 = 0.41)
0.35, 0.77, 0.12, 0.33, 0.88, 0.45, 0.19, 0.25, 0.91, 0.54
Solution
Independence Property:
Independence means each random number is not related to others.
The value of one number does not affect the next.
Ensures no pattern or correlation in the sequence.
Important for valid simulation results.
Tested using runs test, autocorrelation, K–S test.
Uniformity Property:
Uniformity means numbers are evenly distributed in the interval (0,1).
Each value has equal probability of occurring.
Ensures fair representation of randomness.
Tested using K–S test, Chi-square test.
2. Give the concept of the Mid-square method. Calculate the first five random numbers using the mid-square method for seed 117.
Solution
Mid Square Method:
This is one of the earliest methods used for generating pseudo-random numbers.
It is worked by:
a. Taking a seed value: Typically, a number with an even number of digits(like 4).
b. Square the seed value
c. Extracting the middle digits of the result to the same length as the seed number.
d. Use these middle digits as the next seed.
e. Normalizing it (dividing by the largest possible value like 999 or 1000 to get a number between 0 and 1.
3. What are the properties of a random number? The sequence of numbers 0.54, 0.73, 0.98, 0.11, and 0.68 has been generated. Use the Kolmogorov-Smirnov test with a level of significance (α) 0.05 to determine if the hypothesis that the numbers are uniformly distributed on the interval 0 to 1 can be rejected. (Note that the value of D for α = 0.05 and N = 5 is 0.565).
Solution
Properties of Random Numbers
Random numbers are values generated to simulate random behavior in systems.
Good random numbers must satisfy statistical properties.
Mainly, they should be uniform and independent.
Uniformity:
Numbers are evenly distributed over the interval (0,1).
Each value has equal probability.
Independence:
Each number is not influenced by previous numbers.
No correlation or pattern exists.
Randomness (Unpredictability):
Future values cannot be predicted from past values.
Long Period:
The sequence should not repeat quickly.
It should have a large cycle length.
Efficiency:
Numbers should be generated quickly and easily.
Reproducibility (for pseudo-random):
The same seed produces the same sequence when needed.
Example: Random numbers used in simulation, cryptography, and modeling.
Conclusion: Good random numbers must be uniform, independent, and unpredictable for reliable
4. Write short notes on:
a. stationary Poisson process
b. Non-stationary Poisson process
c. Poker test
d.Random Variation
a. Stationary Poisson Process
A stationary Poisson process is a random arrival process with a constant rate (λ) over time.
The probability of arrivals is independent of time interval location.
It is widely used in queueing and simulation models.
Constant arrival rate (λ) over time.
Independent increments (arrivals in disjoint intervals are independent).
Number of arrivals follows Poisson distribution.
Inter-arrival time follows exponential distribution.
Example: Customers arriving at a bank at a steady average rate.
Conclusion: It models random events with constant average rate.
b. Non-Stationary Poisson Process
A non-stationary Poisson process has a time-dependent arrival rate λ(t).
The arrival rate changes with time.
Used for systems with varying demand.
Variable arrival rate λ(t).
Still has independent increments.
Suitable for peak and off-peak situations.
More realistic for real-world systems.
Example: Call arrivals in a call center (busy in day, low at night).
Conclusion: It models random events with changing arrival rates.
c. Poker Test
The poker test is a statistical test used to check the randomness of numbers.
It is based on patterns similar to poker hands.
Used in random number testing.
Digits are grouped and classified into patterns.
Patterns include all different, one pair, two pairs, three of a kind.
Frequencies are compared with expected values.
Uses chi-square test for validation.
Example: Checking whether generated random numbers follow expected pattern distribution.
Conclusion: Poker test verifies uniformity and independence of random numbers.
d. Random Variation
Random variation refers to natural fluctuations in data due to randomness.
It represents unpredictable changes in a system.
It is present in all real-world processes.
Caused by chance factors.
Cannot be completely eliminated.
Measured using variance or standard deviation.
Important in simulation and statistical analysis.
Example: Variation in customer arrival times at a service center.
Conclusion: Random variation reflects uncertainty and natural randomness in systems.
5. Generate 10 random integers using the Linear congruential method where m=1000, a =19, c=6, and X0=13
Soluiton
6. What are the two main properties of random numbers? Test whether the 3rd, 7th, 11th, and so on numbers in the sequence in the following random number sample are autocorrelated. (Zα=0.05 and Z0.025=1.96)
0.12 0.01 0.23 0.28 0.89 0.31 0.64 0.28 0.83 0.93 0.99 0.15 0.33 0.35 0.91 0.41 0.60 0.27 0.75 0.88 0.68 0.49 0.05 0.43 0.95 0.58 0.19 0.36 0.69 0.87
Solution
7. Generate ten 3-digit random integers and corresponding random variables using the Multiplicative Congruential method, where a =7, and X0= 22.
Solution
8. Define and develop a Poker test for four-digit random numbers. A sequence of 1,000 random numbers, each of four digits, has been generated. The analysis of the numbers reveals that in 525 numbers all four digits are different, 419 contain exactly one pair of like digits, 47 contain two pairs, 9 have three digits of a kind, and 7 contain all like digits. Use the Poker test to determine whether these numbers are independent.(Critical value of chi-square for a = 0.05 and N = 4 is 9.49).
Solution
9. Explain the generation of non-uniform random number generation using the inverse method.
SOlution
Generation of Non-Uniform Random Numbers (Inverse Method)
The inverse method is used to generate non-uniform random variates from a given probability distribution.
It transforms uniform random numbers into values following a desired distribution.
It is based on the cumulative distribution function (CDF).
Algorithm / Steps
Generate a uniform random number (U) in (0,1).
Find the CDF F(x) of the target distribution.
Set U = F(x).
Solve for x = F⁻¹(U) (inverse of CDF).
The value of x is the required non-uniform random variate.
Works when the inverse of CDF exists.
Simple and widely used in simulation.
10. Use the mixed congruential method to generate a sequence of random numbers with X0 = 27, n = 17, m = 100, and c = 43.
Solution
11. Difference between chi-square test and KS test for uniformity. Use the KS test to check for the uniformity of the input set of random numbers given below. 0.54, 0.73, 0.98, 0.11, 0.68, 0.45. Assume level of significance to be Dα=0.05 => 0.565
Solution
12. Define true random numbers and pseudo random numbers with their properties. The sequence of numbers 0.64, 0.50, 0.25, 0.58,0.72, 0.90 has been generated. Use KS Test with Da=0.050 => 0512 to determine if the hypothesis that they are uniformly distributed on the interval [0, 1] can be rejected.
Solution
True Random Numbers:
True random numbers are numbers generated from physical or natural processes.
They are completely unpredictable and non-deterministic.
No algorithm is used to generate them.
Generated using physical sources (noise, radiation, etc.).
Have high randomness and no pattern.
Not reproducible.
Slower and may require special hardware.
Example: Random numbers generated from thermal noise or atmospheric noise.
Conclusion: True random numbers provide real randomness, but are harder to generate.
Pseudo Random Numbers:
Pseudo random numbers are generated using mathematical algorithms.
They appear random but are actually deterministic.
Generated from an initial value called seed.
Follow uniformity and independence (approximately).
Reproducible using the same seed.
Faster and easy to generate using computers.
Have a finite period (eventually repeat).
Example: Numbers generated using Linear Congruential Generator (LCG).
Conclusion: Pseudo random numbers are efficient and practical, but not truly random.
Unit 6: Verification and Validation
1. Explain the iterative process of calibrating a simulation model.
Solution
Calibrating a simulation model is the process of tuning it until its output matches real-world historical data.
It is the process of adjusting model parameters so that the model matches real-world data.
It is done through an iterative process.
The goal is to achieve accurate and reliable simulation results.
Step 1: Initialize Model
Start with initial assumptions and parameter values.
Step 2: Run Simulation
Execute the model to generate output results.
Step 3: Compare Results
Compare simulation output with actual system data.
Step 4: Measure Error
Calculate error between simulated and real values.
Step 5: Adjust Parameters
Modify parameters to reduce error.
Step 6: Repeat Process
Run the simulation again and re-evaluate results.
Continue until acceptable accuracy is achieved.
Example:Adjusting arrival rate or service time in a queue model to match real waiting time.
Conclusion: Calibration is an iterative trial-and-error process that improves the accuracy and validity of a simulation model.
2. Describe the process of model building, verification, and validation in detail with examples.
Solution
Model Building
Model building is the process of creating a mathematical or logical representation of a real system.
It simplifies reality to analyze and predict system behavior.
It is the first step in simulation.
Identify problem and objectives.
Define system boundaries and assumptions.
Identify variables, parameters, and relationships.
Formulate mathematical or logical model.
Collect input data.
Example:
Modeling a bank queue system with arrival rate (λ) and service rate (μ).
Conclusion: Model building converts a real system into an analyzable form.
Verification:
Verification is the process of checking whether the model is implemented correctly.
It ensures the model follows the intended logic and design.
It answers: “Are we building the model right?”
Detects coding and logical errors.
Uses debugging and step-by-step testing.
Ensure model follows designed specifications.
Use test cases and step-by-step tracing.
Example: Checking if a simulation correctly computes waiting time.
Validation
Validation checks whether the model accurately represents the real system.
It ensures the model is realistic and reliable.
It answers: “Are we building the right model?”
Compare model output with real system data.
Use statistical tests and historical data.
Check if assumptions are realistic.
Modify model if results are inaccurate.
Example: Comparing simulated results with actual system performance.
Conclusion: Validation ensures the model is realistic and reliable for decision-making.
3. What is the three-step approach for validation of simulation models?
Solution
Three-Step Approach for Validation of Simulation Models
Validation checks whether the model accurately represents the real system.
It checks whether the model is reliable for decision-making.
It focuses on real-world correctness.
1. Face Validation (Conceptual Validation)
Check whether the model appears reasonable to experts.
Validate assumptions, logic, and structure.
Done through expert judgment and discussion.
Ensures model behavior is realistic at a basic level.
Quick and informal validation step.
Example: Experts review a bank queue model and confirm logic is realistic.
2. Validation with Historical Data
Compare model output with actual past data.
Check whether results are close to real observations.
Use statistical analysis.
Helps detect major inaccuracies.
Requires reliable historical data.
Example: Compare simulated waiting time with recorded bank data.
3. Predictive Validation
Test the model by predicting future system behavior.
Compare predictions with actual outcomes.
Measures model accuracy over time.
Strongest form of validation.
Confirms model is useful for forecasting.
Example: Predict future customer load and compare with actual results.
Conclusion: The three-step approach ensures the model is logically correct, matches past data, and predicts future behavior accurately.
4. Define verification and validation. Explain the process of model verification in brief.
Solution
Verification
Verification is the process of checking whether the model is implemented correctly.
It ensures the model is built according to design and logic.
It answers: “Are we building the model right?”
Validation
Validation is the process of checking whether the model represents the real system accurately.
It ensures the model is realistic and reliable.
It answers: “Are we building the right model?”
Process of Model Verification
Review Model Logic:
Check whether equations, assumptions, and flow are correct.
Debugging:
Identify and fix coding and logical errors.
Step-by-Step Tracing:
Execute model with sample inputs and trace outputs.
Consistency Check:
Ensure model behaves correctly under different conditions.
Comparison with Known Results:
Compare outputs with manual calculations or simple cases.
Sensitivity Check:
Change input values slightly and observe logical output changes.
Example:
In a queue model, verify that waiting time increases when arrival rate increases.
Conclusion: Verification ensures the model is free from errors and correctly implemented, forming a base for valid results.
5. Differentiate between validation and calibration. How can we perform validation of a model?
Solution
Validation of a Model
Validation is the process of checking whether a model accurately represents the real system.
It ensures the model is reliable for decision-making.
It answers: “Are we building the right model?”
How to Perform Validation
1. Face Validation (Conceptual Validation)
Check whether the model appears reasonable to experts.
Validate assumptions, logic, and structure.
Done through expert judgment and discussion.
Ensures model behavior is realistic at a basic level.
Quick and informal validation step.
Example: Experts review a bank queue model and confirm logic is realistic.
2. Validation with Historical Data
Compare model output with actual past data.
Check whether results are close to real observations.
Use statistical analysis.
Helps detect major inaccuracies.
Requires reliable historical data.
Example: Compare simulated waiting time with recorded bank data.
3. Predictive Validation
Test the model by predicting future system behavior.
Compare predictions with actual outcomes.
Measures model accuracy over time.
Strongest form of validation.
Confirms model is useful for forecasting.
Example: Predict future customer load and compare with actual results.
Conclusion: The three-step approach ensures the model is logically correct, matches past data, and predicts future behavior accurately.
6. Define the terms verification, calibration, validation, and accreditation of models.
Solution
Verification
Verification is the process of checking whether the model is implemented correctly.
It ensures the model follows the intended logic and design.
It answers: “Are we building the model right?”
Detects coding and logical errors.
Uses debugging and step-by-step testing.
Ensure model follows designed specifications.
Use test cases and step-by-step tracing.
Example: Checking if a simulation correctly computes waiting time.
Calibration
Calibration is the process of adjusting model parameters to match real system data.
It improves the accuracy of the model.
It involves fine-tuning input values.
Uses historical or observed data.
Parameters are modified until outputs are close to reality.
Example: Adjusting arrival rate (λ) in a queue model to match actual data.
Validation
Validation checks whether the model accurately represents the real system.
It ensures the model is realistic and reliable.
It answers: “Are we building the right model?”
Compare model output with real system data.
Use statistical tests and historical data.
Check if assumptions are realistic.
Modify model if results are inaccurate.
Example: Comparing simulated results with actual system performance.
Conclusion: Validation ensures the model is realistic and reliable for decision-making.
Accreditation:
Accreditation is the formal process of officially accepting the model for use.
It is done by authorized individuals or organizations.
Confirms the model is suitable for its intended purpose.
Based on verification and validation results.
Required for critical applications.
Example: Government approving a simulation model for policy decisions.
Conclusion: These processes ensure a model is correct, accurate, adjusted, and officially approved for use.
7. “Building a model right” and “Building a right model”. Discuss the importance of V & V.
Solution
“Building a Model Right” vs “Building a Right Model”:
These two phrases represent verification and validation concepts in modeling.
They are often confused, but they test different aspects of a model.
Both are essential for reliable simulation results.
Building a Model Right (Verification):
Verification is the process of checking whether the model is implemented correctly.
It ensures the model follows the intended logic and design.
It answers: “Are we building the model right?”
Detects coding and logical errors.
Uses debugging and step-by-step testing.
Ensure model follows designed specifications.
Use test cases and step-by-step tracing.
Example: Checking if a simulation correctly computes waiting time.
Building a Right Model (Validation):
Validation checks whether the model accurately represents the real system.
It ensures the model is realistic and reliable.
It answers: “Are we building the right model?”
Compare model output with real system data.
Use statistical tests and historical data.
Check if assumptions are realistic.
Modify model if results are inaccurate.
Example: Comparing simulated results with actual system performance.
Conclusion: Validation ensures the model is realistic and reliable for decision-making.
Importance of Verification & Validation (V & V):
Ensures accuracy and reliability of the model.
Detects errors in design and implementation.
Improves confidence in simulation results.
Helps in better decision-making.
Avoids wrong conclusions and costly mistakes.
Ensures model is fit for its intended purpose.
Example:
Without V&V, a faulty model may give incorrect predictions, leading to poor decisions.
Conclusion
Verification ensures the model is built correctly, while validation ensures it is the correct model.
Both are critical to ensure trustworthy and useful simulation outcomes.
Unit 7: Analysis of Simulation Output
1.Why is it necessary to analyze the simulation output? Explain different estimation methods used in simulation output analysis.
Solution
Need for Simulation Output Analysis
Simulation output analysis is required to interpret raw results produced by a model.
It helps in making statistically valid conclusions.
It ensures results are reliable and not due to randomness.
Simulation outputs are random (stochastic).
Single run may give misleading results.
Helps estimate performance measures (mean, variance).
Supports decision-making and comparison of alternatives.
Example: Estimating average waiting time from multiple simulation runs.
Conclusion: Output analysis converts simulation data into meaningful and reliable information.
Estimation Methods in Simulation Output Analysis
Estimation methods are used to obtain accurate performance measures from simulation data.
They reduce the effect of random variation.
1. Point Estimation:
Provides a single value estimate of a parameter.
Usually the sample mean (average).
Easy to compute.
Does not show uncertainty.
Example: Average waiting time = 5 minutes.
2. Interval Estimation (Confidence Interval):
Provides a range of values within which the true value lies.
Expressed with a confidence level (e.g., 95%).
More reliable than point estimate.
Shows precision of estimate.
Example: Waiting time = 5 ± 1 minutes (95% confidence).
3. Replication Method:
Run simulation multiple times independently.
Compute mean and variance across runs.
Reduces random error.
Improves accuracy.
Example: Run simulation 10 times and average results.
4. Batch Means Method:
Divide a long simulation run into batches.
Treat each batch as a separate observation.
Useful for steady-state simulations.
Reduces correlation between data.
Example: Divide 1000 observations into 10 batches.
Conclusion
Output analysis is essential to ensure valid and accurate simulation results.
Estimation methods like point, interval, replication, and batch means help in obtaining reliable performance measures.
2. Why is a confidence interval needed in the analysis of simulation output? How can we establish a confidence interval?
Solution
Need for Confidence Interval in Simulation Output
A confidence interval (CI) gives a range of values for the true performance measure.
Simulation outputs are random, so a single value is unreliable.
CI shows the accuracy and uncertainty of estimates.
Provides reliability of results.
Helps compare different system designs.
Indicates precision (narrow interval = better estimate).
Avoids misleading conclusions from a single run.
Example: Average waiting time is not just 5 min, but (4.5, 5.5) with 95% confidence.
Conclusion: Confidence interval is needed to make statistically valid and trustworthy decisions.
Establishing a Confidence Interval
- CI is constructed using sample mean and variability.
- Based on multiple observations (replications or batches).
- Step 1: Run simulation n times → get outputs
- Step 2: Compute sample mean (x̄)
- Step 3: Compute sample standard deviation (s)
- Step 4: Choose confidence level (e.g., 95%)
- Step 5: Use t-distribution (for small samples)
- Confidence Interval formula:
- Where:
- x̄ = sample mean
- s = standard deviation
- n = number of observations
- t = critical value
-
Example:
x̄ = 5, s = 1, n = 25 → CI ≈ 5 ± (t × 0.2)
3. Explain different estimation methods used in simulation output analysis.
Solution
Estimation Methods in Simulation Output Analysis:
Estimation methods are used to obtain accurate performance measures from simulation data.
They reduce the effect of random variation.
1. Point Estimation:
Provides a single value estimate of a parameter.
Usually the sample mean (average).
Easy to compute.
Does not show uncertainty.
Example: Average waiting time = 5 minutes.
2. Interval Estimation (Confidence Interval):
Provides a range of values within which the true value lies.
Expressed with a confidence level (e.g., 95%).
More reliable than point estimate.
Shows precision of estimate.
Example: Waiting time = 5 ± 1 minutes (95% confidence).
3. Replication Method:
Run simulation multiple times independently.
Compute mean and variance across runs.
Reduces random error.
Improves accuracy.
Example: Run simulation 10 times and average results.
4. Batch Means Method:
Divide a long simulation run into batches.
Treat each batch as a separate observation.
Useful for steady-state simulations.
Reduces correlation between data.
Example: Divide 1000 observations into 10 batches.
Conclusion
Estimation methods like point, interval, replication, and batch means help in obtaining reliable performance measures.
4. What do you mean by replication of runs? Why is it necessary?
Solution
Replication of Runs:
Replication of runs means executing the same simulation model multiple times with different random inputs.
Each run is independent and produces a separate output.
It is used to obtain reliable statistical results.
Each run gives a value: X₁, X₂, X₃, …, Xₙ
These outputs are used to compute the average and variability.
X̄ (mean) gives the estimated performance measure.
Multiple runs reduce the effect of randomness.
Replication is necessary because of the following:-
Simulation output is random:
One run can give misleading results.
Improves accuracy:
Averaging multiple runs gives better estimates.
Measures variability:
Helps compute variance and confidence intervals.
Reduces bias:
Eliminates dependence on a single random sequence.
Supports comparison:
Useful when comparing different system designs.
Example
Run a queue simulation 10 times
Waiting times: 4, 6, 5, 7, 5, 6, 4, 5, 6, 5
Average gives a more reliable estimate than any single run
Conclusion
Replication ensures statistical reliability by reducing randomness.
Without it, your results are basically guesswork disguised as analysis.
5. Explain the importance of the elimination of initial bias during simulation.
Solution
Elimination is Important because
Initial results are often distorted or unrepresentative.
Affects estimates like average waiting time and queue length.
Leads to wrong conclusions and poor decisions.
Required for accurate steady-state analysis.
Example:
In a queue system starting with an empty queue, waiting time is initially low → not realistic.
Eliminate Initial Bias
Warm-up Period (Transient Removal):
Ignore initial simulation output until the system reaches steady state.
Graphical Method:
Plot performance measure vs time and identify stabilization point.
Statistical Methods:
Use tests to detect when data becomes stable.
Multiple Replications:
Helps reduce the impact of initial conditions.
Conclusion
Eliminating initial bias ensures simulation results reflect true long-term system behavior, not misleading early conditions.
6. Write short notes on:
a. Hypothesis testing
b. Simulation run statistics
c. System and its environment
d. Simulation run statistics
a)Hypothesis Testing:
Hypothesis testing is a statistical method to make decisions based on sample data.
It tests an assumption about a population parameter.
It determines whether to accept or reject a claim.
Define Null hypothesis (H₀) and Alternative hypothesis (H₁).
Choose significance level (α) (e.g., 0.05).
Compute test statistic.
Compare with critical value or p-value.
Example: Testing whether average waiting time = 5 minutes.
Conclusion: Hypothesis testing helps in making statistically valid decisions.
b) Simulation Run Statistics:
Simulation run statistics are measures obtained from one simulation run.
They summarize the performance of the system.
Used for analysis and comparison.
Includes mean, variance, max, min values.
Can be time-based or observation-based.
Helps understand system behavior.
Example: Average queue length in one simulation run.
Conclusion: Run statistics provide quantitative insight into system performance.
c) System and its Environment:
A system is a set of interacting components working together.
The environment includes everything outside the system that affects it.
Both are essential in modeling and simulation.
The system has inputs, processes, and outputs.
The environment provides external influences.
Proper boundary definition is important.
Example:
Bank = system; customers and external factors = environment.
Conclusion: Understanding the system and environment ensures accurate modeling.
Unit 8: Simulation of Computer Systems
-
Develop GPSS block diagram and code for a manufacturing shop problem and explain blocks used.
-
Why is GPSS called transaction flow oriented language?
-
What is storage in GPSS? Describe the blocks associated to storage in GPSS.
-
What is transaction in GPSS? Explain about facility in GPSS.
-
Draw GPSS block diagram for barbershop system and run simulation.
-
Represent the system in GPSS using facility and run simulation for given parts.
-
Draw GPSS block diagram to simulate the inspection system for 100 parts.
-
Draw GPSS block diagram to simulate supply store problem for 100 requisitions.
-
Write short notes on:
a. Simulation tools