Simulation Basics: Definitions, Types, Models & Applications

What is a Simulation?

A simulation is an imitation of a real-world process or system over time. It uses mathematical models, algorithms, and computational tools to analyze behavior without physically interacting with the system.

Why Simulate?

• Cost-effective: Test systems virtually (e.g., crash-testing cars without building prototypes).
• Safe: Study dangerous scenarios (e.g., nuclear reactor failures).
• Time-efficient: Speed up or slow down processes (e.g., aging a product in hours instead of years).
• Flexibility: Modify variables easily (e.g., traffic flow under different weather conditions).

Key Components of a Simulation

Every simulation has three core elements:

1. Real System: The physical or abstract process being studied (e.g., a hospital emergency room, a stock market).
2. Model: A simplified representation of the system (e.g., equations, flowcharts, or agent-based rules).
3. Simulation Engine: The computational tool that runs the model (e.g., Python, MATLAB, or specialized software like AnyLogic).
4. Output: Results (e.g., graphs, statistics, or animations) used for decision-making.

Types of Simulations

Simulations are classified based on the nature of the system and time representation:

How Simulations Work: A Step-by-Step Trace

Let’s simulate a simple bank queue to understand the process.

Step 1: Define the System

• Real System: A bank with 1 teller, customers arriving randomly, and service times varying between 2–5 minutes.
• Assumptions:
  • Customers arrive every 3 minutes on average (Poisson process).
  • Service time is uniformly distributed between 2–5 minutes.

Step 2: Build the Model

We’ll use a discrete-event simulation (customers arrive and leave at specific times). Key Variables:

• Arrival time: Randomly generated using a Poisson distribution.
• Service time: Randomly generated between 2–5 minutes.
• Queue length: Number of customers waiting.

Step 3: Run the Simulation (Trace for 30 Minutes)

Visualization of the Queue Over Time:

timeline
    title Bank Queue Simulation (30 min)
    0: System starts
    2: Customer arrives (Queue: 1)
    4: Service starts (Customer 1)
    5: Customer arrives (Queue: 1)
    7: Customer 1 leaves (Queue: 0)
    8: Customer arrives (Queue: 1)
    11: Customer 2 leaves (Queue: 0)
    12: Customer arrives (Queue: 1)
    13: Customer 3 leaves (Queue: 0)
    ...
    30: End (10 customers served)

In the Real World

Simulations are everywhere in Nepal and globally. Here’s how companies use them:

1. Khalti (Digital Payments)

   • Idea Used: Discrete-event simulation of transaction queues.
   • How: Khalti simulates peak-hour payment processing to optimize server capacity. For example, during Dashain, millions of transactions occur simultaneously. By simulating arrival rates (e.g., 10,000 transactions/minute) and processing times (e.g., 0.5 seconds/transaction), they ensure servers don’t crash.
   • Real Trace:
     • Input: 5000 transactions/minute, 90% success rate.
     • Simulation: If servers handle 3000 transactions/minute, 20% fail → Khalti adds more servers.
     • Output: Scaled to handle 8000 transactions/minute during festivals.

2. NTC (Power Grid Management)

   • Idea Used: Continuous-time simulation of electrical load.
   • How: NTC uses simulations to predict power demand during load-shedding periods. For example, during summer, AC usage spikes. By modeling temperature vs. power consumption (e.g., +5°C → +10% load), they schedule generators efficiently.
   • Real Trace:
     • Input: Historical data (2018–2023) shows July demand = 2500 MW at 35°C.
     • Simulation: Predict 2024 demand at 38°C → 2800 MW.
     • Output: Schedule 300 MW extra from hydropower plants.

3. Pathao (Ride-Sharing)

   • Idea Used: Agent-based simulation of driver-customer matching.
   • How: Pathao simulates driver availability vs. ride requests in Kathmandu. For example, during peak hours (6–9 PM), demand surges in Thapathali. By simulating 1000 drivers with 2-minute response times, they optimize pricing and driver incentives.
   • Real Trace:
     • Input: 5000 ride requests/hour, 80% driver acceptance rate.
     • Simulation: If drivers take >3 min to accept, 30% requests fail → Pathao offers bonuses for faster responses.
     • Output: Reduced wait time from 8 to 4 minutes.

Simulation vs. Real Experiment: A Comparison

Advantages and Disadvantages of Simulations

✅ Advantages

• Cost savings: Avoid expensive real-world trials (e.g., airplane wing testing).
• Risk reduction: Test dangerous scenarios (e.g., nuclear meltdowns).
• Reproducibility: Run the same experiment multiple times.
• Complexity handling: Model systems with thousands of variables (e.g., climate models).
• Education: Train personnel (e.g., pilots in flight simulators).

❌ Disadvantages

• Model limitations: Simplifications may miss real-world nuances.
• Computational cost: Large simulations require powerful hardware.
• Validation needed: Results must be verified against real data.
• Learning curve: Requires expertise in modeling and programming.

Common Applications of Simulation

Worked Example: Simulating a Pendulum (Continuous System)

A pendulum’s motion is governed by physics equations, but we can simulate it digitally.

Step 1: Define the Model

The pendulum’s angle θ(t) follows: $$\frac{d^2\theta }{d{t}^2}+\frac{g}{L}\sin \theta =0$$ where:

• $g=9.8{\text{m/s}}^2$ (gravity),
• $L=1\text{m}$ (length),
• Initial angle $\theta (0)=30°$.

Step 2: Discretize Time

Use Euler’s method to approximate the solution: $$\theta (t+\Delta t)=\theta (t)+\Delta t\cdot \frac{d\theta }{dt}$$ Let $\Delta t=0.1\text{s}$.

Step 3: Simulate for 10 Seconds

Visualization of Pendulum Motion:

Graph of Angle vs. Time:

Simulation Software Tools

Students often use these tools for assignments:

Exam Tip: How to Score Full Marks

This unit is theoretical but practical. Examiners test:

1. Definitions: Know the difference between model, simulation, discrete vs. continuous.
2. Applications: Link simulations to real-world examples (e.g., "How would you simulate Khalti’s payment system?").
3. Diagrams: Draw a system model (like the bank queue) or a timeline (like the pendulum).
4. Worked Examples: Show step-by-step traces (like the bank queue table).
5. Advantages/Disadvantages: Compare simulations to real experiments.

Common Mistakes to Avoid:

• ❌ Confusing discrete (events at specific times) with continuous (smooth changes).
• ❌ Forgetting to mention randomness in simulations (e.g., arrival times in queues).
• ❌ Not validating the model (always check if the simulation matches reality).

Model Answer Structure for Short Questions:

Q: Explain discrete-event simulation with an example. A: A discrete-event simulation models a system where changes occur at distinct points in time (events). Example: A bank queue.

• Events: Customer arrival, service start, service completion.
• State variables: Queue length, teller status (idle/busy).
• Process:
  1. Customer arrives → joins queue.
  2. Teller becomes free → serves next customer.
  3. Service ends → customer leaves.
• Output: Average wait time, queue length distribution. Visual:

Summary Checklist

Before the exam, ensure you can: ✅ Define simulation, model, and system. ✅ Classify systems as discrete/continuous/hybrid. ✅ Trace a simple simulation (e.g., bank queue or pendulum). ✅ List 3 real-world Nepalese examples (Khalti, NTC, Pathao). ✅ Compare simulations to real experiments. ✅ Draw a timeline or state diagram for a simulation.