Lecture 4: Introduction to Discrete Event Simulation

END 322E System Simulation Mehmet Ali Ergün, Ph.D.

1
 Introduction

❖ We will see some simulation examples that can be performed via using a
simulation table
❖ By hand
❖ Spreadsheet
❖ The simulation table provides a systematic method for tracking system state
over time
❖ Provide insights into the methodology of discrete event system simulation
and descriptive statistics

2
 Simulation of a Queueing System
❖ Consider a single server queue (ATM)
❖ Infinite calling population and infinite queue capacity
❖ Single server (Only one customer served at a time)
❖ Customers served in First in First out (FIFO) fashion
❖ No departures from the queue (patient customers )
Arrivals

Departures
Population Server

Waiting Line
3
 Simulation of a Queueing System
❖ Arrivals occur one at a time in a random fashion
❖ Once they join the waiting line they are eventually served
❖ Service times are of some random length according to a probability distribution

Arrivals

Departures
Population Server

Waiting Line
4
 Simulation of a Queueing System
❖ State: Number of units in the system and the status of the server, busy or idle
❖ Event: A circumstance that cause a change in system state
❖ Arrival of a customer
❖ Departure of a customer after service completion
❖ Simulation Clock
Arrivals

Departures
Population Server

Waiting Line
5
 Simulating the clock

❖ Obvious choice: Start counting 1,2,3…

❖ But it is inefficient in most cases

❖ Solution:Discrete Event Simulation

❖ Get to the earliest event and update state and simulation clock

6
 Simulation of a Queueing System

Arrival Event

No Yes
Server
Busy?

Customers Enters queue

Customers Enters Service for service

7
 Simulation of a Queueing System
Departure
Event

No Another Yes
Customer
waiting?
Remove the next
Begin Server Idle customer from the
Time queue

Begin servicing
the customer

8
 Simulation of a Queueing System
❖ Events usually occur at random times
❖ The randomness in simulation imitates uncertainty in real life. It is not
known with certainty
❖ When the next customer will arrive to a grocery store
❖ How long for a bank teller will take to complete a transaction
❖ A statistical model developed from historical data

9
 Simulation of a Queueing System
❖ For simplicity assume that interarrival times (time between arrivals) are the
result of rolling a die.
❖ 1 to 6 equally likely
❖ Assume that service times are generated from a distribution with four values;
1, 2, 3, 4 minutes. They are equally likely
❖ Assume that there is a customer using the ATM machine (server) at time 0.
❖ State: Number of customers in the system
❖ Let’s simulate this simple system

1
0
 Simulation of a Queueing System
❖ Simulation Clock t=0, Number of customers in the system, St=1
❖ Assign service time to the first customer.
❖ 2 minutes (t=2)
❖ Assign interarrival time for next customer (time until next arrival)
❖ 2 minutes (t=2)

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0

1
1
 Simulation of a Queueing System
❖ Simulation Clock t=0, Number of customers in the system, St=1
❖ Assign service time to the first customer.
❖ 2 minutes (t=2)
❖ Assign interarrival time for next customer (time until next arrival)
❖ 2 minutes (t=2)

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2

1
2
 Simulation of a Queueing System
❖ Simulation Clock t=0, Number of customers in the system, St=1
❖ Assign service time to the first customer.
❖ 2 minutes (t=2)

❖ Assign interarrival time for next customer (time until next arrival)
❖ 2 minutes (t=2)

❖ Advance clock to the closest event time, update state

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2

1
3
 Simulation of a Queueing System
❖ Simulation Clock t=2
❖ Customer 1 departs the server, Customer 2 arrives (interarrival time was 2) and starts using
server since the first customer left, St=1
❖ Assign service time for Customer 2: 1 minute (t=3)
❖ Assign interarrival time for Customer 3: 4 minutes (t=6)
❖ Move clock to the nearest event (departure)

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2

1
4
 Simulation of a Queueing System
❖ Simulation Clock t=3
❖ Customer 2 departs, no waiting patients in the queue, server is idle, St=0
❖ Advance the clock to the next closest event (arrival)

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6

1
5
 Simulation of a Queueing System
❖ Simulation Clock t=6
❖ Customer 3 arrives, starts using server, St=1
❖ Assign service time for Customer 3: 3 minutes (t=9)
❖ Assign interarrival time for Customer 4: 1 minute (t=7)
❖ Move the clock to nearest event (arrival)

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6 6 3 9
4 7
1
6
 Simulation of a Queueing System
❖ Simulation Clock t=7
❖ Customer 4 arrives, Customer 3 in server, Customer 4 joins the queue, St=2
❖ Assign interarrival time for Customer 5: 2 minutes (t=9)
❖ Advance clock to the nearest event (arrival and departure)
Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6 6 3 9
4 7
5 9
1
7
 Simulation of a Queueing System
❖ Simulation Clock t=9
❖ Customer 5 arrives, Customer 3 departs, Customer 4 starts service, customer 5 joins the queue,
St=2
❖ Assign a service time for Customer 4: 2 minutes (t=11)
❖ Assign an interarrival time for Customer 6: 6 minutes (t=15)
Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6 6 3 9
4 7 9 2 11
5 9
6 15
1
8
 Simulation of a Queueing System
❖ Simulation Clock t=11
❖ Customer 4 departs, Customer 5 in server starts service, St=1
❖ Assign a service time for Customer 5: 1 minute (t=12)
❖ Arrival for Customer 6 is at t =15
❖ Advance to the next event (departure)
Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6 6 3 9
4 7 9 2 11
5 9 11 1 12
6 15
1
9
 Simulation of a Queueing System
❖ Simulation Clock t=12
❖ Customer 5 departs, no customer in the queue, St=0
❖ Advance to arrival of customer 6

Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6 6 3 9
4 7 9 2 11
5 9 11 1 12
6 15
2
0
 Simulation of a Queueing System
❖ Simulation Clock t=15
❖ Customer 6 arrives, starts getting service, St=1
❖ Assign a service time for Customer 6: 4 minutes (t=19)
❖ END OF SIMULATION
Customer Arrival Time Time Service Begins Service Time Time Service Ends
1 0 0 2 2
2 2 2 1 3
3 6 6 3 9
4 7 9 2 11
5 9 11 1 12
6 15 15 4 19
2
1
Simulation of a Queueing System

22
 Simulation of a Grocery Checkout
❖ A small grocery store has only one checkout counter. Customers arrive
at this checkout at random from 1 to 8 minutes apart. Each possible
value of interarrival time has the same probability of occurrence. The
service time vary from 1 to 6 minutes with the probabilities shown
below. Analyze the system by simulating the arrival of 100 customers.
Service Time Probability Cumulative Probability
1 0.10 0.10
2 0.20 0.30
3 0.30 0.60
4 0.25 0.85
5 0.10 0.95
6 0.05 1.00
23
 Simulation of a Grocery Checkout
❖ How can we generate random interarrival times and service times in
accordance with the given distributions?
❖ Generate a uniform random number R, between 0 and 1, then return
service time x, if it falls in the interval (F(x-1),F(x)]
Service Time Probability Cumulative Probability Random Number Interval
1 0.10 0.10 (0-0.10]
2 0.20 0.30 (0.10-0.3] 0.553
3 0.30 0.60 (0.30-0.60]
4 0.25 0.85 (0.60-0.85] Generated
Service
5 0.10 0.95 (0.85-0.95] time 3
6 0.05 1.00 (0.95-1.00] minutes
24
 Simulation of a Grocery Checkout
❖ Same idea for interarrival times

IA Time Probability Cumulative Probability Random number interval

1 0.125 0.125 (0.000-0.125]
2 0.125 0.250 (0.125-0.250]
3 0.125 0.375 (0.250-0.375]
4 0.125 0.500 (0.375-0.500]
5 0.125 0.625 (0.500-0.625]
6 0.125 0.750 (0.625-0.750]
7 0.125 0.875 (0.750-0.875]
8 0.125 1.000 (0.875-1.000]

25
Generated Inter-arrival Times

26
Generated Service Times

27
 Exercise: Complete the Table
❖ Use the generated numbers given in the previous slides to run the simulation
until the arrival of the 5 customer
th

❖ Use the table template below:

Cust. Arrival Time Service Service Waiting Time Time Time in Server
Time Begins Time in Queue Service system idle time
Ends
1 0 0 0

28
Final Simulated Table

29
 Analysis Using the Simulation Results

❖ The average waiting time for a customer is

❖ The average service time is (E[Service Time] =?)

❖ The average time spent in system

(1.74+3.17) 30
❖ The proportion of times that server is idle

❖ The probability that a customer has to wait in the queue

Note: 46 can be found by counting the number of customers who waited in the queue more than 0
minutes. This cannot be directly observed from the table presented in the slides, since the table is not
complete.
31
 Excel Time

❖ Now Let’s see how we can simulate this grocery store example in Excel…

32