Computer Simulation

Arena Modeling

Dr. Jeff Hong

IELM Dept, HKUST
Fall 2010

Note 05: Arena Modeling 1

Case 1: The Darker Image
„ The Arena Model
Assign Self Self Service
Service Time Copying

Arrival Self Service?

0 T ru e
0 D epa rtu re

0 0
0 Fa ls e Assign Full Full Service
Service Time Copying

Proposal One 0

Assign Self
Service Time

0 T ru e
Arrival Self Service? Copy Machine Departure
0 0
0 0
Fa ls e Assign Full
Service Time

Proposal Two

Note 05: Arena Modeling 2

Launch Arena

Start
‐>All Programs
‐>Rockwell 
Software
‐>Arena 10.0
‐>Arena 10.0

Note 05: Arena Modeling 3

Arena World View
„ Arena takes the process interaction world view.
„ Entities flow through a network of modules that describe
their logical behavior.
„ We describe the network by developing a process
flowchart.
Assign Self
Service Time

0 T ru e
Arrival Self Service? Copy Machine Departure
0 0
0 0
Fa ls e Assign Full
Service Time

Note 05: Arena Modeling 4

About Modules…

„ Arena contains a very large number of modules that are

organized into panels.
„ The panels are structured from high level to low level
concepts:
‰ Basic Process
‰ Advanced Process & Advanced Transfer
‰ Blocks & Elements (a programming language)

Note 05: Arena Modeling 5

Entities

„ Entities are dynamic elements that pass through the

system.
„ Entities are distinguished by their attributes.
„ Ex: people, parts, information, paperwork, etc.
„ Entities must be Created to get them into the model,
and are Disposed when they leave.

Note 05: Arena Modeling 6

Queueing
„ Entities queue when they need processing.

„ In Arena…
‰ An entity tries to Seize a Resource.
‰ The time the entity uses the resource is the Delay.
‰ If the resource is not available, the entity waits in a Queue.
‰ The entity Releases the resource when processing is
complete.

„ Queues are created automatically by some modules

(e.g., Process).

Note 05: Arena Modeling 7

Resources

„ Resources have…
‰ A Name (up to you)
‰ A Capacity (number of identical units of this resource).
‰ And can have a Schedule (how many available when).

„ Resources are automatically defined by some

modules (e.g., Process)
„ Resources can also be defined manually, and the
properties of all resources are changed, via the
Resources spreadsheet on the Basic Process panel.

Note 05: Arena Modeling 8

Basic Process Modules

Push (possibly) batches of entities into

the model with a (possibly) random time
Create 1
between.
0

Models Queue-Seize-Delay-Release of
Proc ess 1 Resource, or any part of this (like pure
Delay).
0
Make decisions about where to go next
based on conditions or chance.

Take entities out of the model and

Dispose 1 (perhaps) record statistics.
0

Note 05: Arena Modeling 9

Basic Process Modules

Assign values (especially Attributes)

when an entity passes through.
Assign 1

Record information when entities pass

Record 1
through, typically statistics on entities.

Combine multiple entities into a single

entity.
Batch 1

Split multiple entities that were combined,

0
Separate 1 or duplicate a single entity.
Original

0 Duplicate

Note 05: Arena Modeling 10

Simulate Proposal One (step by step)

Assign Self Self Service

Service Time Copying

Arrival Self Service?

0 T ru e
0 D epa rtu re

0 0
0 Fa ls e Assign Full Full Service
Service Time Copying

Note 05: Arena Modeling 11

Create arrival of jobs

2 3 6
4

7
5

Note 05: Arena Modeling 12

Determine types of jobs

Note 05: Arena Modeling 13

Assign service times to self service jobs

3
4
1
2

Note 05: Arena Modeling 14

Processing delay and queue for self service
jobs

1
4

2
3
5

7
Note 05: Arena Modeling 15
Assign service times to full service jobs

Note 05: Arena Modeling 16

Processing delay and queue for full service
jobs

1
2

Note 05: Arena Modeling 17

Departure

Note 05: Arena Modeling 18

Run: set model & run parameters
1

Note 05: Arena Modeling 19

Run to completion

Note 05: Arena Modeling 20

Report

Note 05: Arena Modeling 21

Simulate Proposal Two (sketch)

Assign Self
Service Time

0 T ru e
Arrival Self Service? Copy Machine Departure
0 0
0 0
Fa ls e Assign Full
Service Time

Note 05: Arena Modeling 22

Create arrival of jobs

Note 05: Arena Modeling 23

Determine types of jobs

Note 05: Arena Modeling 24

Assign service times to self service jobs

Note 05: Arena Modeling 25

Assign service times to full service jobs

Note 05: Arena Modeling 26

Processing delay and queue

Note 05: Arena Modeling 27

Departure

Note 05: Arena Modeling 28

Resource: set resource capacity (different
from Proposal One)

Note 05: Arena Modeling 29

Summary

The rest steps are similar with Proposal One (you

can finish them now).

So far,
„ you have learned Arena modeling;

„ you are familiar with basic process modules.

We now introduce another example.

Note 05: Arena Modeling 30

Case 2: Warehouse Design
„ A warehouse is open from 7:00am to 4:00pm. The
warehouse is equipped with two forklifts, four loading
bays and a parking lot.
„ Customer (with truck) comes to the warehouse. If there
is empty loading bay in the warehouse, the truck stops in
loading bay and waits for forklift to pick up goods,
otherwise, the truck joins the queue in the parking lot.
„ Forklift picks up goods for trucks in the loading bays
according to first come first serve rule, and unloads the
goods in the loading bay. Truck driver loads goods on
the truck, then the truck leaves.

Note 05: Arena Modeling 31

Warehouse
„ Current situation

Parking Lot Loading Bay

Note 05: Arena Modeling 32
Assumption
„ According to data analysis, we make the following
assumptions for the distributions of input variables：
‰ Customers arrive from 7:00am to 4:00pm. Interarrival time for
customers follows an exponential distribution, with mean 20
minutes.
‰ The pick up time of forklift follows a triangular distribution, with
parameters (20,35,60) minutes.
‰ The loading goods time of truck driver follows a triangular
distribution, with parameters (5,10,20) minutes.

Note 05: Arena Modeling 33

Question
„ Build ARENA model to simulate the warehouse
„ Take a sample of ten days, simulate average number of
trucks in the parking lot, average time that customers
spend in the warehouse and the utilization of forklifts.
„ To improve the quality of service, the warehouse
suggests two proposals
‰ proposal one：add a forklift
‰ proposal two：add a loading bay
„ Which proposal is better for improving the service quality？

Note 05: Arena Modeling 34

Create: customer arrival

Note 05: Arena Modeling 35

Process: occupy loading bay

Note 05: Arena Modeling 36

Process: pick up orders

Note 05: Arena Modeling 37

Process: loading goods

Note 05: Arena Modeling 38

Process: departure

Note 05: Arena Modeling 39

Resource: set resource capacity

Note 05: Arena Modeling 40

Run: set model & run parameters

Note 05: Arena Modeling 41

Report

Note 05: Arena Modeling 42

Average total time spent in warehouse

Note 05: Arena Modeling 43

Average number of trucks in parking lot

Note 05: Arena Modeling 44

Utilization of forklift

Note 05: Arena Modeling 45

 Extension 1: add a forklift

Note 05: Arena Modeling 46

Extension 2: add a loading bay

Note 05: Arena Modeling 47

Extension 3: schedule

„ Resource Schedules dictate when resources are

available.

Note 05: Arena Modeling 48

Schedule rule

Note 05: Arena Modeling 49

Extension 3: schedule (cont’d)

„ The forklift drivers need to have lunch (Suppose each

forklift driver take one-hour lunch break from 12-1 pm) .

„ Assume the resource schedule to be “ignore”.

Note 05: Arena Modeling 50

Extension 3: schedule (cont’d)

2 3
4

Note 05: Arena Modeling 51

Extension 3: schedule (cont’d)

1
2, right click, select
edit via dialog…

Note 05: Arena Modeling 52

Extension 3: schedule (cont’d)

Note 05: Arena Modeling 53

Extension 3: schedule (cont’d)

3
1

Note 05: Arena Modeling 54

Extension 3: schedule (cont’d)

3
1

Note 05: Arena Modeling 55

Extension 3: schedule (cont’d)

Note 05: Arena Modeling 56

Case 3: Simulate Arrival Situations

„ In this case, we will discuss three commonly

seen arrival situations in projects:

1. Joining the shortest queue

2. Balking

3. NSAP: non-stationary arrival process

Note 05: Arena Modeling 57

Joining the shortest queue

„ Imagine you go to McDonald’s, there are

three queues:
‰ Counter A: 4 people
‰ Counter B: 3 people
‰ Counter C: 4 people

„ By probability? No!

Note 05: Arena Modeling 58

In Arena, …

Note 05: Arena Modeling 59

Drawbacks

„ Once customers (entities) join the queue, they

could not change.

„ Some counters will always be higher utilized.

Note 05: Arena Modeling 60

Balking

„ Sometimes, when customers see that the queue

is too long, then they decide not to join the
queue. This is called “balking”.

„ The difficulty of modeling balking is that not

everyone balk, and people balk at different
queue lengths.

Note 05: Arena Modeling 61

In Arena, …

Note 05: Arena Modeling 62

NSAP

„ You can find the arrival rate changes among

time.
„ NSAP: Non-stationary arrival process
‰ Use the Create module to produce arrivals at the smallest
mean TBA (Time between arrival).
‰ Use Decide to randomly accept only enough arrivals to get
the mean TBA we want. Generally accept smallest/current x
100%
‰ Use Create-Assign to change the value of the current mean
TBA over time (hour by hour here).
„ e.g.,

Note 05: Arena Modeling 63

Time dependent

Note 05: Arena Modeling 64

In Arena, …

Note 05: Arena Modeling 65

Exercise

„ A fast food restaurant has 4 counters. Assume all

counters are the same, with service time uniformly
distributed between 1 and 3 minutes.

„ The restaurant operates from 8:00 am to 2:00 pm.

Interarrival time between each customers are
exponentially distributed with mean 1 minutes in the first
two hours, 5 minutes in the following two hours, and 0.5
minutes in the last two hours.

Note 05: Arena Modeling 66

Exercise (cont’d)

„ After customers come to front of counters, they will

decide whether to stay or not, depending on the total
number of people waiting in all 4 queues. Their tolerance
are normally distributed with mean 10 and variance 4.

„ After the customers buy the food, they will leave.

„ All customers will join the shortest queue.

„ Simulate 100 days to see the results.

Note 05: Arena Modeling 67