Intro to Arena

Download Report

Transcript Intro to Arena

Intro to Arena
A Second Simulation
Model 2
We now consider a more complex simulation, from Ch. 5 of
Simulation with Arena.
We consider an assembly plant which receives two types of items,
prepares them, checks them, attempts to repair the faulty ones,
and ships the good finished parts, while scrapping the unfixable
ones.
Part A prep
Part A
EXPO(5)
Rework
TRIA(1,4,8)
Sealer
9%
EXPO(45)
Part B prep
Part B
Batches of 4
EXPO(30) TRIA(3,5,10)
91%
Part A
TRIA(1,3,4)
Part B
NORM(2.4, 0.5)
20%
Scrapped
80%
Salvaged
And
Shipped
Shipped
Model 2
The arrival and preparation segments of the process to be modeled
are no different from the arrival and processing segments of
Model1. The “Sealer” and “Rework” modules are new - and require
an “Inspector” entity for their implementation, since they will
determine whether a part is faulty or not. We will also need to
model different processing times (different distributions of
processing times) depending on the part being processed.
A further change appears because we introduce a “transfer time”
from one station to another (parts don’t move instantaneously) which we had not done before. We will use the “Route” rather than
the “Connect” option.
Model 2
We start, as before, with File/New, click on Attach… and choose
Common.tpo. Then lay down the two arrival modules, the two
servers, the two inspectors, the three departure modules, and a
simulate one. You may have to “zoom in” to be able to see the
whole layout. Notice we have not added any information or
connections, yet.
Model 2
We will introduce the notions of Stations, Station Transfers and
Pictures. A “Station” is simply a location where something occurs.
Each station has a unique name. The Station Transfer is where
the notion of Route will come in. We start by double-clicking on
the “Arrive” modules and filling in the details.
Enter Data
Station: Part A Arrive
Arrival Data
Time Between
EXPO(5)
Mark Time Attribute Arrival Time
Leave Data
Station
Part A Prep
Route Time
2
Model 2
Click on Assign… to bring up a new dialog box. Click on Add…, to
bring up a third one.
In this last one
Attribute
Value
Sealer Time
TRIA(1,3,4)
Check the next slide to see all the dialog boxes and their contents.
Make sure you exit each dialog box by clicking OK… otherwise the
changes will not occur. If you forget, just re-open and go through
it again.
Model 2
The dialog boxes
for the arrival of
Part A: when
you click OK in
the Assignments
dialog, the list in
Assign will
change…
Model 2
Double-click on the second Arrival station, and fill in:
Enter Data
Station
Part B Arrive
Arrival Data
Batch Size
4
Time Between
EXPO(30)
Mark Time Attribute Arrival Time
Leave Data
Station
Part B Prep
Route Time
2
Assignments
Attribute
Sealer Time
Value
NORM(2.4, 0.5)
Model 2
The dialog boxes
for the arrival of
Part B…
When done (all
Ok), double-click
on the upper of
the two Servers.
You will see the
dialog in the
next slide.
Model 2
The dialog box has a
default station name
(Server 1) and a list
of all the known
station names including the ones
you have typed but
not yet assigned to
any station.
Drop the drop list and pick the Part A Prep name. You could type it
in, but you would have to make absolutely sure that what you type
matches the name already there…
Model 2
Right before you hit
OK, the dialog box
should look like this.
You will then need to
open the dialog box
for the second
server, and choose
Part B Prep as
Station; TRIA(1,4,8)
for the Server Data
Process Time; Sealer
as the Station in the Leave Data area; with a Route Time of 2.
Model 2
The Sealer dialog
should be filled as
indicated. Note that
we have a Failure
Probability, a station
to go to on passing,
and another one to
go to on failing.
Notice that we
have not specified the Sealer Time distribution, since each type of
part has a different distribution, which has already been assigned,
under that name, in “Arrive”.
Model 2
The Rework module
comes next, with
appropriate entries.
Notice the Process
Time, the Failure
Probability and the
two Leave Data
areas, both with
Route Times of 2,
and next station
names.
Model 2
We choose the lowest
“Depart” module as
the “Shipping”
module, and fill in as
indicated on this
slide. Click OK to
complete it. The
second Depart from
the bottom becomes
the “Salvaged”
module, while the top
one is the “Scrap”.
All three are filled in the same way, except for choosing the three
different names from the drop-down list for “Station”.
Model 2
Fill in the dialog box for the “Simulate”
module - 2000 minutes of model run.
Now click on the Route button in the
“Animate” toolbar at the right of the
model design area.
You’ll se the dialog:
Model 2
Once you have clicked OK, the cursor changes to a “cross-hair”,
and you can click, starting in a station, clicking placement corners,
and ending in another station. You may have to repeat the Route
button choice for each route segment you want to set up. Rightclick (repeat action) might work, too.
Model 2
You are now ready to run the model - use the “tape player”
buttons, and try out “single-step”, “fast-forward”, “pause”, etc.
At the end you will be asked whether you want to see the data
collected - say Yes and look at the results, which have to be
interpreted. There is little of significance in a single replication:
the results are only “plausible” and “obtainable”. In order to be
able to make statistical validity claims you will need many
replications - with different random number streams.
You will see two sets of variables: Tally Variables and Discrete
Change Variables, and various statistics derived from the single
run, plus some Counters. Let’s go through them.
Model 2
Tally Variables and Discrete Change Variables give average, 95%
confidence-interval half-width, and minimum and maximum
observed values.
The half-width interval is not given unless at least 320
observations have been taken for the Tallies and at least 5 time
units have elapsed for the Discrete Changes. If this is not the
case, Arena will report (Insuf). If this test is passed, then a check
for data correlation is carried out. If independence is violated,
then (Corr) will appear. If all tests are passed, then a half-interval
is computed. No such appeared on our report…
The only meaningful conclusion relates to the capacity of the
Rework station: what is it?
Model 2
Exercises. (From K. S. & S., Ch. 5)
1) Travelers arrive at the main airline door of an airline terminal
according to an exponential interarrival-time distribution with
mean 1.6 minutes. The travel time between the entrance and the
check-in is distributed uniformly between 2 and 3 minutes. At the
check-in counter travelers wait in a single line until one of five
agents is available to serve them. The check-in time follows a
normal distribution with mean of seven minutes and standard
deviation of 2 minutes. Upon completion of the check-in, they are
free to travel to their gates. Create a simulation model, with
animation, of this system. Run the simulation for 16 hours, to
determine the average time in system, number of passegers
completing check-in, and the average length of the check-in
queue.
2) Develop a model of a simple serial two-process system. Items
arrive at the system with a mean time between arrivals of 10
minutes. They are immediately sent to Process 1, which has an
Model 2
unlimited queue and a single resource with a mean service time of
9 minutes. Upon completion, they are sent to Process 2, which is
identical to Process 1. Items depart the system upon completion of
Process 2. Performance measures of interest are the average
numbers in queue at each process and the system cycle time.
Using a replication length of 10,000 minutes, make the following
four runs and compare the results:
Run 1: exponential inter-arrival times and exponential service
times.
Run 2: constant inter-arrival times and exponential service times.
Run 3: exponential inter-arrival times and constant service times.
Run 4: constant inter-arrival times and constant service times.