E211 – Operations Planning II

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Finding the Bottleneck

Recall the utilization versus waiting time:
 The higher the utilization of a process, the longer the
queue in front of that process

In Flexsim, the utilization statistics can be
obtained by examining the ‘state’ result.

However, it is not so straightforward if the arrival
is batched. When there is no item to be processed,
obviously there is no utilization! This is not due to
bottlenecks occurring upstream.

When the current bottleneck operation is alleviated,

other operations with higher utilization may become
bottleneck again.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Study of Cycle Time Improvement

Study each measure one at a time. For
example, adding one more operator/forklift;
adding one more packaging station, etc.

Document the improvement in cycle time and
cost of implementing the measure.

If significant improvement is sought, may have
to combine two or more measures.

Basically a trial-and-error optimization process!

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Simulation Techniques Learnt

Specify exact
arrival of items
using ‘Arrival
Schedule’ under
arrival style

Specify
batch
size for
each
arrival
Specify type of
item for each
arrival

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Simulation Techniques Learnt

After items are

generated from
the source:

To differentiate two
types of items, set a
unique attribute for
each type.

Notify VisualTool of
the entry of each
item into the
processes.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Simulation Techniques Learnt

When modeling with

mobile resources
(operators,
transporters), pay
attention to the
speed, distance as
well as time units.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Simulation Techniques Learnt
When using more than 1 operator (transporter) for one
route, use a dispatcher or the ‘direct’ Technique(using
the <A> key).

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

Simulation Techniques Learnt
To route different
item types to
different objects,
specify a ‘send to
port’ rule under
flow tab of object.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Available/Collected Information from
Running Simulation
Conveyor
Storage ArrivalTime ItemName ItemType Quantity
Packaging Rack
Operator
(Running & (Running) 0 Running shoes 2 30
Trekking)

Common
900 Trekking shoes 1 40
Production Common
Holding Area Storage Bin
2100 Running shoes 2 30
Storage
Packaging Rack
3900 Trekking shoes 1 40
(Running & (Trekking)
Trekking) 4200 Running shoes 2 30
Conveyor
6300 Running shoes 2 30

6900 Trekking shoes 1 40

Forklift
truck 8400 Running shoes 2 30
Setup time at Triangular (5, 15, 10) seconds for both 9900 Trekking shoes 1 40
packaging station types of shoes
10500 Running shoes 2 30
Packaging for
12600 Running shoes 2 30
Trekking shoes Normal ~ (60, 25) seconds
Packaging for 12900 Trekking shoes 1 40

Running Shoes Normal ~ (45, 15) seconds 14700 Running shoes 2 30

Trekking Shoes Batch of 40 products every 50 minutes, 6 15900 Trekking shoes 1 40
Schedule batches daily 16800 Running shoes 2 30
Running Shoes Batch of 30 products every 35 minutes, 18900 Running shoes 2 30
Schedule 10 batches daily

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Current Process Flow

Packaging Station Utilization

Average Cycle Time (based on 10

replications): 1451.7 seconds (24
minutes 12 seconds)

Packaging Station has higher utilization rate,

and is more likely to be the bottleneck
Operator Utilization

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 New Process Flow: Three Test Stations with
One Operators

Packaging Station Utilization

Average Cycle Time (based on 10

replications): 681.3 seconds (11
minutes 21 seconds)

The Operator has now higher

utilization, and is more likely
to be the bottleneck.
Operator Utilization

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 New Process Flow: Three Test Stations with
Two Operators

Packaging Station Utilization

Average Cycle Time (based on 10

replications): 610.7 seconds (10
minutes 11 seconds)

Packaging Station now has

higher utilization, and is more
likely to be the bottleneck
Operator Utilization

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Statistical Comparison: 2-Sample t-test

How can you statistically prove that the
Cycle Time (Seconds)
proposed alternative can improve cycle
3 QC stations + 1 time by at least 700 seconds?
Current
operator
Use 2-Sample t-test under the following
1412.78 675.73 assumptions:
 Collected data are independent of one
1541.9 694.68 another;
1424.99 681.22  Each of the two samples (current and
improved) is drawn from normally
1280.21 661.21 distributed population;
 The two populations have unequal
1592.13 700.21 variance
1475.32 680.04
Statistical hypotheses
 Null hypothesis (H0): The difference in
1501.98 689.56 cycle time between current and improved
method is less and equal to 700 seconds;
1456.35 668
 Alternative hypothesis (H1): The
1376.8 679.67 difference in cycle time between current
and improved method is greater than 700
1454.24 682.75 seconds.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Statistical Comparison: 2-Sample t-test

Excel>Data>Data Analysis>t-Test: Two-Sample Assuming Unequal Variances.

The improvement on cycle time (between current and

Level of significance proposed alternative) you want to statistically prove

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Statistical Comparison: 2-Sample t-test

t-Test: Two-Sample Assuming Unequal Variances

Variable 1 Variable 2  Compare t-statistic

Mean 1451.67 681.307
with critical value: t-
Variance 7579.743711 135.8573344 statistic > critical
Observations 10 10 value.

Hypothesized  Or compare p-value

Mean Difference 700 with significance level:
df 9 p-value < 0.05.
t Stat 2.533141595  Conclusion: Reject H0
P(T<=t) one-tail 0.016034819 and conclude that the
proposed alternative
t Critical one-tail 1.833112923
can improve cycle time
P(T<=t) two-tail 0.032069637 by at least 700
t Critical two-tail 2.262157158 seconds.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

 Conclusion
Configuration Average Cycle Time (based on 10
replications)
Current 1451.7 seconds (24 minutes 12 seconds)
3 packaging stations with 1 operator 681.3 seconds (11 minutes 21 seconds)
3 packaging stations with 2 operators 610.7 seconds (10 minutes 11 seconds)

Packaging station is the bottleneck of the current packaging and stock-in

process.

However the bottleneck operation also may have higher percentage of idle
time. This is because arrivals are batched; sometimes resources may be idle
in between the arrival batches.

By improving the bottleneck operation (add in one more packaging station

and/or operator), average cycle time can be improved.

By statistical comparison, such as 2-Sample t-test, we can conclude whether

the proposed alternative can effectively improve the cycle time.

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II

Learning Objectives

Modeling and simulation of a typical production line using
a simulation software.

Identify bottleneck operations in the process flow through
simulation modeling and analysis.

Recommend how to alleviate bottleneck operations
(improve cycle time); and how to simulate proposed
alternatives using Flexsim.

Use statistical test to conclude whether the proposed
alternative can effectively alleviate the bottleneck
(improve the cycle time).

SCHOOL OF ENGINEERING E211 – OPERATIONS PLANNING II