The Theory of Constraints
Fundamental Exam Review
Applications: DBR Segment
James R. Holt, Ph.D., PE
Professor
jholt@wsu.edu
http://www.engrmgt.wsu.edu/
Engineering & Technology
© Washington State University-2010
1
Management
TOCICO Segmented Fundamentals Exam
Fundamentals Certificate
Multiple Choice Exam
(Identify, Exploit, Subordinate, Elevate, Go to Step 1)
Fundamentals
Certificate of
TOC Philosophy
Fundamentals
Certificate of
TOC Thinking
Processes
•Inherent Potential
•Inherent
Simplicity
•Inherent Win-Win
•Five Focusing
Steps
•Three Questions
•Conflict
Cloud
•Negative
Branch
•Ambitious
Target
Fundamentals
Fundamentals
Certificate of
Certificate of
TOC Applications
TOC Finance &
Measures
•DBR
•Project
Management
•T, I, OE
•PQ Type
Problem
•Replenishment
© Washington State University-2010
2
Topics in TOC Operations
•
•
•
•
•
Locating Capacity Constrained Resource
Sizing Buffer (50% of existing to start)
Rope Control (release)
Buffer Management
Simplified- DBR (finished goods buffers
controlling release)
•
Make to Order
•
Make to Stock
© Washington State University-2010
3
TOC and Physical Process Flows
•
Market
Process
A
B
C
D
E Request
11
FG
RM
Capability
Parts
7
per Day
9
5
FG
8
6
•
•
•
RM
I previously introduced the linear flow
line at left. This type of flow is often
called an “I” line. “I” standing for a
vertical line as shown at the bottom.
Raw material flows from Raw Material
at the bottom upward to Finished
Goods at the top.
In reality, there are few such simple
lines making a single commodity.
But the “I” form itself is common.
Consider: A trip to the hospital (you
are the RM and FG). You go to the
office for processing, see the doctor,
pay, go get the prescription, go
home, get better. I hope it doesn’t
take too long.
© Washington State University-2010
4
Moving the Constraint
FG
• I mentioned, “You can choose the constraint.” How is that
•
•
•
•
RM
possible?
If we progress normally, we find the constraint (Step 1)
exploit the constraint (Step 2), subordinate everything else
to the constraint (Step 3), elevate the constraint (Step 4),
and if the constraint is eliminated, find the new constraint
(Step 5).
If we elevate the constraint far enough, we to move it.
Actually, any one of the processes in our “I” line can be
improved. Any one process can increase its capacity. So,
by picking and choosing which processes to improve
(elevate) we can move the constraint to the location where
we want it to be.
Ah! This creates an interesting question. Where is the best
place to have the constraint? At the middle as shown? At
the beginning (bottom) or at the top (end)?
© Washington State University-2010
5
There are many factors in choosing the
best location for the constraint.
FG
•
•
?
RM
Retail sales, traditional management wants the constraint to be in
the market (buffer Finished Goods) *. If you are selling diamonds,
you probably want the constraint to be in the Raw Material. Hum?
Two selling markets with opposite recommendations? What I am I
do do?
Things to consider.
 Capacity is obviously an issue. How hard is it to get? What does it
cost? Do I want the constraint to be the $10,000 Drill or the $100,000
Mill?
 Predictability is important. You want a constraint that produces 5±5
or 5±1? Which helps you plan ahead?
 Quality and scrap are an issue as well. After the constraint, we want
zero scrape. Resolve quality issues before the constraint.
 Ease of expansion. Try to locate the constraint where you can easily
surge to capture unexpected opportunities
 Location of Inventory is key. You will have buffers. Try to locate
them naturally. Cost of inventory is often a real issue (use raw
material costs).
 Control is the most important factor in locating the inventory. Who is
in control and how responsive is the system?
*TOC tries to avoid this option
© Washington State University-2010
6
Choosing the Constraint for Control
Purposes
FG
•
•
?
•
•
RM
Who is in control if the constraint is in Raw Material? Not
you. Who is in control if the constraint is in the Market?
Not you. Those are both bad choices. If you are stuck in
that world, you should take action to put yourself back in
control (Read Its Not Luck to learn how to create
unrefusable offers to increase your market).
If the constraint is at the beginning, are you more or less
responsive to customer demand? Probably Less. It takes
time to move the product through the system.
If the constraint is at the end, are you more or less
responsible to the customer demand? Could be less if the
material in your buffer is not the right item. You have to
clean out your buffer before you can get the right part to
the customer.
It seems there is no right answer. Well there is, but the
answer depends a lot upon the physical system you are
dealing with. But, we can make some general statements.
© Washington State University-2010
7
Consider the “V” Plant
•
FG FG FG FG FG
•
•
V-Plant
RM
•
•
In a “V” plant structure, relatively few Raw Materials are
used to make a wide variety of Finished Goods. Molten
steel, once rolled into a thickness can be cut into widths,
and treated. But, once it’s 1/2 inch thick and 12 inches
wide, it is very hard to be 2 inches thick and 36 inches
wide.
Oil is refined into many types. Metal cut, formed, drilled
coated is not usually returned to its Raw Material State.
The main problem in “V” Plants is stealing. Individual
processes, in an effort to improve efficiencies or reduce
set-ups will produce more than ordered or change the
production schedule to ‘look better’ locally. This problem
creates havok with too much inventory in some places and
too little in other places.
In the “V” Plant, it is much better to have the control at the
beginning. Buffer an early process and then flow the
product quickly to the Finished Goods.
The “V” Plant unit measure CAN NOT BE QUANTITY
PRODUCED/TIME!!!!!!!!!!!!!!
© Washington State University-2010
8
Consider the “A” Plant
•
•
FG
•
A-Plant
•
RM
RM
RM
RM
RM
In the “A” Plant structure, many Raw Materials
combine to produce few Finished Goods. This is
mostly an assembly process.
The main problem in “A” plants is with coordination.
There are a lot of assembly operations. If one of the
items is not immediately available for an assembly,
the assembly is delayed.
There is better control if the constraint is at the apex
of the A. There are fewer buffers, better response
time and less inventory over all.
It is important to
 Value inventory at the Raw Material cost
(Aluminum sheet with a bunch of holes is worth
less to an outsider than a solid sheet -- not more.
A partly processed part has no extra value until it
is sold.)
 Choke off release of Raw Material to just the
amount needed to replenish buffers. (Excess
inventory delays work, creates quality problems,
prevents improvements.)
© Washington State University-2010
9
Consider the “T” Plant
•
FG FG FG FG FG
•
T-Plant
•
RM
In the “T” Plant structure, work flows
through a common line to a point where
the common parts are ‘customized’ into
specialized Finished Goods.
The main problem in a “T” plant is
having enough of the right materials at
the point where the common
components become assembled (or
become defined) as the different
Finished Goods.
In such a line, it makes sense to control
the flow at the specialization position.
You have the fastest response yet
minimize inventory.
© Washington State University-2010
10
Hybrids “X” and “Diamonds”
•
FG
FG
FG
FG
FG
FG
•
RM
RM
RM
RM
RM
RM
•
There are hybrid combinations
of plant of course.
Looking at the “X”, with many
raw materials brought together
into a single process and then
forming many other products
(abrasive paper for instance)
seems to have an obvious
control point at the center.
The “Diamond” shape should
probably have its constraint
near the end to avoid
cluttering up the complex
internal processes.
© Washington State University-2010
11
Working With Your Plant
• It is not always obvious what type plant you have.
•
•
The
buildings and material flow aren’t always laid out in an
“I” or “A” or “V” or “T” pattern.
It is a theoretical plant that we are discussing. Look at
the conceptual structure. Learn how to function in all
the theoretical structures and then you will be able to
formulate the best policy for your system.
Hum? Is college an “I” Plant, a “A” Plant or an “V”
Plant? That depends. If you are becoming a nurse, its
probably an “I”. If you are studying semi-conductors,
its probably an “A”. If you are studying management, it
is probably a “V”. Well, that is an interesting puzzle.
© Washington State University-2010
12
Let’s Play a Game!
The Dice Game-Set Up
•
Processes (move token
across the line)
Storage areas for Workin-Process
•
•
•
A B C D E F
RM
FG
Storage of Raw
Material and Finished
Goods
•
I want you to simulate the ‘match
game’ found in The Goal, pages
104-112. But we will do it in a
more fun way.
You will need six fair dice (one die
will do if you share).
Gather some tokens (lima beans,
marbles, or toothpicks work well.
Note: Jelly beans and chocolate
chips tend to disappear)
Get two cups, one for Raw
Material and one for Finished
Goods
Arrange your play area as on the
left. Create Six Processes (A, B,
C, D, E, F) between Raw Material
and Finished Goods.
© Washington State University-2010
13
The Play
•
•
•
A B C D E F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
14
The Play
•
Day 1.
A Rolls a 5
•
•
A B C D E F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
15
The Play
•
Day 1.
A Rolls a 5, B Rolls a 3
•
•
A B C D E F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
16
The Play
•
Day 1.
A Rolls a 5, B Rolls a 3, C Rolls a
3
•
•
A B C D E F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
17
The Play
•
Day 1.
A Rolls a 5, B Rolls a 3, C Rolls a
3, D Rolls 6
•
•
A BB CC DD EE F F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
18
The Play
•
Day 1.
A Rolls a 5, B Rolls a 3, C Rolls a
3, D Rolls 6, E Rolls a 2
•
•
A B C D E F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
19
End of Day One
•
Day 1.
A Rolls a 5, B Rolls a 3, C Rolls a
3, D Rolls 6, E Rolls a 2, F Rolls 1
and puts it in the cup.
•
•
A B C D E F
RM
FG
•
You play the game for ten days, one day
at a time(each process gets to roll a die
ten times-once each day-in the order
prescribed).
The work progresses from left to right.
Each day, A rolls first, takes however
many tokens are rolled from the RM cup
and moves them across the line to the
WIP (Work in Process) location between
Line A and B.
B then rolls and tries to move tokens
from the left WIP location between A and
B to the right WIP location between B
and C, if available.
Click a few times and I’ll show you the
sequence.
© Washington State University-2010
20
The Play Continues
•
•
•
•
A B C D E F
RM
FG
Day 2 starts with the WIP (tokens in
places left from Day 1) and continues in
the same fashion.
Starting Day 2, worker A rolls, takes
from the RM cup and moves more
tokens to the space between A and B.
Each process takes it’s turn moving
tokens during the day.
The last process F ends the day by
rolling an moving the number of tokens
shown on the die (if they are available)
into the FG cup.
All players move the maximum of what
they roll or what inventory is available to
process for that day’s roll.
© Washington State University-2010
21
The Results
•
•
?
A B C D E F
RM
FG
•
•
Now, we play for ten days. How
many tokens will be in the
Finished Goods Cup? Take a
guess before you play.
Each die averages 3.5 dots per
roll. There are 10 rolls. Each
process should produce 35 dots
on the average over ten days.
Do you expect 35 in the FG Cup?
To make sure we understand
what is happening in this game,
lets keep some records. For one,
we need to make sure we are
using Fair Die.
© Washington State University-2010
22
Keeping Track of our Rolls
•
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
Day 8
Day 9
Day 10
Process
A
B
C
5
3
3
2
4
2
1
5
3
and so on …
D
6
5
2
E
2
1
4
F
1
4
6
__
__
__
__
__
__
Sum Totals for each Process
•
•
•
•
•
Keep track of the rolls you make.
Make sure each process (die)
produces as the company expects-close to 35.
In fact, lets reward the workers.
Any process who rolls better than 35
dots in 10 days receives a “Superior
Performance Award”
Any process who does better than 40
receives a “Sustained Superior
Performance Award”
But, if any process rolls less than 25
dots, we will have to let them go.
Your Team will probably earn a few
Awards.
© Washington State University-2010
23
Keep Track of Your Inventory
•
•
When you finish the Dice Game, you will have some unexpected results. We need to
produce at least 30 items to make a profit. The plant has such a bad history, one more 10
day period of loss (deliveries below 30) and we will have to close the plant.
So, you’ll probably want to keep records of your inventory so you can figure out what
happened. Record the WIP (Work in Process) remaining between each workstation at the
end of each day as you go. Similar to below:
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
Day 8
Day 9
Day 10
Process
A Wip B Wip
5
2 3 0
2
0 4 2
1
0 5 0
and so on …
C Wip D Wip E
3 0 6 1 2
2 0 5 2 1
3 1 2 0 4
__ __ __ __ __ __ __
Sum Totals for each Process
Wp
1
0
0
F
1
4
6
FG FG Total
1
1
2
3
4
7
You see the WIP changes as the
days go by. If you want, you can
calculate the average WIP for
each position to see who is the
problem.
Knowing the WIP at the End of the
10 periods also helps you
understand your system.
© Washington State University-2010
24
The Report
•
•
•
•
•
Play the Dice Game for ten days. Record the individual Process
results. Determine which of your Processes received which
awards (or were laid off).
Some of you are pretty concerned about my letting a person go if
they produce less than 25 parts. But, this is a very rare case.
Rolling a fair die multiple times soon generates a near normal
distribution. There is less than a 5% chance of producing less than
25 dots in ten rolls.
Besides, this is more than generous in light of our production
demands. Cooperate has told us, “If you don’t produce at least 30
tokens in that FG Cup, they are going to close the plant!”
What we want is everyone to do good so we can keep the plant
open.
Make any recommendations you want on how to improve the plant.
© Washington State University-2010
25
Not Too Good!
• Let’s try again.
• Clearly, you didn’t have enough
•
?
A B C D E F
RM
•
FG
•
Work-In-Process Inventory to
avoid he devastating impact of the
combination of Variability and
Interdependence (which occur in
almost every system).
Let’s start with Six tokens at each
work center and play for ten days.
(AH, much better)!
Now, without starting over,
continue to play for another ten
days (How was production then?)
What would happen if we go ten
more days? Hum?
© Washington State University-2010
26
More on Dice Games
• Some of you will find the Dice Game fascinating.
•
•
You
will play it over and over and dumbfound your friends.
That’s ok, but don’t blame me if your social life goes
to pot.
If you can’t get enough, you can try a few more
options I’ve documented at:
http://www.wsu.edu/~engrmgmt/holt/em530/Docs/DiceGames.htm
© Washington State University-2010
27
Looking at DBR from Several Points of
View
• By now you have:
 Overview of DBR and some theory
 Played the Dice Game
 You’ve Read The Goal at least once (3 or 4
times is probably enough)
 We are comfortable with PQ (exploiting
Octane)
 What’s left? Don’t we know it all?
© Washington State University-2010
28
Understanding DBR
• You still have some work to do.
• We need to understand Buffer Management
• What is that?
• Five Focusing Steps:
• 1. Identify the Constraint
• 2. Decide how to Exploit the Constraint
• 3. Subordinate all else to
Buffer
the above decision
Management
• 4. Elevate the Constraint
• 5. Warning, Warning. If the constraint moves, start
over at step 1. © Washington State University-2010
29
Review--Where are the Buffers?
• There are four buffers to date:
FG
8
6
7
•
5
•
Work Flow
4
1
2
3
RM RM
RM
 Constraint Buffers
 Raw Material Buffers
 Shipping Buffers
 Assembly Buffers
Where is each one on this process
flow?
Do you know know why?
Note: The implementation of the Assembly Buffer is confusing
in many situations. And, its purpose is weak.
The current thinking leans towards eliminating the Assembly
buffer in production.
Then, the Shipping Buffer is the time from material release to
finished goods deliver and subsumes the Constraint Buffer.
© Washington State University-2010
30
Review: Where are the buffers?
FG 13
FG 14
11
12
7
8
10
9
5
6
Work Flow
4
1
2
3
RM RM
RM
• Locate the Four
Buffers
 Constraint Buffers
 Raw Material
Buffers
 Shipping Buffers
 Assembly Buffers
• Why?
Why not just have one
Assembly Buffer at 5?
© Washington State University-2010
31
Review: Where are the buffers?
Market Constraint
90% of
the time
15 FG
 Constraint
 Raw Material
 Shipping
 Assembly
• Where (when) does a
buffer start? When
does it end?
14
13
12
11
10
9
8
Work Flow
• Locate Buffers
6
7
1
2
3
4
5
RM RM RM RM RM
© Washington State University-2010
32
Review: Where are the buffers?
Market Constraint
FG
1
FG
2
FG
3
FG
4
12
11
10
9
Work Flow
8
7
6
FG
5
• Locate Buffers
 Constraint
 Raw Material
 Shipping
 Assembly
• Where (when) does a
buffer start? When
does it end?
13
RM
© Washington State University-2010
33
A Buffer is ‘Time’
• Time for what?
How big should the buffer be?
Process A
1
1
1
2
5
RM
Process B
2
3
Constraint
2
2
5
RM
Constraint
Process C
4
4.5
4
5
2
RM
Constraint
© Washington State University-2010





Safety?
Cushion?
Variability?
Confidence?
Comfort?
Much more buffer is
needed when nonconstraints are near
the constraint
capacity.
HOW TO DECIDE?
34
The Starting Position
To make this work,
the WIP Inventory
must be at least 5
per work center (30)
•
•
?
A B C D E F
RM
FG
•
We experienced in the Dice
Game the devastating Effects
when Interdependence AND
Variability both exist in the
system (ALMOST EVERY
SYSTEM).
The only way to ever hope to get
production anyway near the
average capacity of the
processes is to have Work-InProcess Inventory everywhere,
and Lots of it!
And yet, this solution degrades
to pretty poor performance over
time.
© Washington State University-2010
35
The DBR Approach
To make this work,
the WIP Inventory
must be at least 5
per work center (30)
•
•
?
A B C D E F
RM
FG
•
First, Identify the Constraint. For
this case, let’s make a
constraint.
Modify D’s die. That is, let’s
change the dotes. When D rolls,
if D rolls a 5, change it to a 4. If
D rolls a 6, change it to a 4. So
the new Die looks like,
1,2,3,4,4,4. This is 18 dots
spread over six sides for an
average of 3/side.
Increase the capacity for A,B,C,
E and F to 4/side by just ignoring
the side with the 1. Just assume
the die is five sided: 2, 3, 4, 5, 6.
for 20 dots or average of 4/side.36
© Washington State University-2010
The DBR Approach
To make this work,
the WIP Inventory
must be at least 5
per work center (30)
•
•
?
RM
A B C D E F
4 4 4 3 4 4
X
Constraint
FG
•
First, Identify the Constraint. For
this case, let’s make a
constraint.
Modify D’s die. That is, let’s
change the dotes. When D rolls,
if D rolls a 5, change it to a 4. If
D rolls a 6, change it to a 4. So
the new Die looks like,
1,2,3,4,4,4. This is 18 dots
spread over six sides for an
average of 3/side.
Increase the capacity for A,B,C,
E and F to 4/side by just ignoring
the side with the 1. Just assume
the die is five sided: 2, 3, 4, 5, 6.
for 20 dots or average of 4/side.37
© Washington State University-2010
Buffer Sizing
When is WIP
Inventory a
problem?
•
•
?
RM
A B C D E F
4 4 4 3 4 4
X
Constraint
FG
If there is Not Enough WIP, we
can’t get production volumes.
If there is too much, we lose
control, delay the flow and chaos
occurs.
The Bath Tub Curve WIP Curve
Problems!
•
Anywhere in here is OK. To
start, just Cut WIP (typical
flow times in half)
Too Little
Much
© Washington State University-2010
Too
38
Buffer Sizing
When is WIP
Inventory a
problem?
•
•
?
RM
A B C D E F
4 4 4 3 4 4
X
Constraint
FG
If there is Not Enough WIP, we
can’t get production volumes.
If there is too much, we lose
control, delay the flow and chaos
occurs.
The Bath Tub Curve WIP Curve
Problems!
•
Anywhere in here is OK. To
start, just Cut WIP (typical
flow times in half)
Too Little
Much
© Washington State University-2010
Too
39
DBR Approach
Cutting WIP from 30
to 15 is safe, but…
• Third, Subordination.
• Where should the WIP Be Held?
• We need to protect the
•
Constraint from starving.
Let’s move the WIP to where it
helps.
?
RM
A B C D E F
4 4 4 3 4 4
X
FG
Constraint
© Washington State University-2010
40
DBR Approach
Cutting WIP from 30
to 15 is safe, but…
• Third, Subordination.
• Where should the WIP Be Held?
• We need to protect the Constraint
•
•
•
?
RM
A B C D E F
4 4 4 3 4 4
X
Constraint
FG
•
•
from starving.
Move the WIP to where it helps.
That should do it.
But, We have so much Safety in
our Buffer. And A can out produce
D. We need to prevent OVER
PRODUCTION!
Don’t allow A to produce any more
than D produces. Keep Buffer at or
below 15.
Tie the Rope
© Washington State University-2010
41
Tie the Rope (of DBR)
• Third, Subordination.
• Don’t allow A to produce any
Control the Buffer
size with the Rope
more than D produces. Keep
Buffer at or below 15.
• Each day A examines the WIP.
?
RM
A B C D E F
4 4 4 3 4 4
X
Constraint
•
FG
•
If there are 15 in the Buffer,
DO NOT WORK. If less than
15, work just enough to
replace the Buffer to 15
Pull WIP into the system Only
at the Rate the Constraint
Produces.
What is the average outcome?
© Washington State University-2010
42
Shape of the Buffer
If all products take the same time on the
constraint, we can count parts. Say buffer
is 10 parts. What does it look like?
Buffer Time
RM
Constraint
© Washington State University-2010
• Where should
the buffer
materials be
located?
• What is
protecting
what?
All three views
are possible
views of the
Constraint
Buffer.
43
Replenishment is the Key
What Minimum Buffer Size
would you recommend for
each product?
Market Constraint
Demand
A
B
C
Consider
market
constraints for
products A, B,
C, D, E with
weekly demand
shown
D
E
0 10 20 30 40 50
Per Week
© Washington State University-2010
44
Replenishment is the Key
What Minimum Buffer Size
would you recommend for
each product?
Market Constraint
Demand
A
B
C
Consider
market
constraints for
products A, B,
C, D, E with
weekly demand
shown
D
E
0 10 20 30 40 50
Per Week
© Washington State University-2010
45
Replenishment is the Key
What Minimum Buffer Size
would you recommend for
each product?
Market Constraint
Demand
A
B
C
If Replenishment time is
two weeks!
D
If Replenishment time is
one week!
E
If Replenishment time is
one day!
Consider
market
constraints for
products A, B,
C, D, E with
weekly demand
shown
0 10 20 30 40 50
Per Week
Ropes?
© Washington State University-2010
46
Buffer Reduction Efforts
Solution called S-DBR
Could Adding a Constraint
help?
Market Constraint
Demand
A
B
C
• What can we
do to reduce
buffer
sizes?
• Focus on
Replenishment!
D
50 100
150
Aggregate Constraint
Demand
E
0 10 20 30 40 50
Per Week
Ropes?
© Washington State University-2010
47
Think about the Ropes
• Ropes Pull the
A
B
Product.
• Which Rope Pulls
What?
C
D
E
© Washington State University-2010
48
Consider Possible Advance Warning
Sell an A, Constraint Produces A,
Release Raw Material for an A.
A
Sell a C, Constraint Produces C,
Release Raw Material for a C.
B
Sell a D, Constraint Produces D,
Release Raw Material for an D.
C
Ropes are communication!
Does the communication go
to the Constraint only?
Sell a B, Constraint Produces B,
Release Raw Material for an B.
D
E
Sell an A, Constraint Produces A,
Release Raw Material for an A.
And so on…
Ropes can pull whatever is smart! Read the Precision Power Case on WebCT.
© Washington State University-2010
49
More about the sequencing
of the Work in the Buffer
A
B
C
D
Using DBR Replenishing the FG
Inventory, the production line should
produce as much as possible in the
sequence ordered. Reduce Set-up
Times to make this possible.
Avoid Large Batches that will choke
the balancing of many products.
E
Order Sold:
2 As, 3 Es, 2Cs, 1 A, 10 Bs, 1 A, 2 D, 2 Es, 3 Bs, 2 Es, 2 As, 2 Cs, 4 Ds…
Pure Order Sequence in order sold
AAEEECCABBBBBBBBBBADDEEBBBEEAACCDDDD …
Adjusted Order Sequence to make (Breaking large B batch in two)
AAEEEBBBBBCCADDBBBBBEEBBBEEAACCDDDD …
© Washington State University-2010
50
Now Managing the Buffer?
Adjusted Order Sequence to make
AAEEEBBBBBCCAADDBBBBBEEBBEEAAC...
Buffer
First order
Release Work Order:
…CAAEEBBEEBBBBBDDAACCBBBBBEEEAA
…CAEEB
If the Buffer
was a hopper
filling right to
left, bottom to
top...
BEEB
BBBB
DDAA
CCBB
BBBE
EEAA
Pull the next Work Order from
the bottom of the hopper
Order to the Constraint.
© Washington State University-2010
51
Now Let’s look at the Buffer Correctly As
Time
So, what does this hopper represent?
It is the order, but no time is connected.
What if I spaced the Products out
according to how much time it takes
on the constraint?
…CAEEB
BEEB
BBBB
DDAA
CCBB
BBBE
EEAA
© Washington State University-2010
• Time On
Constraint
 A: 15 Min
 B: 10 Min
 C: 20 Min
 D: 15 Min
 E: 5 Min
52
Convert Letters to Time
…CAAEEB
BEEB
BBBB
DDAA
CCBB
BBBE
EEAA
Third Hour
-> :
Since 5 minutes is the common
denominator, I’ll use 5 mins
for each Letter and add dashes
of 5 mins to space out the
letters in time sequence.
D--D--
-> :
A--A--
The Second hour -> :
C---C-
On hour and half -> :
B-B-B-
Second half hour -> :
B-BEEE
The first half hour on
the constraint is then:
--A--A
Second + half
Time On
Constraint
A: 15 Min
B: 10 Min
C: 20 Min
D: 15 Min
E: 5 Min
This stack will be
tall. But, it is important to
represent the ‘Time blocks’
within the Buffer.
Scheduling the constraint
© Washington State University-2010
53
Creating a 9 Hour
Buffer
…CAAEEB
BEEB
BBBB
DDAA
CCBB
BBBE
EEAA
Since scheduling the constraint is the most
important thing about managing DBR, we
need to do this. Nine working hours before it
is due at the constraint, work for a product is
released.
© Washington State University-2010
--C--D
AE---A
-B-B-B
EE---C
---C-D--D-D--D--C---C
A--AEE
-B-BEE
-B-B-B
-B-B-D--D-A--A-C---CB-B-BB-BEEE
--A--A
Time on
Constraint
A: 15 Min
B: 10 Min
C: 20 Min
D: 15 Min
E: 5 Min
Each line is
an half hour
54
Rearranging the Buffer Stack
That tall stack was unwieldy. But, if we split it into
three smaller stacks, it’s not so oppressive.
Think of the flow moving through a set of tanks so the
Jobs stay in order from beginning to end.
Constraint Buffer
--C--D
AE---A
-B-B-B
EE---C
---C-D--D--
D--D--C---C
A--AEE
-B-BEE
-B-B-B
-B-B--
© Washington State University-2010
D--D-A--A-C---CB-B-BB-BEEE
--A--A
55
Buffer Zones
With the Buffer Split in three sections, we can name the
zones. The one closest to the constraint RED. The
middle one YELLOW. The zone farthest away from the
constraint is Green.
Constraint Buffer
--C--D
AE---A
-B-B-B
EE---C
---C-D--D--
D--D--C---C
A--AEE
-B-BEE
-B-B-B
-B-B--
© Washington State University-2010
D--D-A--A-C---CB-B-BB-BEEE
--A--A
56
We can take a different view of the Buffer
A fully loaded
buffer
CDDD
DCCA
AEEB
EAB
BBE
CD
A
E
--C--D
AE---A
-B-B-B
EE---C
---C-D--D--
BE
EBB
BBB
B
D--D--C---C
A--AEE
-B-BEE
-B-B-B
-B-B--
DAAC
CBBBB
EEEAA
D
D--D-A--A-C---CB-B-BB-BEEE
--A--A
© Washington State University-2010
57
The Parts in Zones
CDDD
DCCA
AEEB
EAB
BBE
CD
A
E
--C--D
AE---A
-B-B-B
EE---C
---C-D--D--
BE
EBB
BBB
B
D--D--C---C
A--AEE
-B-BEE
-B-B-B
-B-B--
DAAC
CBBBB
EEEAA
D
D--D-A--A-C---CB-B-BB-BEEE
--A--A
© Washington State University-2010
58
What if there were problems?
EAB
XBBE
X
CD
A
RM
Freight Truck
problems so RM
released partial
parts over time.
Missing 8 Bs!
CDDD
DCCA
AEEB
E
--C--D
AE---A
X
X
-B-B-B
EE---C
---C-D--D--
BE
XX
EBB
BBB
X
B
D--D--C---C
A--AEE
-B-BEE
X
-B-B-B
X
X-B-B--
DAAC
CBBBB
XX X
EEEAA
D
D--D-A--A-C---CXB-B-BX X
B-BEEE
--A--A
© Washington State University-2010
59
An Easier Way to Envision Buffer Status
I’ll replace the missing Pieces in the buffer with a hole.
What happens if?
Green Zone Hole?
Not to worry.
RM
Yellow Zone Hole?
Watch out-Learn
--C--D
AE---A
X-B-B-B
X
EE---C
---C-D--D--
D--D--C---C
A--AEE
-B-BEE
X
-B-B-B
X
X-B-B--
Red Zone Hole?
Act Immediately! ACT! ACT!
D--D-A--A-C---CX B-B-BX X
B-BEEE
--A--A
© Washington State University-2010
60
What is the allowable
probability of a Zone Hole?
Green Zone
60%
Yellow Zone
30%
Red Zone
5%
UnAcceptable
Prob of Parts
In the Zone
Acceptable Region
Just Released
RM
--C--D
AE---A
-B-B-B
EE---C
---C-D--D--
<-Time Away
D--D--C---C
A--AEE
-B-BEE
-B-B-B
-B-B--
© Washington State University-2010
At Constraint
D--D-A--A-C---CB-B-BB-BEEE
--A--A
61
The Buffer Zone Concept
Works for ALL Buffers
Constraint Buffers, Finished Goods Buffers, Supply
Buffers, Assembly Buffers are reported the same
way.
Green Zone
60%
Yellow Zone
30%
Red Zone
5%
Prob of Parts
In the Zone
UnAcceptable
Acceptable Region
At Release
Point
<-Time Away
At Protection Point
© Washington State University-2010
62
So, what about Buffers?
• The are critical
• They are not optional
• They are integrated with the Constraint and the
Rope
• Buffers are the KEY MANAGEMENT TOOL for
DBR.
• We can overcome bad Statistics, terrible
Interdependency, common Human Behavior
© Washington State University-2010
63
DBR and
Human Behavior Issues
• Drum-Buffer-Rope overcomes the problems of system
structure
 DBR de-couples interdependency
 DBR allows variability to work in our favor
 Protects the throughput capacity of the system
• Measurements should be in place to encourage the right
behaviors
 Maintain Buffers, Increase throughput, Reduce
variability, …
 Focus on Flow; Getting Faster and Faster!
© Washington State University-2010
64
<Add Make to Order vs. Make to Stock
Element about here>
© Washington State University-2010
65
What About Multi Operators?
Demand
100 parts/hr
RM
A
B
C
D
17 min
/part
26 min
/part
22 min
/part
Staffing 34
29
44
37
32
Flow
1.7
p/min
1.7
p/min
1.7
p/min
1.7
p/min
20 min
/part
1.7
p/min
© Washington State University-2010
E
FG
19 min
/part
66
Based Upon Buffers, Where Should the
Constraint Be?
FG
• The RM is Very Expensive?
• There is lots of Scrap in the
Processes?
• A Very Expensive Machine?
• We have highly variable processes?
 Locate at the least variable.
• The Market is sporadic?
RM
© Washington State University-2010
67
Physical Flow Processes
The Goal: Produce.
The Measure: Throughput
The Constraint: Internal or Market
Process
RM
Capability
Parts
per Day
A
B
C
D
E
7
9
5
8
6
© Washington State University-2010
Market
Request
11
FG
68
Physical Flow Processes
The Conflict Cloud:
Process
A
R
The Paradigm Shift: We have
M
Efficiency at the Constraint,
Capability
Parts
per Day
Buffer Delivery, Watch Costs.
B
Market
Request
D11
C
FG
7
9
5
B. Effective use
of Resources
D. Focus on
Efficiency/Cost
C. Effective
Delivery
D’. Ignore
Efficiency/Cost
8
6
A. Excellent
production
© Washington State University-2010
69
The Behavior/Results
The Drum & Buffer Exploit the Constraint
Process
The Buffer and Rope Subordinate
(de-couple and motivate) the System
Capability
Processes
Parts
A
B
Market
Request
D11
C
R
M
FG
7
9
5
8
6
per Day
Common Measures (flow time) Create Teamwork.
Buffer Management measures Buffer Penetration
Red-> Immediate Action
What do we learn
Yellow-> Learn about what to fix next
here to apply to
Green-> Allow the system to run
Daily Lives?
Constraint Focus typically results in 20% improvement.
Road Runner Ethic, Throughput Accounting, Continuous
Improvement, Culture Change.
Very effective when Touch time<< Flow time.
© Washington State University-2010
70
Physical Flow
Lessons Learned
• Look at the Whole--Your whole life
• Select the Limiting Factor (Self-Examination)
• Do the Best You Can
• Subordinate secondary wishes to the Goal
• Simple, Cooperative Measures Lead to Better
Behavior
• Higher levels of Achievement are Possible by
Improving our Current Limiting Factors
• As You Improve, Continue to Get Better
• Be Anxiously Engaged in Good Causes (doing
and supporting)
© Washington State University-2010
71
Strategy and Tactics Tree
• Rapid Reliable Response
• http://www.wsu.edu/~engrmgmt/holt/em534/SandTRRR.pdf
• The Rapid Reliable Response S&T Tree
 Improves operations flow
 Focuses on selling Value
 Includes valuing Rapid Response
© Washington State University-2010
72