The Corrupting Influence of Variability

When luck is on your side, you can do without brains.

– Giordano Bruno,burned
at the stake in 1600

The more you know the luckier you get.

– “J.R. Ewing” of Dallas

1
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Performance of a Serial Line
Measures: Links to Business Strategy:
• Throughput • Would inventory reduction
• Inventory (RMI, WIP, FGI) result in significant cost
• Cycle Time savings?
• Lead Time • Would CT (or LT) reduction
• Customer Service result in significant
competitive advantage?
• Quality
• Would TH increase help
generate significantly more
Evaluation: revenue?
• Comparison to “perfect” • Would improved customer
values (e.g., rb, T0) service generate business over
• Relative weights consistent the long run?
with business strategy? Remember – standards
change over time!
2
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Capacity Laws

Capacity Law: In steady state, all plants will release work

at an average rate that is strictly less than average
capacity.

Utilization Law: If a station increases utilization without

making any other change, average WIP and cycle time
will increase in a highly nonlinear fashion.

Notes:
• Cannot run at full capacity (including overtime, etc.)
• Failure to recognize this leads to “fire fighting”

3
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Cycle Time vs. Utilization
24

22

20

18

16
Cycle Time (hrs)

14

12
High
10
Variability
8

6
Low
4 Variability
Capacity
2

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2
Release Rate (entities/hr)
4
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 What Really Happens:
System with Insufficient Capacity
700

600

500

400
WIP

300

200

100

0
0 5 10 15 20 25 30 35 40 45
Day
5
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 What Really Happens:
Two Cases with Releases at 100% of Capacity

120 120

100 100

80 80
WIP

WIP
60 60

40 40

20 20

0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Day Day

6
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 What Really Happens:
Two Cases with Releases at 82% of Capacity

120 120

100 100

80 80
WIP

WIP
60 60

40 40

20 20

0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Day Day

7
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Overtime Vicious Cycle

1. Release work at plant capacity.

2. Variability causes WIP to increase.
3. Jobs are late, customers complain,…
4. Authorize one-time use of overtime.
5. WIP falls, cycle times go down, backlog is
reduced.
6. Breathe sigh of relief.
7. Go to Step 1!

8
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Mechanics of Overtime Vicious Cycle
50
CT without
45 Overtime

40

35
Cycle Time (hrs)

30

25

20

15
Original
10 Capacity

5 CT with
Capacity
Overtime
with Overtime
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5
Release Rate (entities/hr)
9
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Influence of Variability

Variability Law: Increasing variability always degrades the

performance of a production system.

Examples:
• process time variability pushes best case toward worst case
• higher demand variability requires more safety stock for same
level of customer service
• higher cycle time variability requires longer lead time quotes to
attain same level of on-time delivery

10
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Variability Buffering

Buffering Law: Systems with variability must be buffered by

some combination of:
1. inventory
2. capacity
3. time.

Interpretation: If you cannot pay to reduce variability, you will pay

in terms of high WIP, under-utilized capacity, or reduced customer
service (i.e., lost sales, long lead times, and/or late deliveries).

11
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Variability Buffering Examples

Ballpoint Pens:
• can’t buffer with time (who will backorder a cheap pen?)
• can’t buffer with capacity (too expensive, and slow)
• must buffer with inventory
Ambulance Service:
• can’t buffer with inventory (stock of emergency services?)
• can’t buffer with time (violates strategic objectives)
• must buffer with capacity
Organ Transplants:
• can’t buffer with WIP (perishable)
• can’t buffer with capacity (ethically anyway)
• must buffer with time

12
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Lean Production

Definition: Production of goods or services is lean if it is

accomplished with minimal buffering costs.

Insights:
• Lean is about more than waste elimination.
• Variability reduction is key to lean.
• Choosing the right mix of inventory, capacity and time
buffering is also important.

13
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Simulation Studies
TH Constrained System (push)

1 2 3 4
ra, ca B(1)= te(1), ce(1) B(2)= te(2), ce(2) B(3)= te(3), ce(3) B(4)= te(4), ce(4)

WIP Constrained System (pull)

Infinite
raw 1 2 3 4
materials
te(1), ce(1) B(2) te(2), ce(2) B(3) te(3), ce(3) B(4) te(4), ce(4)

ra arrival rate
ca CV of interarrival times
te (i ) effective process time at station i
ce (i ) effective CV at station i
B (i ) buffer size in front of station i
14
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Push Systems
Case te(i), te(3) c(i), i = 1-4 TH CT WIP CT Comments
i = 1, 2, 4 (min) (unitless) (j/min) (min) (jobs) (min)
(min)
1 1 1.2 0 0.8 4.2 3.4 0.0 best case
2 1 1.2 1 0.8 44.6 35.7 26.8 WIP buffer
3 1 1.0 1 0.8 20.0 16.0 10.3 capacity buffer
4 1 1.2 0.3 0.8 7.8 6.2 3.3 reduced
variability
Notes:
• ra = 0.8, ca = ce(i) in all cases.
• B(i) = , i = 1-4 in all cases.
Observations:
• TH is set by release rate in a push system.
• Increasing capacity (rb) reduces need for WIP buffering.
• Reducing process variability reduces WIP for same TH, reduces CT for
same TH, and reduces CT variability.

15
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Pull Systems
Case te(i), te(3) c(i), i = 1-4 B(3) TH CT WIP CT Comments
i = (min) (unitless) (jobs) (j/min) (min) (jobs) (min)
1,2,4
(min)
1 1 1.2 0 0 0.83 4.6 3.8 0.0 best case
2 1 1.2 1 0 0.48 6.4 3.1 2.4 plain JIT
3 1 1.2 1 1 0.53 7.2 3.8 2.6 inv buffer
4 1 1.2 0.3 0 0.72 5.0 3.6 0.6 var reduction
5 1 1.2 0.3 1 0.76 6.0 4.5 0.8 inv buffer +
var reduction
6 1 1.2 0.3 0 0.73 6.3 4.6 0.7 non-bottleneck
buffer

Notes:
• Station 1 pulls in job whenever it becomes empty.
• B(i) = 0, i = 1, 2, 4 in all cases, except case 6, which has B(2) = 1.
16
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Pull Systems (cont.)

Observations:
• Capping WIP without reducing variability reduces TH.
• WIP cap limits effect of process variability on WIP/CT.
• Reducing process variability increases TH, given same buffers.
• Adding buffer space at bottleneck increases TH.
• Magnitude of impact of adding buffers depends on variability.
• Buffering less helpful at non-bottlenecks.
• Reducing process variability reduces CT variability.
Conclusion: consequences of variability are different in push
and pull systems, but in either case the buffering law implies
that you will pay for variability somehow.

17
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Example – Discrete Parts Flowline

process
buffer process buffer
process
Inventory Buffers: raw materials, WIP between processes, FGI
Capacity Buffers: overtime, equipment capacity, staffing
Time Buffers: frozen zone, time fences, lead time quotes

Variability Reduction: smaller WIP & FGI , shorter cycle times

18
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Example – Batch Chemical Process

reactor reactor reactor

tank tank
column column column

Inventory Buffers: raw materials, WIP in tanks, finished goods

Capacity Buffers: idle time at reactors
Time Buffers: lead times in supply chain

Variability Reduction: WIP is tightly constrained, so target is

primarily throughput improvement, and maybe FGI reduction.
19
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Example – Moving Assembly Line
in-line buffer

fabrication lines
final assembly line

Inventory Buffers: components, in-line buffers

Capacity Buffers: overtime, rework loops, warranty repairs
Time Buffers: lead time quotes

Variability Reduction: initially directed at WIP reduction, but

later to achieve better use of capacity (e.g., more throughput)
20
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Buffer Flexibility

Buffer Flexibility Corollary: Flexibility reduces the amount

of variability buffering required in a production system.

Examples:
• Flexible Capacity: cross-trained workers
• Flexible Inventory: generic stock (e.g., assemble to order)
• Flexible Time: variable lead time quotes

21
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Variability from Batching

VUT Equation:
• CT depends on process variability and flow variability

Batching:
• affects flow variability
• affects waiting inventory

Conclusion: batching is an important determinant of

performance

22
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Process Batch Versus Move Batch

Dedicated Assembly Line: What should the batch size be?

Process Batch:
• Related to length of setup.
• The longer the setup the larger the lot size required for the same
capacity.

Move (transfer) Batch: Why should it equal process batch?

• The smaller the move batch, the shorter the cycle time.
• The smaller the move batch, the more material handling.
Lot Splitting: Move batch can be different from process batch.
1. Establish smallest economical move batch.
2. Group batches of like families together at bottleneck to avoid setups.
3. Implement using a “backlog”.
23
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Process Batching Effects

Types of Process Batching:

1. Serial Batching:
• processes with sequence-dependent setups
• “batch size” is number of jobs between setups
• batching used to reduce loss of capacity from setups

2. Parallel Batching:
• true “batch” operations (e.g., heat treat)
• “batch size” is number of jobs run together
• batching used to increase effective rate of process

24
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Process Batching

Process Batching Law: In stations with batch operations or

significant changeover times:
1. The minimum process batch size that yields a stable
system may be greater than one.
2. As process batch size becomes large, cycle time grows
proportionally with batch size.
3. Cycle time at the station will be minimized for some
process batch size, which may be greater than one.

Basic Batching Tradeoff: WIP versus capacity

25
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Serial Batching

k ts
setup t0
ra,ca

forming queue of
batch batches

Question: what is the impact of batch size k?

26
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Serial Batching

Parameters:
k serial batch size (10)
t time to process a single part (1)
s time to perform a setup (5)
ce CV for batch (parts  setup) (0.5)
ra arrival rate for parts (0.4)
ca CV of batch arrivals (1.0)

Time to process batch: te = kt + s

27
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Serial Batching Effects (cont.)

Arrival of batches: ra/k

Utilization: u = (ra/k)(kt + s) = ra(t + s/k )

sra
For stability: u < 1 requires k  minimum batch size required
1  tra for stability of system...

28
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Serial Batching Effects (cont.)

Average queue time at station: Note: we assume

arrival CV of
 ca2  ce2  u   1  0.5   0.6  batches is ca
CTq  
  1  u  e  2   1  0.6 15 16.875
 t 
 2  regardless of
batch size – an
approximation...
Average cycle time depends on move batch size:
• Move batch = process batch
CTnon  split CTq  t e CTq  s  kt
16.875  15 31.875
Note: splitting move
batches reduces wait
• Move batch = 1 k 1 for batch time.
CTsplit CTq  s  t
2
10  1
16.875  10  (1.0) 27.375
2
29
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Cycle Time vs. Batch Size – 5 hr setup
80.00

70.00

60.00
Cycle Time (hrs)

50.00

40.00

30.00

20.00

10.00

0.00
0 5 10 15 20 25 30 35 40 45 50
Optimum
Batch Size (jobs/batch)
Batch Sizes
No Lot Splitting Lot Splitting

30
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Cycle Time vs. Batch Size – 2.5 hr setup
80.00

70.00

60.00
Cycle Time (hrs)

50.00

40.00

30.00

20.00

10.00

0.00
0 5 10 15 20 25 30 35 40 45 50
Optimum
Batch Size (jobs/batch)
Batch Sizes
No Lot Splitting Lot Splitting

31
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Setup Time Reduction

Where?
• Stations where capacity is expensive
• Excess capacity may sometimes be cheaper

Steps:
1. Externalize portions of setup
2. Reduce adjustment time (guides, clamps, etc.)
3. Technological advancements (hoists, quick-release, etc.)

Caveat: Don’t count on capacity increase; more flexibility will

require more setups.

32
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Parallel Batching

ra,ca k t

forming queue of jobs processed

batch batches in batches

Question: what is the impact of batch size k?

33
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Parallel Batching
Parameters:
k parallel batch size (10)
t time to process a batch (90)
ce CV for batch (1.0)
ra arrival rate for parts (0.05)
ca CV of batch arrivals (1.0)
B maximum batch size (100)
k11
Time to form batch: W 
2 ra

Time to process batch: t e =t

34
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Parallel Batching (cont.)

Arrival of batches: ra/k

Utilization: u = (ra/k)(t)

minimum batch size required

For stability: u < 1 requires k  ra t
for stability of system...

35
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Parallel Batching (cont.)

Average wait-for-batch time:

batch size affects
k  1 1 10  1 1 both wait-for-batch
WT   90
2 ra 2 0.05 time and queue time
Average queue plus process time at station:

 ca2 / k  c02  u   0.1  1   0.45 

CT   
 1  u  t  t    90  90 130.5
 2   2   1  0.45 

Total cycle time:

CT  WT 90  130.5 220.5

36
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
Cycle Time vs. Batch Size in a Parallel Operation

1400.00 queue time due to

wait for batch time
utilization
1200.00
Total Cycle Time

1000.00

800.00

600.00

400.00

200.00

0.00
0 10 20 30 40 50 60 70 80 90 100 110
Optimum Nb B
Batch Size
37
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Variable Batch Sizes

Observation: Waiting for full batch in parallel batch operation may

not make sense. Could just process whatever is there when operation
becomes available.

Example:
• Furnace has space for 120 wrenches
• Heat treat requires 1 hour
• Demand averages 100 wrenches/hr
• Induction coil can heat treat 1 wrench in 30 seconds
• What is difference between performance of furnace and coil?

38
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Variable Batch Sizes (cont.)
Furnace: Ignoring queueing due to variability
• Process starts every hour 100
• 100 wrenches in furnace 50
• 50 wrenches waiting on average
• 150 total wrenches in WIP
• CT = WIP/TH = 150/100 = 3/2 hr = 90 min

Induction Coil: Capacity same as furnace (120 wrenches/hr), but

• CT = 0.5 min = 0.0083 hr
• WIP = TH × CT = 100 × 0.0083 = 0.83 wrenches

Conclusion: Dramatic reduction in WIP and CT due to small batches

—independent of variability or other factors.
39
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching

Move Batching Law: Cycle times over a segment of a routing

are roughly proportional to the transfer batch sizes used over that
segment, provided there is no waiting for the conveyance device.

Insights:
• Basic Batching Tradeoff: WIP vs. move frequency
• Queueing for conveyance device can offset CT reduction from
reduced move batch size
• Move batching intimately related to material handling and layout
decisions

40
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching

ra,ca(1) te(1),ce(1) k te(2),ce(2)

Station 1 Station 2

single job
Question: What is the impact
batch of batch size k?

41
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching

Problem:
• Two machines in series
• First machine receives individual parts at rate ra with CV of ca(1)
and puts out batches of size k.
• First machine has mean process time of te(1) for one part with CV
of ce(1).
• Second machine receives batches of k and put out individual parts.
• How does cycle time depend on the batch size k?

42
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching Calculations

Time at First Station:

• Average time before batching is:
ca2 (1)  ce2 (1) u (1) regular VUT
te (1)  te (1)
2 1  u (1) equation...

• Average time forming the batch is:

k11 k1 first part waits (k-1)(1/ra),
 te (1) last part doesn’t wait, so
2 ra 2u (1)
average is (k-1)(1/ra)/2
• Average time spent at the first station is:
ca2 (1)  ce2 (1) u (1) k1
CT (1)  te (1)  te (1)  te (1)
2 1  u (1) 2u (1)
k1
CT(1, no batching)  te (1)
2u (1)
43
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching Calculations (cont.)

Output of First Station:

• Time between output of individual parts into the batch is ta.
because
• Time between output of batches of size k is kta.
cd2(1)=d2/ta2
• Variance of interoutput
2
times
2
of
2
parts is2 c2d2(1)ta2, where by def of CV
cd (1) (1  u (1) )ca (1)  u (1) ce (1)

because departures are

• Variance of batches of size k is kcd (1)ta . independent, so variances add
2 2

• SCV of batch arrivals to station 22 is:2

kcd (1)t a variance divided by
ca2 (2) 
k 2t a2 mean squared...
cd2 (1)

k

44
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching Calculations (cont.)
Time at Second Station: independent
process times...
• Time to process a batch of size k is kte(2).
• Variance of time to process a batch of size k is kce2(2)te2(2).
2 2 2
kc
• SCV for a batch of size k is: e ( 2 ) t e ( 2 ) c e ( 2)

k 2 t e2 (2) k

• Mean time spent in partial batch of size k is: first part doesn’t wait,
k1
t e (2) last part waits (k-1)te(2),
2 so average is (k-1)t (2)/2 e

• So, average time spent at the second station is:

2 2
cd (1) / k  ce (2) / k u (2) k1
CT (2)  kt e (2)  t e ( 2)  t e ( 2)
2 1  u ( 2) 2
VUT equation to
compute queue time k1
CT(2, no batching )  t e ( 2)
of batches... 2
45
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Move Batching Calculations (cont.)

Total Cycle Time:

k1 k1
CT(batching) CT(no batching)  te (1)  te (2)
2u (1) 2
 k  1   te (1) 
CT(no batching)      te (2) 
 2   u (1) 

inflation factor due

Insight: to move batching
• Cycle time increases with k.
• Inflation term does not involve CV’s
• Congestion from batching is more bad control than randomness.

46
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Assembly Operations

Assembly Operations Law: The performance of an assembly

station is degraded by increasing any of the following:
1. Number of components being assembled.
2. Variability of component arrivals.
3. Lack of coordination between component arrivals.

Observations:
• This law can be viewed as special instance of variability law.
• Number of components affected by product/process design.
• Arrival variability affected by process variability and production
control.
• Coordination affected by scheduling and shop floor control.
47
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Improving Performance

Key Questions:

• What causes poor performance?

• What levers exist for improvement?

• How to prioritize policies?

48
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 The Easy Answer – Waste

• Rate: anything that slows pace of line

– Unnecessary steps
– Errors
– Material handling
– Etc.

• Time: anything that limits utilization of line

– Downtime
– Setups
– Etc.

49
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Impact of Reducing Waste at Bottleneck

Improved

Best Case
Original
Throughput

Actual
Improved

Original

WIP
50
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Impact of Reducing Waste at Nonbottleneck

Best Case

Improved Original
Throughput

Improved Actual

Original

WIP 51
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 The Subtle Answer – Variability

• Process Variability: disruptions at processes

– Failures
– Setups
– Operator Variability
– Etc.

• Flow Variability: disruption in arrivals

– Scheduling/Flow Control
– Propagation of Process Variability

52
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Impact of Reducing Variability

Best Case

Reduced Variability
Throughput

Original

Worst Case

WIP
53
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Implementation Strategy

Continual Improvement Cycle

Substitute
Eliminate Reduce
capacity for Reduce
direct capacity
inventory variability
waste buffers
buffers

Phase 0: Phase 1: Phase 2: Phase 3:

54
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Basic Improvement Framework

• Look at capacity first.

• Consider variability second.

• Address structural issues:

– Human considerations
– Layout
– Batching
– Etc.

55
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Attacking Variability

Objectives
• reduce cycle time
• increase throughput
• improve customer service

Levers
• reduce variability directly
• buffer using inventory
• buffer using capacity
• buffer using time
• increase buffer flexibility

56
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Cycle Time

Definition (Station Cycle Time): The average cycle time

at a station is made up of the following components:
cycle time = move time + queue time + setup time + delay times
process time + wait-to-batch time + typically
make up
wait-in-batch time + wait-to-match time
90% of CT

Definition (Line Cycle Time): The average cycle time in a

line is equal to the sum of the cycle times at the individual
stations less any time that overlaps two or more stations.

57
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Reducing Queue Delay

CTq = V U t

 ca2  ce2



 u 
 2   
   1 u 

Reduce Variability Reduce Utilization

• failures • arrival rate (yield, rework, etc.)
• setups • process rate (speed, time,
• uneven arrivals, etc. availability, etc)

58
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Reducing Batching Delay

CTbatch = delay at stations + delay between stations

Reduce Process Batching Reduce Move Batching

• Optimize batch sizes • Move more frequently
• Reduce setups • Layout to support
– Stations where capacity is material handling (e.g.,
expensive cells)
– Capacity vs. WIP/FT tradeoff

59
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Reducing Matching Delay

CTbatch = delay due to lack of synchronization

Reduce Variability Improve Coordination Reduce Number

• High utilization • scheduling of Components
fabrication lines • pull mechanisms • product redesign
• Usual variability • modular designs • kitting
reduction methods

60
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Increasing Throughput

TH = P(bottleneck is busy)  bottleneck rate

Reduce Blocking/Starving Increase Capacity
• buffer with inventory (near • add equipment
bottleneck) • increase operating time (e.g.
• reduce system “desire to queue” spell breaks)
• increase reliability
CTq = V U t • reduce yield loss/rework

Reduce Variability Reduce Utilization

Note: if WIP is limited, then system degrades
via TH loss rather than WIP/CT inflation
61
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Customer Service

Elements of Customer Service:

• lead time
• fill rate (% of orders delivered on-time)
• quality

Law (Lead Time): The manufacturing lead time for a routing

that yields a given service level is an increasing function of
both the mean and standard deviation of the cycle time of the
routing.

62
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Improving Customer Service

LT = CT + z CT

Reduce CT Visible Reduce Average CT Reduce CT Variability

to Customer • queue time generally same as methods
• delayed • batch time for reducing average CT:
differentiation • match time • improve reliability
• assemble to order • improve maintainability
• stock components • reduce labor variability
• improve quality
• improve scheduling
• etc…

63
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Cycle Time and Lead Time
0.18

0.16 CT = 10 Lead Time = 14 days

CT = 3
0.14

0.12
Densities

0.1
CT = 10 Lead Time = 27 days
0.08 CT = 6
0.06

0.04

0.02

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Cycle Time in Days
64
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Diagnostics Using Factory Physics
Situation:
• Two machines in series; machine 2 is bottleneck
• ca2 = 1
• Machine 1: t 0 19 min
c02 0.25
MTTF 48 hr, MTTR 8 hr
• Machine 2: t 22 min
0

c02 1
MTTF 3.3 hr, MTTR 10 min
– Space at machine 2 for 20 jobs of WIP
• Desired throughput 2.4 jobs/hr, not being met

65
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Diagnostic Example (cont.)

Proposal: Install second machine at station 2

• Expensive
• Very little space
Analysis Tools: ca2  ce2 u
CTq  te VUT equation
2 1 u
cd2 u 2 ce2  (1  u 2 )ca2 propogation equation
Analysis: Ask why five times...
Step 1: At 2.4 job/hr
• CTq at first station is 645 minutes, average WIP is 25.8 jobs.
• CTq at second station is 892 minutes, average WIP is 35.7 jobs.
• Space requirements at machine 2 are violated!

66
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Diagnostic Example (cont.)

Step 2: Why is CTq at machine 2 so big?

• Break CTq into
 ca2  ce2   u 
CTq     te (3.16)(12.22)(23.11 min)
 2  1  u 
• The 23.11 min term is small.
• The 12.22 correction term is moderate (u  0.9244)
• The 3.16 correction is large.

Step 3: Why is the correction term so large?

• Look at components of correction term.
• ce2 = 1.04, ca2 = 5.27.
• Arrivals to machine are highly variable.
67
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Diagnostic Example (cont.)

Step 4: Why is ca2 to machine 2 so large?

• Recall that ca2 to machine 2 equals cd2 from machine 1, and
cd2 u 2 ce2  (1  u 2 )ca2 (0.887 2 )(6.437)  (1  0.887 2 )(1.0) 5.27
• ce2 at machine 1 is large.

Step 5: Why is ce2 at machine 1 large?

• Effective CV at machine 1 is affected by failures,
mr
ce2 c02  2 A(1  A) 0.25  6.18 6.43
t0

• The inflation due to failures is large.

• Reducing MTTR at machine 1 would substantially improve
performance.

68
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Procoat Case – Situation

Problem:
• Current WIP around 1500 panels
• Desired capacity of 3000 panels/day
• Typical output of 1150 panels/day
• Outside vendor being used to make up slack

Proposal:
• Expose is bottleneck, but in clean room
• Expansion would be expensive
• Suggested alternative is to add bake oven for touchups

69
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Procoat Case – Layout

Loader Clean Coat 1 Coat 2 Unloader

IN
Touchup
D&I
Unloader Bake Inspect Develop Loader

Manufacturing Inspect Expose

Clean Room

© Wallace J. Hopp, Mark L. Spearman, 1996, 2000

OUT [Link] 70
 Procoat Case – Capacity Calculations
Process Std Dev
or Load Process Conveyor Number
Machine Time Time Trip Time of Setup Rate Time
Name (min) (min) (min) Machines MTTF MTTR Avail Time (p/day) (min)
Clean1 0.33 0 15 1 80 4 0.95 0 3377 37
Coat1 0.33 0 15 1 80 4 0.95 0 3377 37
Coat2 0.33 0 15 1 80 4 0.95 0 3377 37
Expose 103 67 - 5 300 10 0.97 15 2879 118
Develop 0.33 0 2.67 1 300 3 0.99 0 3510 23
Inspect 0.5 0.5 - 2 - - 1.00 0 4680 1
Bake 0.33 0 100 1 300 3 0.99 0 3510 121
MI 161 64 - 8 - - 1.00 0 3488 161
Touchup 9 0 - 1 - - 1.00 0 7800 9
2879 542

rb = 2,879 p/day
T0 = 542 min = 0.46 days
W0 = rbT0 = 1,334 panels
71
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Procoat Case – Benchmarking

TH Resulting from PWC with WIP = 37,400:

w 1,500 Higher than

TH  rb  2,879 1,524
w  W0  1 1,500  1,334  1 actual TH

Conclusion: actual system is significantly worse than PWC.

Question: what to do?

72
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Procoat Case – Factory Physics Analysis

1) Bottleneck Capacity
- rate:
- time:

2) Bottleneck Starving
- process variability:
- flow variability:

73
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Corrupting Influence Takeaways

Variance Degrades Performance:

• many sources of variability
• planned and unplanned

Variability Must be Buffered:

• inventory
• capacity
• time

Flexibility Reduces Need for Buffering:

• still need buffers, but smaller ones

74
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Corrupting Influence Takeaways (cont.)

Variability and Utilization Interact:

• congestion effects multiply
• utilization effects are highly nonlinear
• importance of bottleneck management

Batching is an Important Source of Variability:

• process and move batching
• serial and parallel batching
• wait-to-batch time in addition to variability effects

75
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Corrupting Influence Takeaways (cont.)

Assembly Operations Magnify Impact of Variability:

• wait-to-match time
• caused by lack of synchronization

Variability Propagates:
• flow variability is as disruptive as process variability
• non-bottlenecks can be major problems

76
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]