Basic Factory Dynamics

Physics should be explained as simply as possible,

but no simpler.

– Albert Einstein

1
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 HAL Case

Large Panel Line: produces unpopulated printed circuit boards

Line runs 24 hr/day (but 19.5 hrs of productive time)

Recent Performance:
• throughput = 1,400 panels per day (71.8 panels/hr)
• WIP = 47,600 panels
• CT = 34 days (663 hr at 19.5 hr/day)
• customer service = 75% on-time delivery

Is HAL lean?
What data do we need to decide?
If not, what are levers for improvement?
2
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 HAL - Large Panel Line Processes
Lamination (Cores): press copper and prepreg into core blanks
Machining: trim cores to size
Internal Circuitize: etch circuitry into copper of cores
Optical Test and Repair (Internal): scan panels optically for defects
Lamination (Composites): press cores into multiple layer boards
External Circuitize: etch circuitry into copper on outside of composites
Optical Test and Repair (External): scan composites optically for defects
Drilling: holes to provide connections between layers
Copper Plate: deposits copper in holes to establish connections
Procoat: apply plastic coating to protect boards
Sizing: cut panels into boards
End of Line Test: final electrical test

3
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 HAL Case - Science?

External Benchmarking
• but other plants may not be comparable

Internal Benchmarking
• capacity data: what is utilization?
• but this ignores WIP effects

Need relationships between WIP, TH, CT, service!

4
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Basic Definitions
Workstation: a collection of one or more identical machines.
Part: a component, sub-assembly, or an assembly that moves through
the workstations.
End Item: part sold directly to customers; relationship to constituent
parts defined in bill of material.
Consumables: bits, chemicals, gasses, etc., used in process but do not
become part of the product that is sold.
Routing: sequence of workstations needed to make a part.
Order: request from customer.
Job: transfer quantity on the line.

5
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Basic Measures

Throughput (TH): for a line, throughput is the average quantity of good

(non-defective) parts produced per unit time.
Work in Process (WIP): inventory between the start and endpoints of a
product routing.
Raw Material Inventory (RMI): material stocked at beginning of
routing.
Crib and Finished Goods Inventory (FGI): crib inventory is
material held in a stockpoint at the end of a routing; FGI is material
held in inventory prior to shipping to the customer.
Cycle Time (CT): time between release of the job at the beginning of the
routing until it reaches an inventory point at the end of the routing.

6
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Factory Physics

Definition: A production system is a goal-oriented network of

processes and stockpoints through which parts flow.

Structure: Network is made up of routings (lines), which in turn are

made up of processes.

Focus: Factory Physics is concerned with the network and flows at

the routing (line) level.

7
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Parameters

Descriptors of a Line:
1) Bottleneck Rate (rb): Rate (parts/unit time or jobs/unit time)
of the workstation with the highest long-term utilization.

2) Raw Process Time (T0): Sum of the long-term average

process times of each station in the line.

3) Congestion Coefficient (): A unitless measure of

congestion.
• Zero variability case,  = 0. Note: we won’t use  quantitatively,
• “Practical worst case,”  = 1. but point it out to recognize that lines
• “Worst possible case,”  = W0. with same rb and T0 can behave very
differently.
8
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Definition of Bottleneck

Why do we use highest utilization instead of slowest to define

the bottleneck? rate in
u
capacity
1 min 2 min
r
r A B u (1)  r
y 1
r y
u ( 2)  2ry
1-y 0 .5
u (1) u (2) if y 0.5

If yield loss is greater than 50% then station 1 becomes the

bottleneck because it processes more jobs than station 2. The same
thing happens with systems that have multiple routings.
9
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 A Simple Relationship

Critical WIP (W0): WIP level in which a line having no

congestion would achieve maximum throughput (i.e., rb) with
minimum cycle time (i.e., T0).

W0 = rb T0

10
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 The Penny Fab

Characteristics:
• Four identical tools in series.
• Each takes 2 hours per piece (penny).
• No variability.
• CONWIP job releases.

Parameters:
rb =
T0 =
W0 =
 =

11
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 The Penny Fab

WIP TH CT TH x CT
1
2
3
4
5
6
7
8
9
12
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 TH vs. WIP: Best Case

.5
Throughput (Jobs/hr)

.4

.3

.2

.1

0
0 2 4 6 8 10 12 14

WIP (Jobs)

13
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 CT vs. WIP: Best Case

20
Cycle time (Hours)

16

12

0
0 2 4 6 8 10 12 14

WIP (Jobs)
14
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Best Case Performance

Best Case Law: The minimum cycle time (CTbest) for a given
WIP level, w, is given by
T0 , if w W0
CTbest ( w) 
w / rb , otherwise.

The maximum throughput (THbest) for a given WIP level, w is

given by,
w / T0 , if w W0
TH best ( w) 
 rb , otherwise.

15
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Best Case Performance (cont.)

Example: For Penny Fab, rb = 0.5 and T0 = 8, so W0 = 0.5  8 =

4,
8, if w 4
CTbest ( w) 
2 w, otherwise.

w / 8, if w 4
TH best ( w) 
0.5, otherwise.

which are exactly the curves we plotted.

16
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 A Manufacturing Law

Little's Law: The fundamental relation between WIP, CT, and

TH over the long-term is:
WIP TH CT

parts
parts  hr
hr

Insights:
• Fundamental relationship
• Simple units transformation
• Definition of cycle time (CT = WIP/TH)

17
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Penny Fab Two

2 hr
5 hr 3 hr

10 hr

18
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Penny Fab Two

Station Number of Process Station

Number Machines Time Rate
1 1 2 hr j/hr
2 2 5 hr j/hr
3 6 10 hr j/hr
4 2 3 hr j/hr

rb = ____________ T0 = ____________ W0 = ____________

19
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Worst Case

Observation: The Best Case yields the minimum cycle time and
maximum throughput for each WIP level.

Question: What conditions would cause the maximum cycle time

and minimum throughput?

Analysis:
• throughput cannot be worse than one job in empty line
• use Little’s law to get worst possible cycle time
• Results are exactly what you get if every job in the line is moved
in a single batch between stations.

20
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Worst Case Performance
0.6
Best Case
rb
Throughput: TH  1/T0 0.5
0.4

TH
0.3
0.2 Worst Case
1/T0 0.1
0
0 1 2 3 4 5 6 7 8 9 10 11 12
W0 WIP

32 Worst Case
Little’s Law: 28
24
WIP WIP 20
CT   wT0
CT

16 Best Case
TH 1 / T0 12
T0 8
4
0
0 1 2 3 4 5 6 7 8 9 10 11 12
W0 WIP 21
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Worst Case Performance

Worst Case Law: The worst case cycle time for a given WIP
level, w, is given by,

CTworst(w) = w T0

The worst case throughput for a given WIP level, w, is given

by,

THworst(w) = 1 / T0

Randomness? None - perfectly predictable, but bad!

22
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Practical Worst Case

Observation: There is a BIG GAP between the Best Case and

Worst Case performance.

Question: Can we find an intermediate case that:

• divides “good” and “bad” lines, and
• is computable?

Experiment: consider a line with a given rb and T0 and:

1. single machine stations
2. balanced lines
3. variability such that all WIP configurations (states) are equally
likely

23
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 PWC Example – 3 jobs, 4 stations
clumped
up states
State Vector State Vector
1 (3,0,0,0) 11 (1,0,2,0)
2 (0,3,0,0) 12 (0,1,2,0)
3 (0,0,3,0) 13 (0,0,2,1)
4 (0,0,0,3) 14 (1,0,0,2)
5 (2,1,0,0) 15 (0,1,0,2)
6 (2,0,1,0) 16 (0,0,1,2)
7 (2,0,0,1) 17 (1,1,1,0)
8 (1,2,0,0) 18 (1,1,0,1)
9 (0,2,1,0) 19 (1,0,1,1)
10 (0,2,0,1) 20 (0,1,1,1) spread
out states
Note: average WIP at any station is 15/20 = 0.75,
so jobs are spread evenly between stations.
24
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Practical Worst Case

Let w = jobs in system, N = no. stations in line, and t =

process time at all stations:

CT(single) = (1 + (w-1)/N) t
CT(line) = N [1 + (w-1)/N] t
= Nt + (w-1)t
= T0 + (w-1)/rb

TH = WIP/CT From Little’s Law

= [w/(w+W0-1)]rb

25
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Practical Worst Case Performance

Practical Worst Case Definition: The practical worst case

(PWC) cycle time for a given WIP level, w, is given by,
w 1
CTPWC ( w) T0 
rb

The PWC throughput for a given WIP level, w, is given by,

w
TH PWC ( w)  rb ,
W0  w  1

where W0 is the critical WIP.

26
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 TH vs. WIP: Practical Worst Case

0.6
Best Case
rb 0.5
0.4 Good (lean)
PWC
TH

0.3
0.2 Bad (fat) Worst Case
1/T0 0.1
0
0 1 2 3 4 5 6 7 8 9 10 11 12
W0 WIP

27
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 CT vs. WIP: Practical Worst Case

32 Worst Case PWC

28
24
20 Bad (fat)
Best Case
CT

16
12 Good
(lean)
T0 8
4
0
0 1 2 3 4 5 6 7 8 9 10 11 12
W0 WIP

28
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Penny Fab Two Performance
0.5
Why is PF2
Best Case
rb 0.4 better than
PWC?
Fab 2
Penny
0.3

TH a l Wo r
st Case
Practic
0.2

0.1
1/T0
Worst Case
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26
W0 WIP
29
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Penny Fab Two Performance (cont.)
80

70
Worst Case
60
Case
rs t
o
50 al W 2
c ti c Fab
Pra ny
CT 40 Pen 1/rb

30

T0 20
Best Case
10

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26
W0
WIP
30
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Back to the HAL Case - Capacity Data

Process Rate (p/hr) Time (hr)

Lamination 191.5 4.7
Machining 186.2 0.5
Internal Circuitize 114.0 3.6
Optical Test/Repair - Int 150.5 1.0
Lamination – Composites 158.7 2.0
External Circuitize 159.9 4.3
Optical Test/Repair - Ext 150.5 1.0
Drilling 185.9 10.2
Copper Plate 136.4 1.0
Procoat 117.3 4.1
Sizing 126.5 1.1
EOL Test 169.5 0.5
rb, T0 114.0 33.9

31
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 HAL Case - Situation

Critical WIP: rbT0 = 114  33.9 = 3,869

Actual Values:
• CT = 34 days = 663 hours (at 19.5 hr/day)
• WIP = 47,600 panels
• TH = 71.8 panels/hour

Conclusions:
• Throughput is 63% of capacity
• WIP is 12.3 times critical WIP
• CT is 24.1 times raw process time

32
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 HAL Case - Analysis

TH Resulting from PWC with WIP = 47,600?

Conclusion?

33
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Labor Constrained Systems

Motivation: performance of some systems are limited by labor or

a combination of labor and equipment.

Full Flexibility with Workers Tied to Jobs:

• WIP limited by number of workers (n)
• capacity of line is n/T0
• Best case achieves capacity and has workers in “zones”
• ample capacity case also achieves full capacity with “pick and
run” policy

34
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Labor Constrained Systems (cont.)

Full Flexibility with Workers Not Tied to Jobs:

• TH depends on WIP levels
• THCW(n)  TH(w)  THCW(w)
• need policy to direct workers to jobs (focus on downstream is
effective)

Agile Workforce Systems

• bucket brigades
• kanban with shared tasks
• worksharing with overlapping zones
• many others

35
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]
 Factory Dynamics Takeaways

Performance Measures:
• throughput
• WIP
• cycle time
• service
• flexibility
Range of Cases:
• best case
• practical worst case
• worst case
Diagnostics:
• simple assessment based on rb, T0, actual WIP,actual TH
• evaluate relative to practical worst case

36
© Wallace J. Hopp, Mark L. Spearman, 1996, 2000 [Link]