Variability Basics

Dr. Jeff Hong

IELM Dept, HKUST
Fall 2010

Note02: Variability Basics 1

Motorola Case Study

Motorola Semiconductor Products Sector produces integrated circuits both

for use in Motorola products (e.g., cell phones) and for other OEMs
(original equipment manufacturers). They do this in vastly complex wafer
fabs that can cost US$2 billion or more to construct. Not surprisingly,
efficient utilization of these enormously expensive resources is a key
concern in the semiconductor industry. Despite this, Motorola deliberately
sizes capacity of each process in a wafer fab so that utilization will be no
higher than a specified limit, typically in the range of 75-85%.

Discussion: Why Motorola keeps a 15-25% excessive

capacity? Note that the excessive capacity is very costly.

Note02: Variability Basics 2

VIA Example

„ Current situation
‰ Average throughput: 10 passengers/min
‰ Average queue length: 16.5 passengers
‰ Average waiting time: 1.65 min

„ Proposed system
‰ Average throughput: 10 passengers/min
‰ Average queue length: 0 passengers
‰ Average waiting time: 0 min

„ What makes the difference?

Note02: Variability Basics 3

Tortoise and Hare Example

„ Two machines:
‰ subject to same workload: 69 jobs/day (2.875 jobs/hr)
‰ subject to unpredictable breakdowns (availability = 75%)

„ Machine 1:
‰ long, but infrequent breakdowns

„ Machine 2:
‰ short, but more frequent breakdowns

„ Which one has better performance?

Note02: Variability Basics 4

Variability Views

„ Variability:
‰ Any departure from uniformity
‰ Random versus controllable variation

„ Mean Effects:
‰ Throughput increases with machine speed
‰ Throughput increases with availability
‰ Our intuition is good for mean effects

„ Variability Effects:
‰ Which are more disruptive – long, infrequent failures or short
frequent ones?
‰ Our intuition is less secure for variability effects

Note02: Variability Basics 5

Variability

„ Definition: Variability is anything that causes the system

to depart from regular, predictable behavior.

„ Sources of Variability:
□ setups □ work pace variation
□ equipment failures □ differential skill levels

□ materials shortages □ yield loss

□ rework □ product differentiation

□ operator unavailability

Note02: Variability Basics 6

Measuring Process Variability

te = mean process time of a job

σ e = standard deviation of process time

σe
ce = = coefficient of variation, CV
te

Note: we often use the “squared

coefficient of variation” (SCV), ce2

Note02: Variability Basics 7

Variability Classes
Low variability Moderate variability High variability
(LV) (MV) (HV)
ce
0 0.75 1.33

„ Effective Process Times:

‰ actual process times are generally LV
‰ effective process times include setups, equipment
breakdowns, etc.
‰ HV, LV, and MV are all possible in effective process times

Note02: Variability Basics 8

Natural Variability

„ Definition: variability without explicitly analyzed cause

„ Sources:
‰ operator pace
‰ material fluctuations
‰ product type (if not explicitly considered)
‰ product quality

„ Observation: natural process variability is usually in the

LV category.

Note02: Variability Basics 9

Variability Inflators
„ Sources:
‰ equipment failure
‰ machine setup
‰ operator unavailability
‰ rework
‰ material unavailability

„ Effects:
‰ inflate effective processing time
‰ inflate process variability

„ Consequences: Effective process variability can be LV,

MV, or HV.

Note02: Variability Basics 10

Variability Analysis - Failure

„ Definitions:

t0 = base process time

c0 = base process time coefficient of variation
1
r0 = = base capacity (rate, e.g., parts/hr)
t0
m f = mean time to failure
mr = mean time to repair
cr = coefficent of variation of repair times (σ r / mr )

Note02: Variability Basics 11

Variability Analysis - Failure

„ Availability: Fraction of time machine is up

mf
A=
m f + mr

„ Effective Processing Time and Rate:

re = Ar0

te = t0 / A

Note02: Variability Basics 12

Variability Analysis - Failure

„ Machine 1: „ Machine 2:
t0 = 15 min t0 = 15 min
σ0 = 3.35 min σ0 = 3.35 min
c02 = σ02 /t02 = 3.352/152 = 0.05 c02 = σ02 /t02 = 3.352/152 = 0.05
mf = 12.4 hrs (744 min) mf = 1.9 hrs (114 min)
mr = 4.133 hrs (248 min) mr = 0.633 hrs (38 min)
cr = 1.0 cr = 1.0

„ Availability:

mf 744 mf 114
= = 0.75 = = 0.75
m f + mr 744 + 248 m f + mr 114 + 38
A= A=

No difference between machines in terms of availability.

Note02: Variability Basics 13

Variability Analysis - Failure
„ Effective Variability:
te = t0 / A
⎛ σ 0 ⎞ (mr + σ r )(1 − A)t0
22 2
σe = ⎜ ⎟ +
2

⎝ A⎠ Amr Variability
depends on
σ e2 mr
c =
2
e = c + (1 + c ) A(1 − A)
2
0
2
r
repair times
te2 t0 in addition to
availability
„ Conclusions:
‰ Failures inflate mean, variance, and CV of effective process time
‰ Mean (te) increases proportionally with 1/A
‰ SCV (ce2) increases with mr
‰ SCV (ce2) increases in cr2
‰ For constant availability (A), long infrequent failures increase SCV
more than short frequent ones

Note02: Variability Basics 14

Variability Analysis - Failure

„ Machine 1: „ Machine 2

t0 15 t0 15
te = = = 20 min te = A 0.75 = 20 min
=
A 0.75
mr mr
= c 0 + (1 + c r ) A(1 − A) =
2 2
= c + (1 + c ) A(1 − A) =
2 2
ce2 0 r ce2
t0 t0
248 38
0.05 + (1 + 1)0.75(1 − 0.75) = 0.05 + (1 + 1)0.75(1 − 0.75) =
15 15
6.25 high variability 1.0 moderate variability

Machine 1 is much more variable than Machine 2!

Note02: Variability Basics 15

Variability Analysis - Failure

„ Managerial Insights

‰ Shorter, more frequent breakdowns induce less variability.

‰ Shorter, more frequent setups induce less variability.

Note02: Variability Basics 16

 Illustrating Flow Variability

Low variability arrivals

t

smooth!

High variability arrivals

t

nonsmooth!

Note02: Variability Basics 17

Measuring Flow Variability

t a = mean time between arrivals

1
ra = = arrival rate
ta

σ a = standard deviation of time between arrivals

σa
ca = = coefficient of variation of interarrival times
ta

Note02: Variability Basics 18

Propagation of Variability
ce2(i)
ca2(i) cd2(i) = ca2(i+1)
i i+1

departure variance
depends on
„ Single Machine Station: arrival variance
and process
cd2 = u 2 ce2 + (1 − u 2 )ca2 variance

where u is the station utilization given by u = rate

Note02: Variability Basics 19

Propagation of Variability –
High Utilization Station
LV HV HV

HV HV HV

LV LV LV

HV LV LV

Conclusion: flow variability out of a high utilization station is

determined primarily by process variability at that station.
Note02: Variability Basics 20
Propagation of Variability –
Low Utilization Station
LV HV LV

HV HV HV

LV LV LV

HV LV HV

Conclusion: flow variability out of a low utilization station is

determined primarily by flow variability into that station.
Note02: Variability Basics 21
Variability Interactions - Queueing

„ Importance of Queueing:
‰ Queueing is a common behavior in all operation systems
‰ Queueing and waiting time comprise majority of cycle time

„ System Characteristics to analyze queues:

‰ Arrival process
‰ Service process
‰ Number of servers
‰ Maximum queue size (blocking)
‰ Service discipline (FCFS, LCFS, EDD, SPT, etc.)

Note02: Variability Basics 22

Kendall's Classification

A/B/C
B
A: arrival process A
B: service process
C: number of machines

M: exponential (Markovian) distribution Queue C

Server
G: completely general distribution
D: constant (deterministic) distribution.

Note02: Variability Basics 23

Queueing Parameters

ra = the rate of arrivals in customers (jobs) per unit

time (ta = 1/ra = the average time between
arrivals).
Note: a station
ca = the CV of inter-arrival times. can be
m = the number of machines. described
with 5
re = the rate of the station in jobs per unit time = parameters.
m/te.
ce = the CV of effective process times.

u = utilization of station = ra/re.

Note02: Variability Basics 24

Queueing Measures
„ Measures:
CTq = the expected waiting time spent in queue.
CT = the expected time spent at the process center, i.e., queue
time plus process time.
WIP = the average WIP level (in jobs) at the station.
WIPq = the expected WIP (in jobs) in queue.

„ Relationships:
CT = CTq + te
WIP = ra × CT
WIPq = ra × CTq

„ Result: If we know CTq, we can compute WIP, WIPq, CT.

Note02: Variability Basics 25

The G/G/1 Queue

„ Formula:
CTq ≈ V × U × t
⎛ ca2 + ce2 ⎞⎛ u ⎞
≈ ⎜⎜ ⎟⎜
⎟⎝ 1 − u ⎟⎠t e
⎝ 2 ⎠

„ Observations:
‰ Useful model of single machine workstations
‰ Separate terms for variability, utilization, process time.
‰ CTq (and other measures) increase with ca2 and ce2
‰ Flow variability, process variability, or both can combine to
inflate queue time.
‰ Variability causes congestion!

Note02: Variability Basics 26

The G/G/m Queue

„ Formula:
CTq ≈ V × U × t
⎛ ca2 + ce2 ⎞⎛⎜ u 2 ( m +1) −1 ⎞⎟
≈ ⎜⎜ ⎟⎟ te
⎜
⎝ 2 ⎠⎝ m(1 − u ) ⎠ ⎟

„ Observations:
‰ Useful model of multi-machine workstations
‰ Extremely general.
‰ Fast and accurate.
‰ Easily implemented in a spreadsheet

Note02: Variability Basics 27

Motorola Case Study
24

22

20

18

16

14
Cycle Time

12

10

4
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
Utilization

Note02: Variability Basics 28