Operations & Supply Chain Management Class Participation Guidelines

Let’s Get to Know Each Other

• Use Your Name Card
• Place your name card in front of you during each class
Class 3
Fundamentals of Process Analysis II Staying Focused Together
• No Cell phones (unless for QR codes)
• We’ll keep our focus on the class, so let's minimize
distractions

Laptops with Purpose

• Limited Laptop Use
Ali Shantia • Your laptops are awesome tools for learning—let’s use them
for taking notes and tackling problems

Punctuality matters
• Be on Time
• Arriving on time shows respect for everyone’s time and helps
us all get the most out of every session. Let’s make it a habit!
1 2

The story so far

• The way you Manage your Operations and Supply Chain can have a significant
impact on your firm’s performance.
• Four dimensions, Compromise is unavoidable
• The ones you should focus on — depends on your strategy

What have we seen so far in the course?

• Cycle Time: The average time between the completion of successive units
• Flow (or Throughput) Time: The length of time a unit spends at a given
stage/process
• Flow (or Throughput) Rate: The rate at which units flow through the
stage/process
• Capacity: The rate with which units flow through the stage/process
• Inverse of flow or cycle time
To-do
=
Can do
This Class Process Analysis Example: Capacity of bread-making on Two Lines
• More practice on analyzing processes
Production in batches of 100 loaves at a time
• Little’s Law
• Relationship between inventory, throughput rate, and flow time Mix Proof Bake

• Useful formula used in many different contexts

Raw 45 min / 45 min / 60 min /
batch batch batch WIP FGI
Mat. Pack
• Metalarm Case
40 min /
batch
Mix Proof Bake

30 min / 45 min / 60 min /

batch batch batch
FGI: Finished Good
Inventory
7

Process Analysis Example: Capacity of bread-making on Two Lines Bread Baking Example, cont.

• What is the minimum TT (for a rush order)?

Production in batches of 100 loaves at a time

• What is the throughput rate

Mix Proof Bake
– if the process runs at capacity?
Raw 45 min / 45 min / 60 min /
batch batch batch WIP FGI – if you launch one batch every 2 hours?
Mat. Pack

– if you launch three batches every hour

40 min /
batch – What happens if input > capacity?
Mix Proof Bake
• What is the process utilization
30 min / 45 min / 60 min /
batch batch batch – if the process runs at capacity?

Process capacity = Min (2, 60/40) = 1.5 batches/hour – if you launch one batch every 2 hours?
8 9
 Bread Baking Example, cont. Croissant Baking: An Assembly Operation!

• What is the minimum TT (for a rush order)?

TTMin = 30 min + 45 min + 60 min + 40 min = 175 min WIP
Raw Mat.
Dough Roll & Dough
Mix Proof Fill & 5 min / 50
• What is the throughput rate Cut
Fold croissants

– if the process runs at capacity? 1.5 batches / hour 5 min / 50 15 min / 50 5 min / 50
croissants croissants croissants
– if you launch one batch every 2 hours? ½ batch / hour Bake 20 min / 50
croissants
– if you launch three batches every hour? 1.5 batches/hour Raw Mat. WIP
Filling Mix Filling
10 min / 50
– What happens if input > capacity? Pack
croissants

• What is the process utilization 10 min / 50

croissants
– if the process runs at capacity? 100% FGI
Production in batches of 50 croissants at a time
– if you launch one batch every 2 hours? 33%
10 11

Croissant Baking cont. Croissant Baking cont.

• What is the capacity of the process? (batches per hour) • What is the capacity of the process? (batches per hour)
Bottleneck: Baking
Bottleneck rate = process capacity = 3 batches per hour

• What is the minimum process throughput time? • What is the minimum process throughput time?
Dough available after: 5 min + 15 min + 5 min = 25 min
Filling available after : 10 min
TTMin = 25 min + (5 min + 20 min + 10 min) = 60 min = 1 hr
• What is the cycle time of the process? • What is the cycle time of the process?
CT = 1 / 3 batches per hour * 60 min / hr = 20 min / batch
• If you start a batch of croissants every 40 minutes, what is the • If you start a batch of croissants every 40 minutes, what is the
process utilization? process utilization?
% of time busy: 20min/40min=50%
or TR/Capacity: 1 * (60/40) batches per hour / 3 batches per hour = 50%
12 13
What is cycle time of this process? Little’s law
Can we tell how long you need to
M1 wait before being served?
15 min/u
Can we obtain an estimate of the
M2 average number of customers
10 min/u who are kept waiting?

Capacity M1 : 60 / 15 = 4 unit / hr Do Absolutely!

Capacity M2 : 60 / 10 = 6 units / hr NOT
add or
Process capacity: 10 units / hr average CT: John Little
MIT Professor
In
Process CT: 60 / 10 = 6 min / unit CT = 1 / Capacity Operations
Research

14

Little’s Law Some Examples using Little’s Law

Inventory Throughput Rate WIP = Throughput rate * Throughput time

(Work in Process) [units/hr]
[units]
... ... ... ... ...
Ex: 1
ThroughputTime T [hrs] 6000 claims per year (50 Weeks). Processing time 2 weeks. # of
applications in the process?

TR = 6000 [claims / year] / 50 [weeks / year]

= 120 [claims/ week]
Inventory = throughput rate * throughput time TT = 2 [weeks]
WIP = 120 [claims/ week] * 2 [weeks]
Easier to remember: [units ] = [ units / time ] * [ time ] = 240 claims

If you know 2, you can determine 3rd

Valid for steady state only!!!

16 17
Some Examples using Little’s Law Some Examples using Little’s Law

WIP = Throughput rate * Throughput time WIP = Throughput rate * Throughput time

Ex: 2
A company has an average balance of
Ex: 3 (ATM)
Date Transaction Amount Balance
$30,000 in its Accounts Payable.
5 Deposit 1200.00 4533.37 The company pays off the balance six times Five people are in line ahead of you.
7 Check # 123 -77.37 4456.00 per year. The average transaction time is 2 minutes.
8 Check # 124 -100.00 4356.00 How much is the total purchase ($) per
year? How long does it take for you to get the money?

WIP = 30000 [$] TT = 1/6 [year]

TR = 30000 [$] / (1/6) [year] = 180,000 [$] / [year]

WIP = (5+1) customer

TR = 1/2 [customer/minutes]
TT = 6/(1/2) = 12 minutes

18 19

Some Examples using Little’s Law Little’s Law: practice problems

• Practice 1: Insurance company
WIP = Throughput rate * Throughput time
• An insurance company processes 10,000 claims per year. The average
processing time is 3 weeks. How many claims are in the system on
average? (Assume 50 weeks in a year)
Ex: 4 (Restroom at Rock Concert)
• If the company reduces its processing time to 80%, how many claims
15 people ahead of you waiting to be served.
It takes on average 2 minutes to serve. will be in the system on average?
How long does it take for you before you go back to the concert?
TR = 10,000 claims / year = 200 claims / week
TT = 3 weeks
WIP = 3 * 200 = 600 claims are on the system

New TT= 0,8*3=2,4 weeks

New WIP=2,4*200=480 claims are on the system after reduction in TT

WIP = (15+1) customer

TR = 1/2 [customer/minutes]
TT = 16/(1/2) = 32 minutes
21
20
Little’s Law: practice problems Little’s Law: practice problems
• Practice 2: Supply Chain • Practice 3: Cash Flows
• Wal-Mart imports 3000 pairs of shoes from an overseas factory every • Motorola sells $300 million worth of cellular equipment per year. The
month. Each order from Wal-Mart goes through several stages before it average accounts receivable in the cellular group is $45 million. What is
gets to the stores (as shown below). Assuming 30 days in a month: the average time from billing to collection?
• How much inventory of shoes is tied up at the warehouse?
• How much inventory is tied up in the entire supply chain?

TR=$300 Million / year

WIP = $45 Million
Factory Port Ship Warehouse TT = 45 / 300 = 0,15 year = 1,8 Month (1 month and 24 days)
2 days 1 day 5 days 3 days
TR=3000 pairs per month = 100 pairs per day
TT (warehouse) = 3 days
WIP = 3 * 100 = 300 pairs of shoes tied up in the Warehouse

TT (supply chain) = 11 days

WIP = 11*100=1100 pairs of shoes tied up in the supply chain

22 23

Little’s Law: practice problems Little’s Law: practice problems

• Practice 4: Capacity Calculation • Practice 5: Starbucks
• A fast-food restaurant has 20 seats. An average customer spends half • David goes to the same Starbucks every morning on the way to work.
an hour in the restaurant. What is the capacity rate of the restaurant? On an average day, there are 4 people waiting to order a coffee (except
the person who’s already ordering) and David waits for 10 minutes
before ordering a coffee. As soon as David arrives at Starbucks, how
long does it take for the next customer to queue behind David?
Queueing Process:
TT = 0,5 hour WIP = 4 + 1 = 5 people in the queue
WIP = 20 TT = 10 minutes waiting in the queue
TR = 20 / 0.5 = 40 customer per hour TR = 5 / 10 = 0,5 people per minute
Average Cycle time = 1/0,5 = 2 minutes the average time
• Half the customers stay inside Starbucks after getting a coffee, sitting
on a chair for one hour on average (the other half leave immediately).
How many chairs are occupied on average?
Sitting Process: (half of the TR of the waiting process)
TR = 0,5/2 customer per minutes = 0.25 customer/minute
TT = 60 minutes
WIP = 60* 0,25 = 15 Customers are sitting on average
24
Little’s Law: Inventory Turns Little’s Law: Inventory Turns
• Inventory Turns (turnover) is a financial ratio showing how many • Inventory turns in practice:
times a company turned over its inventory. • Dell: 5-8 times more than competitors
• Toyota: Spare parts inventory 10 times faster than competitors
• Wal-Mart: 7.5 turns in 2002 (K-Mart was less than 5)
• Example: Material Flow at Wendy’s
• A Five Guys branch processes an average of 5,000 lbs of hamburgers • Why are higher inventory turns good?
weekly. The typical inventory of raw meat is 2,500 lbs. • Shorter flow time or a more responsive system
• What is the average hamburger’s flow time? • Revenue is generated faster
• How many is Wendy’s inventory turn?

TR = 5000 lbs/week
WIP = 2500 lbs
TT = =2500/5000 = 0.5 week
Inventory Turn = 1/TT = 1/0.5 = 2 turns/ week
26 27

Little’s Law: Summary

• Provides a relationship between the inventory, throughput rate
and flow time of a process

• Always holds (on average)

• Can be applied to any part of the process

• Provides insight into how to improve system responsiveness:

Process Analysis in Action
Team up to work on the case, the Metalarm.
• Allows one to identify what causes inventory / flow time to grow
• Allows one to identify the levers on which improvement efforts should
be focused

28
Metalarm case: Metalarm case:
Solution Solution

Cutting Stamping Soldering Painting Assembly Packaging Cutting Stamping Soldering Painting Assembly Packaging

Cycle time in minutes per unit Cycle time in minutes per unit
0.6 3.53 10 1 10 5
Hourly max. production rate (capacity) per workstation in units Hourly max. production rate (capacity) per workstation in units
100 17 100 17 6 60 6 12
Maximum number of machines/workstations Maximum number of machines/workstations

Hourly max production rate per processing step Hourly max production rate per processing step

Metalarm case:
Solution
Capacity, a different view!
Cutting Stamping Soldering Painting Assembly Packaging
Consider
• Different number of shifts per day
Cycle time in minutes per unit
• Different number of days worked per week
0.6 3.53 10 1 10 5
• Different machine maintenance requirements
Hourly max. production rate (capacity) per workstation in units
• Different setup time requirements
100 17 6 60 6 12
• Yield Rates
Maximum number of machines/workstations
• ...
1 4 10 1 infinite 6
Hourly max production rate per processing step
Sometimes it is easier to calculate the capacity in time available for
100 68 60 60 infinite 72 production and compare to need capacity in time
Where is the bottleneck? • particularly useful, if a set of products are being produced on
the same machines
Example: The actual capacity of Painting Metalarm case:
Station Solution
• Monthly theoretical availability:
• 8 h * 20 d = 160 h
Cutting Stamping Soldering Painting Assembly Packaging
• Monthly setup workload: (effective capacity)
• 20h = 5 color setups of 1 hour per week each Available hours (per month and machine)
• Monthly capacity loss due to breakdowns (actual capacity) 160 160 160 120 160 160

• 20 h = 1 hour of breakdown per day Maximum number of machines/workstations

1 4 10 1 infinite 6
Monthly capacity in hours (per processing step)
• Available monthly time of production: 120 h 160 640 1600 120 infinite 960

Anything Else to Consider? Alternative: Capacity and workload in hours

Particularly useful, if you produce several DIFFERENT
Rework of 5% of the produced cabinets in paint shop:
products on the same line!
105 = 95 ok + 5 not ok + 5 rework
100 95 Cutting Stamping Soldering Painting Assembly Packaging
100
Painting
Monthly capacity in hours (per processing step)
5
160 640 1600 120 infinite 960
Planned Workload (in units) - January
7,600 7,600 7,600 7,980 7,600 3,800
Cutting Stamping Soldering Painting Assembly Packaging
Hourly max. production rate per workstation in units
100 17 6 60 6 12

Planned Workload (in units) - January 7,600*1.05 Planned Workload (in hours) – January: planned workload in units / production rate
76 447 1267 133 1267 317
7,600 7,600 7,600 7,980 7,600 3,800
Planned Capacity Utilization – January (workload/capacity)
 Alternative: Capacity and workload in hours
Can Metalarm meet the
Particularly useful, if you produce several DIFFERENT
products on the same line! target in January?
Cutting Stamping Soldering Painting Assembly Packaging
• Capacity Utilization: 110.8 % (hypothetical)
Monthly capacity in hours (per processing step) • Only 6 857 cabinets can be delivered
160 640 1600 120 infinite 960
instead of 7600
Planned Workload (in units) - January
7,600 7,600 7,600 7,980 7,600 3,800 • What solutions do you suggest?
Hourly max. production rate per workstation in units
100 17 6 60 6 12 Two possible actions sets:
Planned Workload (in hours) – January: planned workload in units / production rate Reduce the workload (in hours required –
76 447 1267 133 1267 317
not units!)
Planned Capacity Utilization – January (workload/capacity)
47.5% 69.85 % 79.19% 110.83% --- 33%
Increase the capacity (in hours available)

Possible actions to reduce the planned Possible actions to increase capacity (120
workload (133 hours): hours:

1 2 3 4 5
1 2 Invest in a second Do overtime / work Increase reliability Sub-contract the Reduce the lost time in
tunnel Saturdays? • Preventive maintenance painting color setups
• Utilization rate would be • Consider cost (labor, • Consider transporting and • . Reduce the number of
Increase the tunnel speed Reduce the defect ratio low (Consider the cost) inventory) handling problems, cost,
etc.
colors ( marketing)
• But could increase setup • Reduce the setup
• Upgrade the rate from 60 to 65 (-> frequency (consider frequency (planning)
planning) • Reduce the setup time
quality?) (technical problems).
 Approaches to increase bottleneck capacity
Manage Capacity through Bottleneck

Resources
Labor Machines New bottleneck(s) may be created after we increase the capacity
of old bottleneck(s)
Cross-train workers Invest in new equipment
Long Term Adjust number of Outsourcing
workers Improve reliability Bottleneck Bottleneck

Reduce setup time Input Input

Time Subcontracting
Horizon Medium Term Flexible time Preventive maintenance Capacity Capacity
Subcontracting
Hire part-time
workers
Short Term Priority management
Use overtime
Add shifts

42

Let’s consider the preventive maintenance Throughput Time Analysis

plan:
• Why potentially useful?
• Capacity utilization > 100%
30 € 90 € 150 €
• Cost?
0.5M € 1M € 0.5M €
• 5000 €/ month for 5 hours or 1000 €/ hour
• Compared to? 16,666 11,111 3,333
• Production cost of 250 €/ hour 6,666 6,666 6,666
• Overhead cost certainly > 250 €/ hour – but 4x more?
• Probably cheaper to use overhead, if a possibility 2.5 1.67 0.5
• If overtime not possible?
• Lost margin (225-150)=75 €/ for 60*5/1.05 = 285 products
Avg. Throughput rate: 18M € / 225 € = 80,000 units /y à 6,666 u / m
• Additional Revenues 21,375 € - so worth it!!
TT=WIP/TR => Total: 4.67 months
45
 What happens to capacity, when the batch size increases?

• Capacity increases with batch size …

0.5
Capacity
0.45

The paintshop has setup times each time you switch the colour. But you can 0.4

CHOOSE how often to switch, that is you can choose the batch size!
0.35

0.3

0.25
What is the effect of setups on capacity, inventory, and throughput time? 0.2

0.15

0.1

0.05

0 Batch Size
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 210 220 230

46 47

What happens to WIP inventory, when the batch size increases? Optimal Batch Size: Balance Capacity and Inventory

• … but so does the inventory (and thus avg. TT)

• At a bottleneck resource capacity is extremely valuable:
Inventory Profile
– Increase batch size to increase capacity up to desired TR.

• At a non-bottleneck resource setups are free!

ST ST ST ST ST ST ST ST time
– Decrease batch size to lower inventory and avg. TT – without creating a
new bottleneck (possible only if buffer space exists!)
Inventory Profile

B
A

ST ST ST ST ST ST time

48 49
Choose batch size that provides desired TR at bottleneck! Manage Capacity through Bottleneck

• Bottleneck determines release of material to shop floor, otherwise WIP increases

• Ensure bottleneck time is not wasted
– Eliminate non-value added work from bottlenecks
0.5
Capacity 1/p – Increase batch size at bottleneck to exploit economies of scale or reduce /
0.45

0.4 eliminate setups and changeovers

0.35 – Add inventory buffer before bottleneck (bottleneck running despite breakdown,
0.3
worker lunch, setups, etc. in non-bottleneck resources)
0.25
• ⇓ load of bottleneck
0.2

0.15 – Avoid rework (do it right first time)

0.1 – Move work to non-bottleneck resources or to 3rd party
0.05
• ⇑ availability of bottleneck
0 Batch Size
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 210 220 230 – Work longer: increase scheduled availability
– Increase number of resources at the bottleneck: invest

50 51

Next Class

• More on processes:
– the impact of variability
– managing the service-cost trade-off
• Read Benihana