Lectures Op

Today’s class
• Introduction
Gemma Berenguer
Operations Management
Introduction
• Syllabus
• What is OM?
• Form groups
Fall 2021
Department of Business Administration
UC3M 2
Professor Gemma Berenguer
About me About you?

• Professor Gemma Berenguer • Country/Provincia
• Office 6.0.11
• Main interest
• Office phone 91 624 8679
Gemma Berenguer
Gemma Berenguer
• Background
• I am from Spain
• BS in Mathematics
• Ms in Economics
• MEng in Logistics and SCM
• PhD in Operations Research (U.C. Berkeley)
• Professor at Purdue University
• Experience/Interests
• Optimization and Supply chain design
• Operations management for public and nonprofit institutions
• Sustainable and socially responsible operations 3 4
Syllabus Materials
• Case packet
• Case questions on Aula Global
• Textbooks:
• David Bamford and Paul Forrester. “Essential Guide to Operations
Gemma Berenguer
Gemma Berenguer
Management”. Wiley. 2012
• Jay Heizer, Barry Render and Chuck Munson "Operations Management:

sustainability and supply chain management". Pearson, 12th Edition.
• Slides
• Posted on Aula Global after each session
• Quiz solutions
• Sample exams
5 • Supplemental files 6
Submissions
Course Grading
Continuous Evaluation (70%)
• Class attendance and participation (5%)

Gemma Berenguer
Gemma Berenguer
• Related to practical sessions (30 %)
o 3 in-class quizzes (10% each)
• Related to theoretical sessions (magistrales) (35%)
o 3 Case deliverables (10% each, but only counting best 2, so 20%)
o Final project (15%)
Final Examination (30%): December 20th! 7 8

Goals of the course How will we do it?
• An overview of Operations Management and Operations • Lectures
Strategy • Examples
• A process view of the firm and tools for analyzing processes • Case discussions
• Tools for managing manufacturing and service processes
Gemma Berenguer
Gemma Berenguer
• Quizzes
• Tools for understanding and managing the impact of variability • Final project
• Examples on how companies can use (and have used) the OM
tools from this course to significantly enhance their
competitiveness
9 10
What is Operations
If excessive supply…
Management?
• It is an area of business concerned with the production
of goods and services, and involves the responsibility of
Gemma Berenguer
Gemma Berenguer
ensuring that business operations are efficient in terms
of using as little resource as needed, and effective in
terms of meeting customer requirements.
• In short: Matching supply with demand
11 12
Goods-service continuum If excessive demand…
Goods Service
Surgery, teaching
Gemma Berenguer
Song writing, software development
Computer repair, restaurant meal
Automobile Repair, fast food
Home remodeling, retail sales

13 14
Automobile assembly, steel making
Operations Management Supply Chain Management

The management of systems or processes that create goods
and/or provide services.
Sequence of activities and organizations involved in
people, equipment, money, raw Goods/ producing and delivering a good or service.
material, components services
Gemma Berenguer
Gemma Berenguer
Manufacturer Distributor Wholesaler Retailer
Output
Input
Operations
Control
15 16
Supply Chain Management What defines a ‘good process’?
Ultimately, all organizations compete on delivered value.
Delivered value of process =

benefit to customers – total process cost
Gemma Berenguer
Benefit driven
by customer value
Price p Variety V
Quality Q: Time T (Response time):

▪ of product ▪ rapid, reliable delivery
17 ▪ outcome of service ▪ new product 18
development
How do we measure OM
ROA
success?
• Absolute measures:
Operating Profit
• Revenues, costs, operating income, net income ROA = =
Total Assets
Gemma Berenguer
Gemma Berenguer
• Relative measures: Sales - Costs
= =
• ROA Total Assets
• ROI, ROE Sales - Costs Sales
= * =
Total Assets Sales
• Survival measure: Sales Sales - Costs
• Cash flow = *
Total Assets Sales
19 20
Is it enough? Competitive measures
• However, it is very difficult to understand the exact impact of • Cost vs. quality
operations on assets and sales.
• Why? • Cost vs. variety
•
Gemma Berenguer
Gemma Berenguer
Aggregate measure – hard to separate the impact;
• No details; • Cost vs. time
• Direct connection is not transparent;
• Many dependencies; • Time vs. quality
• External factors;
21 22
Cost Cost
• Fixed (overhead) costs • Fixed (overhead) costs
• DOES NOT CHANGE WITH QUANTITY

• Variable costs
Gemma Berenguer
Gemma Berenguer
Cost
Cost
• Variable costs
• CHANGES WITH QUANTITY
Q*
Quantity Quantity
Need to produce more than Q*
to justify new technology
• Total costs = fixed costs + quantity x variable costs 23 24
• Average costs = Total costs/quantity • Total costs = fixed costs + quantity x variable costs
• Average costs = Total costs/quantity
Variety Is variety good or bad?
• E.g. HP Workstations -- 500,000 options • Manufacturing Costs vs. Customer Choice
• RAM cards, video cards, graphic cards, monitors, disk drives, CD-
ROM drives
Higher Complexity Cost:
Gemma Berenguer
Gemma Berenguer
• Why?
more parts, changeovers, etc.
• Market Segmentation • High variety
• Competition Happier Customers:
more choices
• Technology
• International Differences
• Government Requirements
25 26
Understanding tradeoffs Caterpillar Inc.

• Engine options: 200 hp, 250 hp, 300 hp;
Cost
• Stick options: Integrated or Standard;
High Inefficient • Cabin Options: Canopy or Deluxe with A/C…
Gemma Berenguer
Gemma Berenguer
• 2,275,000 Configurations
Not attainable
Low
Before After
Low High Product Variety Configurations sold 569 76
27 Sales Volume 3825 3812 28
Profit $63M $67M

Shifts Shifts
Cost Cost
High Caterpillar move: High Caterpillar move:
Gemma Berenguer
Gemma Berenguer
Inefficient Inefficient
Shifting the frontier

requires innovation
Low
Not attainable Low
Not attainable
Low High Product Variety Low High Product Variety
29 30
Type of decisions in OM Outline of the course

• Part 1 (Quiz 1)
• Operations Strategy
• Project Management
• Forecasting
Gemma Berenguer
Gemma Berenguer
• Part 2 (Quiz 2)
• Process Analysis
• Part 3 (Quiz 3)
• Statistical Process Control
• Lean Operations
• SCM
• Humanitarian Logistics
• Part 4 (Final project)

31 • Circular Economy/Sustainability 32
Next day (Wednesday!)
2 08/09 Topic: Operations and Strategy
• Form study teams of 4 or 5 students!
Case: American Connector Company
Gemma Berenguer
Gemma Berenguer
1. What are the key differences in the operations at American Connector’s
Sunnyvale plant and DJC's Kawasaki plant?
2. How does ACC compete? Do the operations choices made by ACC
support ACC’s competitive strategy?
3. How does DJC compete? Do the operations choices made by DJC
support DJC’s competitive strategy?
4. What should ACC do to respond to the threat of DJC starting operations 33 34
in the US?
Today’s Outline
Gemma Berenguer
• American Connector
Session 2 • Product/Process Matrix
American Connector Company • Start with Project Management…
Fall 2021
UC3M 2
American Connector Company Industry environment
• Main discussion points:
• Electrical connectors found in virtually every electrical
1. What are the key differences in the operations at American Connector’s and electronic device. Market for both standard and
Gemma Berenguer
MGMT660
Introduction to Operations Management
Sunnyvale plant and DJC's Kawasaki plant? custom designs. Trend towards standardization.
2. How does ACC compete? Do the operations choices made by ACC support • Worldwide sale of $16 billion in 1991. There were
ACC’s competitive strategy? 1200 suppliers, but top 10 accounted for 6.67 billion of
3. How does DJC compete? Do the operations choices made by DJC support sale.
DJC’s competitive strategy? • Market intensely competitive in the U.S. Sales
4. What should ACC do to respond to the threat of DJC starting operations in declining in the recent years. OEM’s were demanding
the US? lower prices, improved quality, and delivery.
3 • Conditions in the U.S. were expected to lead to 4
consolidation of suppliers from 900 to 400.
Competitive position Compare Manufacturing Costs

DJC 1986 ACC 1986 DJC 1991 ACC 1991
ACC
DJC Raw Material, 14.32 10.40 12.13 9.39
Product
• Among top 10 in the • Among top 10 in Raw Material,
Packaging
3.27 2.25 2.76 2.10
world, $800M in sale world, sale between Labor, Direct 7.63 ... 3.02
Gemma Berenguer
MGMT660

in 1991 $500 to $800M Labor, Indirect 2.30 ... 0.75
Total Labor ... 8.53 10.30
• Gross margin 52% in • Gross margin 50% Electricity 2.47 1.80 1.40 0.80
1984, 43% in 1991 • Standardization Depreciation 7.63 5.52 1.80 5.10
• Customization and • Efficiency in

Other 4.12 4.41 4.24 6.10
TOTAL 41.74 32.91 26.10 33.79
quality production Exhibit 8 Cost Indices, United
States/Japan (1991)
• High WIP/ Low FG • Low WIP/ High FG Expense Item Index
Raw Material, Product 0.60
Raw Material, Packaging 0.60
5 Labor, Direct 1.10
Labor, Indirect 1.10
Electricity 0.80
Depreciation 1.00
Other 1.00
Static Costs Comparison Static Cost comparison
DJC/Kawa ACC DJC
Index DJC/US ACC
saki • 33.79 • 12.13*0.6 +
RM, $ 12.13 0.6 $ 7.28 $ 9.39 2.76*0.6+3.77*1.1+1.4*
Product
0.8+1.8+4.24=20.241
RM, $ 2.76 0.6 $ 1.66 $ 2.10
Packaging
X = VS
Labor $ 3.77 1.1 $ 4.15 $ 10.30
(total)
Electricity $ 1.40 0.8 $ 1.12 $ 0.80
Depreciati $ 1.80 1 $ 1.80 $ 5.10
on
Other $4.24 1 $ 4.24 $ 6.10
7 8
Total $26.10 $ 20.25 $ 33.79
Dynamic Costs Comparison Operations Strategies

3% more ACC DJC
DJC/Kawasaki ACC/Sunnyval DJC/Kawasa ACC/Sunnyv
70% utilization 100% utilization
1986 e 1986 ki 1991 ale 1991
RM, Product $ 14.32 $ 10.40 $ 12.13 $ 9.39 Work with customers to Standardized products
customize products
RM, Packaging $ 3.27 $ 2.25 $ 2.76 $ 2.10
5 days a week, 50 weeks a year 24/7 operations
Labor (total) $ 9.93 $ 8.53 $ 3.77 $ 10.30
Facilities are organized by Facilities are organized by
Electricity $ 2.47 $ 1.80 $ 1.40 $ 0.80
process (cellular structure) product (assembly line)
Depreciation $ 7.63 $ 5.5.2 $ 1.80 $ 5.10
Make to order Make to stock (24/7)
Other $ 4.12 $ 4.41 $4.24 $ 6.10
9 10
Total $ 41.74 $ 32.91 $26.10 $ 33.79 Quality by inspection Quality by design
37% less
Competitive strategy IS ACC actually on the frontier?
• E.g. compare labor productivity
Cost • Kawasaki: 700 million units/94 workers = 7.4M per worker
• Sunnyvale: 420 million units/396 workers = 1.1M per worker
ACC ?
Gemma Berenguer
Gemma Berenguer
• Working hours
• Sunnyvale: 6000 (=5x24x50) hours : 2400 short of DJC
(7x24x500=8400)
• Adding 2400 hours at ACC productivity rate of .07 (=420/6000)
DJC per hour, they can add 168 million (=2400x0.07) to 420 million
• Total of 588 million units vs 700 at DJC
• Hence, even of ACC runs for as many hours as DJC, they cannot
catch up with DJC’s volumes
11 12
Variety
What should ACC do? Production strategies

• Keep competitive strategy
• Change competitive strategy towards standardization Make To Assemble Make To
Order To Order Stock
• Improve operational efficiency
Gemma Berenguer
Gemma Berenguer
• Hybrid plant (to appeal to both markets) MTO: Begin production ATO: Build modules MTS: Manufacture entire
• Customized line only after receiving a prior to receiving order; product prior to receiving
customer order. assemble product after customer order.
• Standardized line receiving customer order.
Slower Response Time Faster

Raw Inventory Finished
Materials Goods
High Labor Skill Level Low
13 General 14
Type of Equipment Dedicated
purpose
Types of Processes Work Center / Job shop
PRODUCT STRUCTURE
• A collection of general
Few High volume,
Low volume,
Multiple major high standardi- purpose processing resources,
low standardi-
PROCESS
zation, one of a
products, products, zation, together with a highly
STRUCTURE: low volume higher commodity
kind structured information
Gemma Berenguer
Gemma Berenguer
volume products
processing system which
Job Shop
allows those resources to be
Disconnected Line continuously recombined in
Flow (Batch) the production of different
outputs.
Connected Line
Flow (Assembly
Line)
Continuous Flow 15 16
Source: Link Manufacturing Process and Product Life Cycles by Robert H. Hayes and Steven C.
Wheelwright, HBR, January 1979
Manufacturing Cell / Batch

Assembly Line (Line Flow)
Flow
• A somewhat standardized work center-- something • Fixed resources, typically specialized or narrow in function,
intermediate between a work center and an assembly line. through which work flows in serial fashion. The system is
Similar equipment or functions are grouped together. Work balanced for a particular overall flow rate and a particular
frequently is processed in batches. product mix, and only small variations around those nominal
Gemma Berenguer
values are tolerable.
17 18
Impact of process choice on facility
Continuous Flow layout
• An even more standardized product (vis a vis assembly line)
Product A 1
where the product is not a number of discrete parts but rather 2
Milling 5
a continuous flow of undifferentiated product. Product B 1
Equipment 4 3
Drills
Gemma Berenguer
Gemma Berenguer
Batch Shop
2 4 3
6
Finishing
Lathes 7
Equipment
Forester Power Body Engine Paint

Line Train Assembly Mount Shop
Assembly Line
19 Camry 20
Power Body Engine Paint
Line Train Assembly Mount Shop
A Continuum of Process Types Types of Processes

PRODUCT STRUCTURE
Few High volume,
Low volume,
Multiple major high standardi-
low standardi-
PROCESS products, products, zation,
PROCESS FLEXIBILITY STRUCTURE:
zation, one of a
low volume higher commodity
kind
Gemma Berenguer
Gemma Berenguer
volume products
Job Shop
Batch Shop Continuous Process
Disconnected Line
Job Shop Assembly Line Flow (Batch)
Connected Line
Flow (Assembly
PRODUCT STANDARDIZATION Line)
21 Continuous Flow 22
Source: Link Manufacturing Process and Product Life Cycles by Robert H. Hayes and Steven C.
Wheelwright, HBR, January 1979
Project characteristics
► The job is unique or somewhat unfamiliar
► Many related activities
► Temporary but critical to the organization
Session 2
► Cuts across organizational lines
► The job requires high labor skills
Project Management 1
Fall 2021
UC3M 2
Professor Gemma Berenguer Source: Jay Heizer, Barry Render and Chuck Munson "Operations Management: sustainability and
supply chain management". Pearson, 12th Edition.(chapter 3)
Examples of projects Management of projects

1. Planning - goal setting, defining the project, team
► Building Construction organization
2. Scheduling - relate people, money, and supplies to

specific activities and activities to each other
3. Controlling - monitor resources, costs, quality, and

budgets; revise plans and shift resources to meet time
and cost demands
► Research Project 3 4
Project planning Project scheduling
1) Ensure that all activities are planned for

2) Their order of performance is
accounted for
5 3) The activity time estimates are recorded 6
4) The overall project time is developed
Time/cost estimates
Project controlling Project controlling Budgets
Engineering diagrams
Cash flow charts
Material availability details
Budgets
Delayed activities report
Slack activities report
CPM/PERT
► Close monitoring of resources, costs, quality, Gantt charts
budgets Milestone charts
► Feedback enables revising the project plan and 7 Cash flow schedules 8
shift resources
► Computerized tools produce extensive reports
The role of the project The role of the project
manager manager
Highly visible Highly visible Project managers should be:
Responsible for making sure that: Responsible for making sure that:
► Good coaches
1) All necessary activities are finished in order and on 1) All necessary activities are finished in order and on
► Good communicators
time time
► Able to organize activities
2) The project comes in within budget 2) The project comes in within budget
from a variety of disciplines
3) The project meets quality goals 3) The project meets quality goals
4) The people assigned to the project receive motivation, 4) The people assigned to the project receive motivation,
direction, and information direction, and information
9 10
Ethical issues Work Breakdown Structure

► Project managers face many ethical decisions on a
daily basis Level
1. Project
► The Project Management Institute has established an
ethical code to deal with problems such as: 2. Major tasks in the project
3. Subtasks in the major tasks
1) Offers of gifts from contractors
2) Pressure to alter status reports to mask delays 4. Activities (or “work packages”)
3) False reports for charges of time and expenses to be completed
4) Pressure to compromise quality to meet schedules
11 12
Work Breakdown Structure Project Scheduling Techniques
Develop Windows 10
Level 1 1.0
Operating System
Software Cost Management System

Level 2 1.1 1.2 1.3
Design Plan Testing
► Gantt chart
Develop Design Cost Module
Level 3
GUIs 1.1.1
Tracking Reports
1.2.1 Testing 1.3.1 ► Critical Path Method (CPM)
► Program Evaluation and Review Technique (PERT)
Ensure Compatibility Develop Defect
with Earlier Versions 1.1.2 Cost/Schedule 1.2.2 Testing 1.3.2
Interface
Compatible with
Level 4 1.1.2.1
Windows 8
(Work packages)
Compatible with
1.1.2.2
Windows 7
Compatible with 13 14
Windows Vista 1.1.2.3
A Simple Gantt Chart Service For a Delta Jet

Deplaning
Passengers
Baggage claim
Baggage Container offload
Pumping
Time Fueling
Engine injection water
J F M A M J J A S Cargo and mail Container offload
Main cabin door
Galley servicing
Aft cabin door
Design Lavatory servicing Aft, center, forward
Drinking water Loading
Prototype First-class section
Cabin cleaning
Economy section
Test Cargo and mail Container/bulk loading
Galley/cabin check
Flight services
Revise Receive passengers
Operating crew Aircraft check
Production Baggage Loading
Passengers Boarding
15 0 10 20 30 4016
Time, Minutes
Figure 3.4
Project management software PERT and CPM
► There are several popular packages for managing ► Network techniques
projects ► Developed in 1950s
► CPM by DuPont for chemical plants (1957)
► Oracle Primavera ► PERT by Booz, Allen & Hamilton with the U.S. Navy,
► MindView for Polaris missile (1958)
► HP Project ► Consider precedence relationships and interdependencies
► Fast Track ► Each uses a different estimate of activity times
► Microsoft Project
17 18
Variability in Activity Times Six Steps PERT and CPM
► CPM assumes we know a fixed time estimate for each

1. Define the project and prepare the work
activity and there is no variability in activity times breakdown structure
2. Develop relationships among the activities –
► PERT uses a probability distribution for activity times to decide which activities must precede and
allow for variability
which must follow others
3. Draw the network connecting all of the
activities
20
Six Steps PERT and CPM Questions PERT and CPM
Can Answer
4. Assign time and/or cost estimates to each
activity 1. When will the entire project be completed?
2. What are the critical activities or tasks in the project?
5. Compute the longest time path through the 3. Which are the noncritical activities?
network – this is called the critical path 4. What is the probability the project will be completed
by a specific date?
6. Use the network to help plan, schedule,

monitor, and control the project
21 22
Questions PERT and CPM A Comparison of AON and

AOA Network Conventions
Can Answer
Activity on Activity Activity on
5. Is the project on schedule, behind schedule, or Node (AON) Meaning Arrow (AOA)
ahead of schedule?
6. Is the money spent equal to, less than, or greater C and D cannot
A C begin until both A C
than the budget? A and B are
(d)
7. Are there enough resources available to finish the completed
B D B D
project on time?
8. If the project must be finished in a shorter time, what C cannot begin until
is the way to accomplish this at least cost? both A and B are
A C completed A C
(e) D cannot begin until Dummy activity
B is completed
23 B D A dummy activity is
introduced in AOA B D
A Comparison of AON and Activity on node (AON) Example
AOA Network Conventions
Table 3.1 Milwaukee Paper Manufacturing’s Activities and Predecessors
Activity on Activity Activity on IMMEDIATE
Node (AON) Meaning Arrow (AOA) ACTIVITY DESCRIPTION PREDECESSORS
A Build internal components —
B and C cannot B Modify roof and floor —
begin until A is
completed C Construct collection stack A
A B D D cannot begin A B D D Pour concrete and install frame A, B
until both B and C
(f) are completed Dummy
E Build high-temperature burner C
A dummy activity activity
C F Install pollution control system C
C
is again G Install air pollution device D, E
introduced in AOA
H Inspect and test F, G
Copyright © 2020, 2017, 2014 Pearson Education, Inc. 3 - 26
AON Network for Milwaukee AON Network for Milwaukee

Paper Paper
Activity A Precedes Activity C
Activity A
A A C
(Build Internal Components)
Start Start
Activity B
Start B B D
(Modify Roof and Floor)
Activity
Figure 3.5 27 Activities A and B Figure 3.6
28
Precede Activity D
AON Network for Milwaukee Determining the Project Schedule
Paper
Perform a Critical Path Analysis
F ► The critical path is the longest path
A C through the network
► The critical path is the shortest time in
E
Start H
which the project can be completed
► Any delay in critical path activities delays
the project
B D G
► Critical path activities have no slack time
Arrows Show Precedence Figure 3.7
29 30
Relationships
Determining the Project Schedule Determining the Project Schedule

Perform a Critical Path Analysis
Table 3.2 Time Estimates for Milwaukee Paper Manufacturing
ACTIVITY DESCRIPTION TIME (WEEKS) Earliest start (ES) =earliest time at which an activity can start, assuming
all predecessors have been completed
A Build internal components 2
ES = Max {EF of all immediate predecessors}
B Modify roof and floor 3
Earliest finish (EF) =earliest time at which an activity can be finished
C Construct collection stack 2
D Pour concrete and install frame 4
EF = ES + Activity time
E Build high-temperature burner 4 Latest start (LS) =latest time at which an activity can start so as to not
delay the completion time of the entire project
F Install pollution control system 3
LS = LF – Activity time
G Install air pollution device 5
Latest finish (LF) =latest time by which an activity has to be finished so as
H Inspect and test 2 to not delay the completion time of the entire
Total time (weeks) 25 project
31 32
LF = Min{LS of all immediate following activities}
Computing Slack Time Computing Slack Time
After computing the ES, EF, LS, and LF times for TABLE 3.3 Milwaukee Paper’s Schedule and Slack Times
all activities, compute the slack or free time for EARLIEST EARLIEST LATEST LATEST ON
START FINISH START FINISH SLACK CRITICAL
each activity ACTIVITY ES EF LS LF LS – ES PATH
A 0 2 0 2 0 Yes
► Slack is the length of time an activity can be B 0 3 1 4 1 No

delayed without delaying the entire project C 2 4 2 4 0 Yes
D 3 7 4 8 1 No
E 4 8 4 8 0 Yes
F 4 7 10 13 6 No
Slack = LS – ES or Slack = LF – EF
G 8 13 8 13 0 Yes
H 13 15 13 15 0 Yes
33 34
Activities with zero slack are on the critical path. It:

Computing Slack Time
► Starts at the first activity in the project Determining the Project Schedule
► Terminates at the last activity in the project
► Includes only critical activities

TABLE 3.3 Milwaukee Paper’s Schedule and Slack Times Activity Format
EARLIEST EARLIEST LATEST LATEST ON
START FINISH START FINISH SLACK CRITICAL
ACTIVITY ES EF LS LF LS – ES PATH
Activity Name
A 0 2 0 2 0 Yes
or Symbol
B 0 3 1 4 1 No A Earliest
Earliest ES EF
C 2 4 2 4 0 Yes Finish
Start
D 3 7 4 8 1 No
E 4 8 4 8 0 Yes
F 4 7 10 13 6 No Latest LS LF Latest
G 8 13 8 13 0 Yes Start 2 Finish
H 13 15 13 15 0 Yes
35 36
Activity Duration
Critical Path for Critical Path for Forward Pass
ES = Max {EF of all immediate predecessors}
Milwaukee Paper Milwaukee Paper EF = ES + Activity time
A C F A C F
0 2 2 4 4 7
2 2 3 2 2 3
Start Start
E H 0 0 E H
4 8 13 15
0 4 2 0 4 2
B D G B D G
0 3 3 7 8 13
3 4 5 37 3 4 5 38
Backward Pass
Critical Path for LS = LF – Activity time Critical Path for
Milwaukee Paper Milwaukee Paper
LF = Min{LS of all immediate
following activities}
A C F A C F
0 2 2 4 4 7 0 2 2 4 4 7
0 2 2 2 2 4 10 3 13 0 2 2 2 2 4 10 3 13
Start Start
0 0 E H 0 0 E H
4 8 13 15 4 8 13 15
0 0 0 4 4 8 13 2 15 0 0 0 4 4 8 13 2 15
B D G B D G
0 3 3 7 8 13 0 3 3 7 8 13
1 3 4 4 4 8 8 5 13 1 3 4 4 4 8 8 5 13
39 40
ES – EF Gantt Chart LS – LF Gantt Chart
for Milwaukee Paper for Milwaukee Paper
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
A Build internal A Build internal

components components
B Modify roof and floor B Modify roof and floor
C Construct collection C Construct collection
stack stack
D Pour concrete and D Pour concrete and
install frame install frame
E Build high-temperature E Build high-temperature
burner burner
F Install pollution control F Install pollution control
system system
G Install air pollution G Install air pollution
device device
41 42
H Inspect and test H Inspect and test
Variability in Activity Times

► CPM assumes we know a fixed time estimate for each
activity and there is no variability in activity times
Session 3 ► PERT uses a probability distribution for activity times to

allow for variability
Project Management 2
Fall 2021
UC3M 2
Professor Gemma Berenguer Source: Jay Heizer, Barry Render and Chuck Munson "Operations Management: sustainability and
supply chain management". Pearson, 12th Edition.(chapter 3)
Variability in Activity Times Variability in Activity Times
Estimate follows beta distribution
► Three-time estimates are required
Expected activity time:
► Optimistic time (a) – if everything goes according to plan
t = (a + 4m + b)/6
► Pessimistic time (b) – assuming very
unfavorable conditions
Variance of activity completion times:
► Most likely time (m) – most realistic estimate
v = [(b – a)/6]2
3 4
Variability in Activity Times Computing Variance

TABLE 3.4 Time Estimates (in weeks) for Milwaukee Paper's Project
Estimate follows beta distribution MOST
OPTIMISTIC LIKELY PESSIMISTIC EXPECTED TIME VARIANCE
ACTIVITY a m b t = (a + 4m + b)/6 [(b – a)/6]2
Figure 3.11
Expected activity time: A 1 2 3 2 .11
t = (a + 4m + b)/6 B 2 3 4 3 .11
Probability tof =
1 in(a + 4m + b)/6 C 1 2 3 2 .11
100
Probability
v = [(b
of < a occurring
Variance of activity − a)/6]2Probability
completion times:
of 1 in D 2 4 6 4 .44
100 E 1 4 7 4 1.00
of > b occurring F 1 2 9 3 1.78
v = [(b – a)/6]2 Activity G 3 4 11 5 1.78
Time
H 1 2 3 2 .11
Optimistic Most Likely Pessimistic 5 6

Time (a) Time (m) Time (b)
Probability of Project Completion Probability of Project Completion
Project variance is computed by summing
the variances of critical activities
Project variance is computed by
summing the variances of critical
activities Project variance
s2p = .11 + .11 + 1.00 + 1.78 + .11 = 3.11
s2p = Project variance
= (variances of activities
Project standard deviation
on critical path)
sp = Project variance
7 = 3.11 = 1.76 weeks 8

Figure 3.12
PERT makes two more assumptions: Standard deviation = 1.76 weeks
► Total project completion times follow a

normal probability distribution
► Activity times are statistically independent

When activity times cannot be assumed to be 15 Weeks
independent, simulation is often used 9 10
(Expected Completion Time,
found via Critical Path Analysis)
From Appendix I
What is the probability this project can .00 .01 .07 .08
be completed on or before the 16 week .1 .50000 .50399 .52790 .53188
deadline? .2 .53983 .54380 .56749 .57142
Z = Due – Expected date /sp .5 = due .69497

Z.69146 − expected.71566
date /s
.71904
p
date of completion date of completion
.6 .72575 .72907 .74857 .75175
= (16 weeks – 15 weeks)/1.76 = (16 wks − 15 wks)/1.76
= 0.57 Where Z is the number of = 0.57 Where Z is the number of

standard deviations the due standard deviations the due
date or target date lies from the 11 date or target date lies from the 12
mean or expected date mean or expected date
Probability of Project Completion Variability of Completion Time for

Noncritical Paths
0.57 Standard deviations
Probability ► Variability of times for activities on noncritical paths
(T ≤ 16 weeks) must be considered when finding the probability of
is 71.57% finishing in a specified time
► Variation in noncritical activity may cause change in
critical path
We are not including this consideration in our

probability calculations but it is something to
15 16 Time keep in mind, specially for those noncritical
Weeks Weeks activities with large variances
13 14
Figure 3.13
What Project Management Has Budget Control
Provided So Far • Budget control is an important aspect of project management
1. The project’s expected completion time is • Costs can exceed budget
• Overly optimistic time estimates
15 weeks • Unforeseen events
2. There is a 71.57% chance the equipment • Unless corrective action is taken, serious cost overruns can occur
will be in place by the 16-week deadline
3. Five activities (A, C, E, G, and H) are on
the critical path
4. Three activities (B, D, F) are not on the
critical path and have slack time
15 16
5. A detailed schedule is available
Time-Cost Trade-Offs Time-Cost Trade-Offs: Crashing

• Activity time estimates are made for some given level of • Crashing
resources • Shortening activity durations
• It may be possible to reduce the duration of a project by • Typically, involves the use of additional funds to support additional
personnel or more efficient equipment, and the relaxing of some
injecting additional resources work specifications
• Motivations: • The project duration may be shortened by increasing direct
• To avoid late penalties expenses, thereby realizing savings in indirect project costs
• Monetary incentives
• Free resources for use on other projects
17 18
Cost–Time Trade-Offs and Project Factors to Consider When
Crashing Crashing a Project
It is not uncommon to face the following
situations: ► The amount by which an activity is crashed is, in fact,
permissible
► The project is behind schedule ► Taken together, the shortened activity durations will
enable us to finish the project by the due date
► The completion time has been ► The total cost of crashing is as small as possible
moved forward
Shortening the duration of the

project is called project crashing 19 20
Steps in Project Crashing Steps in Project Crashing

Step 3: If there is only one critical path, then
Step 1: Compute the crash cost per time period. If select the activity on this critical path
crash costs are linear over time:
that (a) can still be crashed, and (b) has
Crash cost (Crash cost – Normal cost) the smallest crash cost per period. If
per period = (Normal time – Crash time) there is more than one critical path,
then select one activity from each
Step 2: Using current activity times, find the critical critical path such that (a) each selected
path(s) and identify the critical activities activity can still be crashed, and (b) the
total crash cost of all selected activities
21 is the smallest. Crash each selected 22
activity by one period.
Steps in Project Crashing Crashing The Project
TABLE 3.5 Normal and Crash Data for Milwaukee Paper Manufacturing
Step 4: Update all activity times. If the desired TIME (WEEKS) COST ($)
CRASH COST CRITICAL
due date has been reached, stop. If not, ACTIVITY NORMAL CRASH NORMAL CRASH PER WEEK ($) PATH ?
A 2 1 22,000 22,750 750 Yes
return to Step 2.
B 3 1 30,000 34,000 2,000 No
C 2 1 26,000 27,000 1,000 Yes
D 4 3 48,000 49,000 1,000 No
E 4 2 56,000 58,000 1,000 Yes
F 3 2 30,000 30,500 500 No
G 5 2 80,000 84,500 1,500 Yes
H 2 1 16,000 19,000 3,000 Yes

23 24
Crash and Normal Times and Critical Path and Slack Times for
Costs for Activity B Milwaukee Paper
Figure 3.16
Activity
Cost Crash A C F
Crash Cost – Normal Cost 0 2 2 4 4 7
$34,000 — Crash Cost/Wk =
Normal Time – Crash Time 0 2 2 2 2 4 10 3 13
Crash $33,000 — $34,000 – $30,000
=
Cost $32,000 — 3–1 Start Slack = 0 E Slack = 6 H
$4,000 0 0 Slack = 0 4 8 13 15
= = $2,000/Week
$31,000 — 2 Wks
0 0 0 4 4 8 13 2 15
$30,000 —
Normal B D Slack = 0 G Slack = 0
Normal — 0 3 3 7 8 13
Cost | | | 1 3 4 4 4 8 8 5 13
25 26
Figure 3.15 1 2 3 Time
Crash(Weeks) Normal Slack = 1 Slack = 1 Slack = 0
Time Time
Crashing Activity A One Week Crashing Activity A One Week
Figure 3.16 (revised)
TIME (WEEKS) COST ($) A C F

ACTIVITY NORMAL CRASH NORMAL CRASH
CRASH COST
PER WEEK ($)
CRITICAL
PATH ?
0 1 1 3 3 6
A 2 1 22,000 22,750 750 Yes 0 1 1 1 2 3 9 3 12
B 3 1 30,000 34,000 2,000 No
C 2 1 26,000 27,000 1,000 Yes

Start
Slack = 0 Slack = 0 E Slack = 6 H
0 0 3 7 12 14
D 4 3 48,000 49,000 1,000 No
E 4 2 56,000 58,000 1,000 Yes

0 0 0 3 4 7 12 2 14
F 3 2 30,000 30,500 500 No
B D Slack = 0 G Slack = 0
G 5 2 80,000 84,500 1,500 Yes 0 3 3 7 7 12
H 2 1 16,000 19,000 3,000 Yes
0 3 3 3 4 7 7 5 12
27 28
Slack = 0 Slack = 0 Slack = 0
Crashing Activity G One Week Crashing Activity G One Week

TIME (WEEKS) COST ($) A C F

ACTIVITY NORMAL CRASH NORMAL CRASH
CRASH COST
PER WEEK ($)
CRITICAL
PATH ?
0 1 1 3 3 6
A 2 1 22,000 22,750 750 Yes 0 1 1 1 2 3 9 3 12
B 3 1 30,000 34,000 2,000 Yes
C 2 1 26,000 27,000 1,000 Yes

Start
Slack = 0 Slack = 0 E Slack = 6 H
0 0 3 7 12 14
D 4 3 48,000 49,000 1,000 Yes
E 4 2 56,000 58,000 1,000 Yes

0 0 0 3 4 7 12 2 14
F 3 2 30,000 30,500 500 No
B D Slack = 0 G Slack = 0
G 5 2 80,000 84,500 1,500 Yes 0 3 3 7 7 12
H 2 1 16,000 19,000 3,000 Yes
0 3 3 3 4 7 7 5 12
29 30
We prefer crashing G because the total crashing per week Slack = 0 Slack = 0 Slack = 0
cost is 1,500 because G is part of both critical paths!
Crashing Activity G One Week Advantages of PERT/CPM
A C F
1. Especially useful when scheduling and
0 1 1 3 3 6 controlling large projects
0 1 1 1 2 3 8 3 11 2. Straightforward concept and not
Start
Slack = 0 Slack = 0 E Slack = 5 H mathematically complex
0 0 3 7 11 13
3. Graphical networks help highlight
0 0 0 3 4 7 11 2 13
relationships among project activities
B D Slack = 0 G Slack = 0 4. Critical path and slack time analyses help
0 3 3 7 7 11
pinpoint activities that need to be closely
0 3 3 3 4 7 7 4 11
31 watched 32
Slack = 0 Slack = 0 Slack = 0
Advantages of PERT/CPM Limitations of PERT/CPM

1. Project activities have to be clearly defined,
5. Project documentation and graphics point independent, and stable in their
out who is responsible for various activities relationships
6. Applicable to a wide variety of projects 2. Precedence relationships must be specified
7. Useful in monitoring not only schedules but and networked together
costs as well 3. Time estimates tend to be subjective and
are subject to fudging by managers
4. There is an inherent danger of too much
emphasis being placed on the longest, or
33 critical, path 34
Using Microsoft Project Using Microsoft Project
Program 3.1 35 36
Program 3.2
Using Microsoft Project
Session 5
Forecasting 1
Pollution Project Percentage Completed on April 9
ACTIVITY COMPLETED ACTIVITY COMPLETED Program 3.3
A 100 E 20
Fall 2021
B 100 F 20 Department of Business Administration
C 100 G 0 37 UC3M
D 10 H 0
Agenda Characteristics of Forecasts
• Forecasting 1 1. Forecasts are always inaccurate
2. Long-term forecasts are usually less accurate than short-
term forecasts
Gemma Berenguer
Gemma Berenguer
3. Aggregate forecasts are usually more accurate than
disaggregate forecasts
4. In general, the farther up the supply chain a company is,
the greater is the distortion of information it receives
2 3
Components and Methods Components and Methods

1. Time Series
• Companies must identify the factors that
• Use historical demand only
influence future demand and then ascertain the
• Best with stable demand
relationship between these factors and future
Gemma Berenguer
Gemma Berenguer
demand 2. Qualitative
• Primarily subjective
• Rely on judgment
3. Causal
• Relationship between demand and some other factor
4. Simulation
4 • Imitate consumer choices that give rise to demand 5
Components of an Observation Behaviors
Observed demand (O) = systematic component (S) • Level
Demand
+ random component (R) No Pattern
• Systematic component – expected value of demand
Gemma Berenguer
Gemma Berenguer
• Trend Time
− Level (current deseasonalized demand) Linear (default) or Nonlinear

− Trend (growth or decline in demand)
− Seasonality (predictable seasonal fluctuation)
• Random component – part of forecast that deviates • Seasonality
from systematic component Repetition at Fixed Intervals
• Forecast error – difference between forecast and actual
demand 6 7
Components of Demand Time Series Models

Trend
component
Forecasting Method Applicability

Demand for product or service
Seasonal peaks
Gemma Berenguer
Moving average No trend or seasonality
Actual demand
line Exponential smoothing No trend or seasonality
Average demand Holt’s model Trend but no seasonality

over 4 years
Winter’s model Trend and seasonality (not in this course)
Random variation
| | | |
1 2 3 4
Time (years) 9
Figure 4.1
Steps in Forecasting Moving Average
• Initialize ❑Used when demand has no observable trend or seasonality
• Compute initial estimates of level (L0), trend (T0), and ❑Systematic component of demand = level
seasonal factors (S1,…,Sp), p: periodicity of seasonal
demand ❑The level in period t is the average demand over the last N
Gemma Berenguer
Gemma Berenguer
• Forecast periods
• Forecast demand for period t + 1 (Ft+1) Lt = (Dt + Dt-1 + … + Dt–N+1) / N
• Estimate error ❑The forecast in period t+1 is
• Compute error Et+1 = Ft+1 – Dt+1 Ft+1 = Lt
• Revise estimates (adaptive)
• Modify the estimates of level (Lt+1), trend (Tt+1), and ❑After observing the demand for period t + 1, revise the
seasonal factor (St+p+1) 10 estimates 11
Lt+1 = (Dt+1 + Dt + … + Dt-N+2) / N, Ft+2 = Lt+1
Moving Average Example Moving Average Example

❑Forecast demand for Period 5
• A supermarket has experienced weekly demand of milk of D1 =
120, D2 = 127, D3 = 114, and D4 = 122 gallons over the past four
weeks:
❑Error if demand in Period 5 = 125 gallons
Gemma Berenguer
Gemma Berenguer
• Forecast demand for Period 5 using a four-period moving average
120,75
• What is the forecast error if demand in Period 5 turns out to be 125

gallons? ❑ Next, step revised demand
- 4,25
12
Comments on MA Exponential Smoothing
• What historical data (how many past periods) does this ❑Used when demand has no observable trend or seasonality
method use?
❑Systematic component of demand = level
N
❑The level in period t is
Gemma Berenguer
Gemma Berenguer
• What weight is given to old and new observations?
Lt = a Dt +(1– a )Ft , L0 = D0
The same ❑The forecast in period t+1 is
• What is the impact of N? Ft+1 = Lt

As N increases, the model is more stable, but less responsive to the most recently observed demand ❑After observing the demand for period t + 1, revise the estimates
Ft+2 = Lt+1 = a Dt+1 +(1– a )Ft+1

14 15
Example Exponential
Comments on ES
Smoothing
• A supermarket has experienced weekly demand of milk of D1 = 120, • What historical data (how many past periods) does this
D2 = 127, D3 = 114, and D4 = 122 gallons over the past four weeks method use?
• Forecast demand for Period 5 using exponential smoothing with a = 0.1 ALL
Gemma Berenguer
Gemma Berenguer
• What weight is given to old and new observations?
More weight to the most recent one
• What is the forecast error if demand in Period 5 turns out to be 125 • What is the impact of α?
gallons? As "alpha" increases, the process is more responsive to changes in recent demand
16 17
Trend-Corrected Exponential
Example Holt’s Model
Smoothing (Holt’s Model)
• Appropriate when the demand is assumed to have a level and • MP3 player demand
trend in the systematic component of demand but no seasonality D1 = 8,415, D2 = 8,732, D3 = 9,014,
Systematic component of demand = level + trend D4 = 9,808, D5 = 10,413, D6 = 11,961
a = 0.1, b = 0.2
Gemma Berenguer
Gemma Berenguer
• Obtain initial estimate of level and trend by running a linear • Using regression analysis we obtain
regression
Dt = at + b, T0 = a, L0 = b L0 = 7,367(intercept) and T0 = 673(slope)
• Forecast for period 1 is
• In Period t the forecast for future periods is
F1 = L0 + T0 = 7,367 + 673 = 8,040
Ft+1 = Lt + Tt and Ft+n = Lt + nTt
• After observing demand of time t+1, revise
Lt+1 = aDt+1 + (1 – a)(Lt + Tt), Tt+1 = b(Lt+1 – Lt) + (1 – b)Tt 18 19
Example Holt’s Model Measures of forecast error

• What is the forecast of period 7?
• Error E = F – D
t t t
𝑛
• Bias 𝑏𝑖𝑎𝑠𝑛 = ෍ 𝐸𝑡
Gemma Berenguer
Gemma Berenguer
𝑡=1
• Tracking Signal (TS) = Bias/MAD <= |6|
𝑛
1
• Mean absolute deviation (MAD) 𝑀𝐴𝐷𝑛 = 𝑛 ෍ 𝐴𝑡 , 𝐴𝑡 = |𝐸𝑡 |
𝑛 𝑡=1
1
• Mean squared error (MSE) 𝑀𝑆𝐸𝑛 = ෍ 𝐸𝑡2
𝑛
𝑡=1 𝐸𝑡
σ𝑛𝑡=1 100
• Mean absolute percent error (MAPE) 𝐷𝑡
20 𝑀𝐴𝑃𝐸𝑛 = 21
𝑛
Several pointers for quantitative
Criteria for selecting method
methods forecasting
• Plot your historical demand and look for patterns: level, trend,
and seasonality • Don’t forget to distinguish between sales data and demand
• Forecast errors provide important information for choosing data
the best forecasting method • Bias is the worst kind of forecast error; strive for zero bias
Gemma Berenguer
Gemma Berenguer
• Forecast errors can also be used to select the best values for • Whenever possible, forecast at aggregate levels. Forecast in
the parameters needed for the method detail only where necessary
• Criteria used:
• Minimizing bias;
• Minimizing MAPE, MAD or MSE;
• Meeting managerial expectations
22 23
Time Series Models
Forecasting Method Applicability
Management
MGMT660 Introduction to Operations
Session 6
Moving average No trend or seasonality
Exponential smoothing No trend or seasonality
Forecasting 2 Holt’s model Trend but no seasonality
Winter’s model Trend and seasonality

Fall 2021
UC3M 2
Trend- and Seasonality-Corrected Adaptive Smoothing
Exponential Smoothing (Winter’s model)
• Appropriate when the systematic component of demand is assumed to have ► It’s possible to use the computer to continually
a level, trend, and seasonal factor
monitor forecast error and adjust the values of
Systematic component = (level + trend) x seasonal factor the a and b coefficients used in exponential
F1 = (L0 + T0) x S1 , Ft+i = (Lt + i*Tt) x St+i i=1,…,p
smoothing to continually minimize forecast
error
• After observing demand for period t + 1, revise estimates
for level, trend, and seasonal factors ► This technique is called adaptive
Lt+1 = a(Dt+1/St+1) + (1 – a)(Lt + Tt) smoothing
Tt+1 = b(Lt+1 – Lt) + (1 – b)Tt
St+p+1 = g(Dt+1/Lt+1) + (1 – g)St+1
a = constant for level, b = constant for trend, g = constant for seasonal 3 4
factor
Qualitative forecasting
Components and Methods
methods
1. Time Series
• Use historical demand only
• Grass roots – talk to sale force, talk to customers
• Best with stable demand
• Market research – survey of customers, test markets
2. Qualitative
• Primarily subjective • Panel consensus – bring in experts
• Rely on judgment • Executive judgment – pool input from executives
3. Causal • Historical analogy – find a similar product
• Relationship between demand and some other factor
• Delphi Method
4. Simulation
• Imitate consumer choices that give rise to demand 5 6
Jury of Executive Opinion Delphi Method Decision Makers
► Iterative group process, (Evaluate
► Involves small group of high-level experts and managers continues until consensus is responses and
make decisions)
► Group estimates demand by working together reached
► Three types of participants
► Combines managerial experience with statistical models
► Decision makers
► Relatively quick
► Staff Staff
► ‘Group-think’ (Administering survey)
disadvantage ► Respondents
7 8
Respondents
(People who can make
valuable judgments)
Several pointers for qualitative

Components and Methods
methods forecasting
1. Time Series
• Judgment forecasting is clearly needed when no quantitative • Use historical demand only
data are available • Best with stable demand
• These methods are also good to adjust quantitative forecasts
Management
Management

2. Qualitative
to improve forecast quality
• when the quantitative method forecast tends to be inaccurate • Primarily subjective
and the decision maker has important contextual knowledge • Rely on judgment
• when a specific event might occur such as an advertising
campaign, the actions of competitors, or international
3. Causal/Associative
developments • Relationship between demand and some other factor
4. Simulation
9 • Imitate consumer choices that give rise to demand 10
Causal (Associative) Causal (Associative) methods
Forecasting
Forecasting an outcome based on predictor
variables using the least squares technique Let
• Y be the quantity to be forecasted
• (X1, X2, . . . , Xn) are n variables that have predictive power for Y
ŷ = a + bx A causal model is Y = f (X1, X2, . . . , Xn).
A typical relationship is a linear one:
where ŷ = value of the dependent variable (in our Y = a0 + a1X1 + . . . + an Xn
example, sales)
a = y - axis intercept
b = slope of the regression line
x = the independent variable 11 12
Values of Dependent Variable (y-values)
Least Squares Method Least Squares Method

Actual observation Deviation7 Equations to calculate the regression variables
(y-value)
Deviation5 Deviation6 ŷ = a + bx
Deviation3
å
Least squares method minimizes
the sum of the
Deviation4 squared errors xy - nxy
(deviations) b=
Deviation1
(error) Deviation2
Trend line, y ^= a + bx
å x - nx 2 2
|
1
|
2
|
3
|
4
|
5
|
6
|
7
a = y - bx
Figure 4.4
Time period
Copyright © 2020, 2017, 2014 Pearson Education, Inc. 4 - 13 Copyright © 2020, 2017, 2014 Pearson Education, Inc. 4 - 14
Least Squares Example Least Squares Example
ELECTRICAL POWER
YEAR (x) DEMAND (y) x2 xy
ELECTRICAL ELECTRICAL
1 74 1 74
YEAR POWER DEMAND YEAR POWER DEMAND
2 79 4 158
1 74 5 105
3 80 9 240
2 79 6 142 4 90 16 360
3 80 7 122 5 105 25 525
4 90 6 142 36 852
7 122 49 854
Σx = 28 Σy = 692 Σx2 = 140 Σxy = 3,063
x=
å x = 28 = 4 y=
å y = 692 = 98.86
n 7 n 7
15 16
Least Squares Example Least Squares Example

- nxy 3,063 - ( 7) ( 4) (98.86) 295
å xyELECTRICAL Trend line,
b= = POWER = = 10.54 – y^ = 56.70 + 10.54x
YEAR (x) å x - nxDEMAND (y)140 - (7) ( 4 ) x 28
2 2 2 2 xy
160
1 74 1 74 Power demand (megawatts) 150 –
() –
2 79 4 158
a = y - bx = 98.8680
-10.54 4 = 56.70 140
3 9 240
4 90 16 360 130 –
5 105 ŷ = 56.70 +10.54x25
Thus, 525 120 –
6 142 36 852
110 –
7 122 49 854
Σx = 28 Σy = 692 Σx2 = 140 Σxy = 3,063 100 –
90 –
x=
å x =in28year
Demand
=4 y=
å y+=10.54(8)
8 = 56.70 692
= 98.86 80 –
n 7 = 141.02,
n or 7141 megawatts
17 70 – | | | | | | | | | 18
60 – 1 2 3 4 5 6 7 8 9
Year Figure 4.5
50 –
Associative Forecasting Example Associative Forecasting Example
NODEL’S SALES AREA PAYROLL NODEL’S SALES AREA PAYROLL SALES, y PAYROLL, x x2 xy
(IN $ MILLIONS), y (IN $ BILLIONS), x (IN $ MILLIONS), y (IN $ BILLIONS), x 2.0 1 1 2.0
2.0 1 2.0 2 3.0 3 9 9.0
3.0 3 2.0 1 2.5 4 16 10.0
2.5 4 3.5 7 2.0 2 4 4.0
2.0 1 1 2.0
4.0 –
3.5 7 49 24.5
Nodel’s sales
(in $ millions)
3.0 – Σy = 15.0 Σx = 18 Σx2 = 80 Σxy = 51.5
2.0 –
x=
å x = 18 = 3 y=
å y = 15 = 2.5
1.0 – 6 6 6 6
| | | | | | |
19 b=
å xy - nxy = 51.5 - (6)(3)(2.5) = .25 a = y - bx = 2.5 - (.25)(3) = 1.75 20
0 1 2 3 4 5
Area payroll (in $ billions)
6 7
å x - nx
2 2
80 - (6)(3 ) 2

SALES, y PAYROLL, x x2 xy SALES, y PAYROLL, x x2 xy
2.0 1 1 2.0 2.0 1 1 2.0
4.0 –
3.0 3
ŷ = 1.75
9
+ .25x 9.0 3.0 3
ŷ = 1.75
9
+ .25x 9.0
2.5 4 16 10.0 2.5 4 16 10.0
Nodel’s sales
Sales = 1.75
4 + .25(payroll) Sales = 1.75
4 + .25(payroll)
(in $ millions)
2.0 2 4.0 2.0 3.0 – 2 4.0
2.0 1 1 2.0 2.0 1 1 2.0
2.0 –
3.5 7 49 24.5 3.5 7 49 24.5
Σy = 15.0 Σx = 18 Σx2 = 80 Σxy = 51.5 Σy = 15.0 Σx =–
1.0 18 Σx2 = 80 Σxy = 51.5
x=
å x = 18 = 3 y=
å y = 15 = 2.5 x=
å x 0= 18 1=| 3 |
2 y =3
| åy
=4
|
15
=52.5 6
| | |
7
6 6 6 6 6 6 6 (in 6$ billions)
Area payroll
b=
å xy - nxy = 51.5 - (6)(3)(2.5) = .25 a = y - bx = 2.5 - (.25)(3) = 1.75 21 b=
å xy - nxy = 51.5 - (6)(3)(2.5) = .25 a = y - bx = 2.5 - (.25)(3) = 1.75 22
å x - nx
2 2
80 - (6)(3 ) 2
å x - nx
2 2
80 - (6)(3 ) 2
If payroll next year is estimated to be $6 billion, If payroll next4.0year

– is estimated to be $6 billion,
then: then: 3.25
Nodel’s sales
(in $ millions)
3.0 –
2.0 –
Sales (in $ millions) = 1.75 + .25(6) Sales (in$ millions) = 1.75 + .25(6)
= 1.75 + 1.5 = 3.25 1.0 – = 1.75 + 1.5 = 3.25
| | | | | | |
0 1 2 3 4 5 6 7
Sales = $3,250,000 Sales = $3,250,000
23 24
Standard Error of the Estimate Standard Error of the Estimate

► A forecast is just a point estimate of a future
value
S y,x =
å( y - y ) c
2
► This point is
actually the
n-2
4.0 –
mean or 3.25
Nodel’s sales
(in $ millions)
3.0 –
expected Regression line, where y = y-value of each data point
value of a 2.0 – ŷ =1.75+.25x yc = computed value of the dependent variable, from
probability the regression equation
1.0 –
distribution n = number of data points
| | | | | | |
0 1 2 3 4 5 6 7
Figure 4.9 Area payroll (in $ billions) 25 26
Standard Error of the Estimate Standard Error of the Estimate
σ 𝑦 2 − 𝑎 σ 𝑦 − 𝑏 σ 𝑥𝑦 39.5 − 1.75 15.0 − .25 51.5
𝑆𝑦,𝑥 = =
Computationally, this equation is 𝑛−2 6−2
considerably easier to use
= .09375
= .306 𝑖𝑛 $ 𝑚𝑖𝑙𝑙𝑖𝑜𝑛𝑠
σ 𝑦 2 − 𝑎 σ 𝑦 − 𝑏 σ 𝑥𝑦
𝑆𝑦,𝑥 = The standard error of 3.6 –
– + .306
𝑛−2 3.5
Nodel’s sales
(in $ millions)
the estimate is 3.4 –
3.3 –
$306,000 in sales 3.2 –
3.1 –
We use the standard error to set up 3.0 –
2.9 – – .306
prediction intervals around the point There is a 68.27% chance
of sales being + or –
estimate 27 $306,000 from the point
5
6
28
estimate $3,250,000
Correlation Correlation Coefficient

Figure 4.10
y y
► How strong is the linear relationship between the
variables?
► Correlation does not necessarily imply causality! x x
(a) Perfect negative (e) Perfect positive
► Coefficient of correlation, r, measures degree of correlation, r = –1 y y correlation, r = 1
association
► Values range from −1 to +1
y
x x
(b) Negative correlation (d) Positive correlation
x
(c) No correlation, r = 0
High Moderate Low Low Moderate High

| | | | | | | | |
29 –1.0 –0.8 –0.6 –0.4 –0.2 0 0.2 0.4 0.6 0.8 1.0
30
Correlation coefficient values
Correlation Coefficient Correlation Coefficient
y x x2 xy y2
2.0 1 1 2.0 4.0
3.0 3 9 9.0 9.0
2.5 4 16 10.0 6.25
2.0 2 4 4.0 4.0
2.0 1 1 2.0 4.0
3.5 7 49 24.5 12.25
Σy = 15.0 Σx = 18 Σx2 = 80 Σxy = 51.5 Σy2 = 39.5
309 − 270 39 39
= = = = .901
31 (156)(12) 1,872 43.3 32
Multiple-Regression Analysis Multiple-Regression Analysis

In the Nodel example, including interest rates in the model
If more than one independent variable is to be used in gives the new equation:
the model, linear regression can be extended to
multiple regression to accommodate several ŷ = 1.80 +.30x1 - 5.0x2
independent variables
An improved correlation coefficient of r = .96 suggests this
ŷ = a + b1x1 + b2 x2 model does a better job of predicting the change in
construction sales
Computationally, this is quite complex
and generally done on the computer Sales = 1.80 + .30(6) − 5.0(.12) = 3.00
33 Sales = $3,000,000 34
Sales & Operations Planning Sales & Operations Planning
• Prepare for the • Coordinate the firm
future • Production Plans
Data Gathering Demand Supply Planning Pre-Meeting Executive • Gather available • Sales Plans
Planning Meeting
•Demand •Forecast •Inventory •Sales Forecast •Sales Forecast

information • Marketing Plans
• Obtain additional
•Leading •Sales Targets •Orders •Inventory •Inventory
Indicators
•Assumptions
•Promotions
•Pricing
•Capacity
•Rationing
Projections
•Capacity Plan
Projections
•Capacity Plan • Procurement Plans
•Customer •Customer
information
Service
Targets
Service
Targets • Cash Flow Plans
•Annual •Annual
• Make effective use of
Budget
Projections
Budget
Projections • Investor Relations
•New Product •New Product information
Intros Intros
• Capacity Plans
36
Productivity Improvements General pointers

• Demand forecasting is being done in virtually every company.
• Coca Cola: a 15% reduction in The challenge is to do it better than the competition
forecast errors lead to a 25%
reduction in inventory (Clarke 2006) • Better forecast result in better customer service and lower
costs, as well as better relationships with suppliers and
• Hewlett Packard: a 10% reduction in customers
forecast errors lead to an increase in
on-time deliveries from 60% to 95% • The forecast can and must make sense based on the big
(Shan et al. 2009) picture, economic outlook, market share, and so on
• Collaboration and communication can be more fruitful than

37 using the most advanced technique or model 38
- Clark, S. 2006. Managing the introduction of a structured forecast process: Transformation lessons from Coca-
Cola Enterprises Inc. Foresight, 4: 21-25.
- Shan, J. Z., J. Ward, S. Jain, J. Beltran, F. Amirjalayer, Y.-W. Kim. 2009. Spare-parts forecasting: A case study at
Hewlett-Packard. Foresight, 14: 40-47.
What is a process?
• A process is a series of independent tasks that transform an
input into output material of higher value for the organization
Gemma Berenguer
Session 8 Input Transformation Output
Process Analysis 1
Fall 2021
UC3M 2
Process examples Process Analysis

• Let’s look at the “transformation” in more detail…
• Honda: Steel → Assembly → Cars
• Why do we need to analyze the process?
Gemma Berenguer
Gemma Berenguer
• Bank: Deposits → Investments → More money • To identify inefficient tasks
• To spot possible effectiveness improvement tasks
• Hospital: Sick people → Treatment → Healthy • To understand where value can be added
• How can we analyze a process?

• Etc.
• Step 1: Map It
3 4
Make-to-Order vs. Make-to-
Process Flow Charts
stock
• Graphical description of a process: Task 1 Task 2
Demand
FG
• Holding:
Gemma Berenguer
Gemma Berenguer
• Raw Materials, RM Task 1 Task 2
• Work in Process, WIP
• Finished Goods Inventory, FGI
• Flow of material or work If demand is satisfied by FG then the system is make-to-stock,
otherwise it is a make-to order system
• Processing step
5 6
Make-to-Order vs. Make-to-

Performance measure
stock
Make to stock
Old Process at McDonald’s Customer
Order
• How do we evaluate a process?
Finished
Raw
Gemma Berenguer
Gemma Berenguer
Mat. Cook Assemble
goods
Deliver • What is its capacity? How many units per unit time go through
each task? The process as a whole?
• What is the bottleneck? Which production step limits the
Make to order process capacity?
• What is the flow time? How long does it take for a unit to go
New Process at
McDonald’s Customer order through the system?
Raw
Matl. Cook Assemble Deliver
7 8
Capacity Process Example
• Cycle Time: Average time for completion of a unit at a • Let’s bake cookies…
production step or process. Measured as time/unit.
• Throughput Rate: Average number of units processed over a Spoon onto Load and Unload and
time interval. Measured as units/time. Cool
Tray Bake 1 tray Pack
Gemma Berenguer
Gemma Berenguer
7 min
2 min per tray 12 min 1 min per tray
Wash (2 min per tray)

• How long does it take a tray to go through the system?
1 • 24 minutes
Throughput rate =
Cycle Time • How many times we can process a tray per hour?
• 60/24 = 2.5 times
Capacity = throughput rate 9 10
Process Gantt Chart Process Example

Spoon onto Load and Unload and
Cool
Tray Bake 1 tray 7 min
Pack
2 min per tray 12 min 1 min per tray Spoon onto Load and Unload and
Cool
Gemma Berenguer
Gemma Berenguer
Wash (2 min per tray) 7 min
Wash (2 min per tray)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Spoon • What is the maximum number of trays of cookies we can make

Bake
per hour?
Cool
Unload • 60/12=5 (trays/hour)

Wash
• What is the cycle time?
11
• 60/5 = 12 minutes 12
What is a bottleneck? Process Gantt Chart
• Bottleneck is the process stage with the smallest throughput • How many trays do we need?
rate (longest cycle time)
• Which task is the bottleneck? Spoon onto Load and Unload and
Cool
Pack
Gemma Berenguer
Gemma Berenguer
Spoon onto Load and Unload and Wash (2 min per tray)
Cool
Pack
2 min per tray 12 min 1 min per tray 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Spoon
Wash (2 min per tray) Bake
Cool
• Capacity of the process = capacity of the bottleneck Unload
• Capacity: 5 trays per hour 13 Wash

14
Process Example Summarize terminology

• How many trays do we need? • Throughput rate R (1 / Cycle time)
• On average, how many units pass through the system?
Spoon onto Load and Unload and • Flow time T
Cool • On average, how much time does one unit spend in the system?
Gemma Berenguer
Gemma Berenguer
7 min
2 min per tray 12 min 1 min per tray • Inventory I
Wash (2 min per tray) • On average, how many units are in the system?
• Cycle time: 12 minutes

• Little’s Law: I = R * T = 5 trays/hour * (24/60) = 5 *
• Flow time: 24 minutes 0.4 = 2 or
• How many trays do we need? I = T / Cycle time
• 24/12 = 2 trays 15 16
Little’s Law Examples Little’s Law Examples
• 1 car enters a tunnel every 5 seconds, it takes 20 minutes to • 20 customers come to a restaurant every hour. On average,
drive through a tunnel. How many cars are in the tunnel on there are 28 customers in the restaurant at a point time. How
average? long does an average customer stay in the restaurant?
• T=20 minutes
Gemma Berenguer
Gemma Berenguer
• R = 20 customers per hour
• R=60 sec/min * (1/5sec)=12 cars per
• I = 28 customers on average
minute
• T = 28/20 = 1.4 hours
• I = 20 min*12 cars/min = 240 cars in the
tunnel on average
17 18
Little’s Law Examples Little’s Law Examples

• BoilerBank loan department takes an average of 6 days to • A fast food hamburger restaurant uses 3,500 pounds of
process an application. Internal audit found that at any given hamburger mince each week. The manager of the restaurant
point in time, there are about 100 applications in the various wants to ensure that the meat is always no more than two
stages of the process. What is the monthly throughput of the days old on average when used. How much hamburger mince
Gemma Berenguer
Gemma Berenguer
department (assume 30 day months)? should be kept in the refrigerator as inventory?
• T = 6 days = 0.2 months
• I = 100 R = 3,500 pounds per week =
• R = I/T = 100/0.2 = 500 applications per month 500 pounds per day)
T = 2 days
19
I= 500 x 2 = 1,000 pounds 20
Last class review
 Throughput rate = Capacity R (aka Flow rate)
 Units/time
 Cycle Time (aka Takt time, Tk Time)
Session 10
 Time/unit it takes from a unit processed to the next one
 Flow time T (aka Throughput time)

 Time it takes for a unit to go through the process
Process Analysis 2  Inventory I
 Average volume of units in the system
Fall 2021
 Little’s Law: I = R * T or I = T / Cycle time
UC3M 2
 Bottleneck: Step with the longest cycle time
What is a bottleneck? What is a bottleneck?
Demand can be a
Input Input bottleneck
Throughput
Flow Rate
Rate Throughput
Flow Rate
Rate
Excess Excess
capacity capacity
Input can be a Throughput rate or Flow rate

3 bottleneck 4
= min{available input,
Demand, process capacity}
Impact of product mix on Impact of product mix on
bottlenecks bottlenecks
• Utilization = Flow rate / Capacity • Utilization = Flow rate / Capacity
• Example: • Example:
Cold Sandwich Cold Sandwich

Demand: 70 per hour Demand: 70 per hour
Assemble Assemble
Hot Sandwich Grill 120/hr Hot Sandwich Grill 120/hr
Demand: 50 per hour 60/hr Demand: 50 per hour 60/hr
• Utilization of the grill: • Utilization of the grill:
• 50/60 = 83.3%
• Utilization of the assembly 5 • Utilization of the assembly 6

• 120/120 = 100%
Impact of product mix on Impact of product mix on

bottlenecks bottlenecks
? per hour ? per hour
Assemble Assemble
50 per hour 60/hr 50 per hour 60/hr
Demand Demand Grill Assembly Cold Hot Grill Assembly

Cold Hot 70 50 50/60 =83.3% 120/120 = 100%
70 50 50 50 50/60 =83.3% 100/120 = 83.3%
50 50 40 50 50/60 =83.3% 90/120 = 75%
40 50
7 8
Demand is the
bottleneck
Implied Utilization Implied Utilization
Capacity requested
• Implied Utilization = Demand / Capacity available • Implied Utilization =
Capacity available
• Example: • Example:
90 per hour 90 per hour
Assemble Assemble
50 per hour 60/hr 50 per hour 60/hr
• Implied utilization of the grill: • Implied utilization of the grill: 50/60 = 83.3%
• Implied utilization of the assembly: 9 • Implied utilization of the assembly: 140/120 = 117% 10
Compare Utilization Imbalanced utilization

• Cereal processing:
• Implied Utilization • Resource (or
• Captures the Process) Utilization Flake prepping:
Packaging:
mismatch between 1.5 tons per
the demand and the • Measures how much the hour
0.5 ton per hour
capacity currently resource/process can 8 am – 12 pm 8 am – 8 pm

available at the produce, opposed to • Inventory Buildup
resource how much the
resource/process
• Can be greater than actually does produce
one 4
• Always less than or equal
to one
• Bottleneck resource 8 am noon 8 pm
utilization is equal to one 11 12
At what rate does inventory buildup?
1.5 tons/hour - 0.5 ton/hour = 1 ton/hour
Example batch process Direct Labor Content
• Custom shirts business • Assume that we have 4 workers on the cutting operation; they
work 8 hour shifts and receive $10 an hour. Current
production rate is 800 shirts a day.
Cutting Sewing Package
MGMT660
It takes 1.5 minutes to • What is the direct labor content per shirt?
roll out one layer of
fabric. Each layer may 4 * (1.5 L + 30) / 8L minutes per shirt
have up to 8 designs.
You can lay up to 60
layers. It take 30 • For L=10?
minutes to cut. 4 * (1.5 *10 + 30) / 80 = 2.25 minutes per shirt
• For L=60?
• What is the processing time of the cutting operation?
4 * (1.5 * 60 + 30) / 480 = 1 minute per shirt
1.5 x L + 30 min, where L is the number of layers 13 14
Direct Labor Content

• Assume that we have 4 workers on the cutting operation; they
work 8 hour shifts and receive $10 an hour. Current
production rate is 800 shirts a day. Assume L=60.
• What is the direct labor utilization?

Time used
=
800 shirts / day *1min ute / shirt
= 0.42
Session 12
Time available 4 wor ker s * 480 min utes / wor ker
Production Planning
• Direct labor cost (i.e. labor cost per shirt)?
Fall 2021
Direct labor cost per day 4 workers *8 hours * $10 Department of Business Administration
= = $0.4
Number of shirts per day 800 shirts per day 15 UC3M
Capacity Planning Over a Time Horizon
► The throughput, or the number of units a facility can

hold, receive, store, or produce in a period of time
► Determines
fixed costs
► Determines if
demand will
be satisfied
► Three time horizons
2 3
Source: Jay Heizer, Barry Render and Chuck Munson "Operations Management:
sustainability and supply chain management". Pearson, 12thEdition.(chapter 7s)
Design and Effective Capacity Design and Effective Capacity

TABLE S7.1 Capacity Measurements
► Design capacity is the maximum theoretical output of a MEASURE DEFINITION EXAMPLE
system
Design capacity Ideal conditions exist Machines at Frito-Lay are designed to
► Normally expressed as a rate during the time that produce 1,000 bags of chips/hr., and the plant
the system is operates 16 hrs./day.
available Design Capacity = 1,000 bags/hr. × 16 hrs.
= 16,000 bags/day
► Effective capacity is the capacity a firm expects to achieve
given current operating constraints
► Often lower than design capacity
4 5
Design and Effective Capacity Design and Effective Capacity
TABLE S7.1 Capacity Measurements TABLE S7.1 Capacity Measurements
MEASURE DEFINITION EXAMPLE MEASURE DEFINITION EXAMPLE
Effective capacity Design capacity Frito-Lay loses 3 hours of output per day Actual output Effective capacity On average, machines at Frito-Lay are not
minus lost output (= 0.5 hrs./day on preventive maintenance, minus lost output running 0.25 hr./day due to late parts and
because of planned 1 hr./day on employee breaks, and 1.5 during unplanned machine breakdowns.
resource hrs./day setting up machines for different resource idleness Actual Output = 13,000 bags/day
unavailability (e.g., products). (e.g., absenteeism, – (1,000 bags/hr.)
preventive Effective Capacity = 16,000 bags/day machine breakdowns, (0.25 hr./day)
maintenance, – (1,000 bags/hr.) unavailable parts, = 13,000 bags/day
machine (3 hrs./day) quality problems) – 250 bags/day
setups/changeovers, = 16,000 bags/day = 12,750 bags/day
changes in product – 3,000 bags/day
mix, scheduled = 13,000 bags/day
breaks)
6 7
Design
Utilization and Efficiency Bakery Example Capacity
Utilization is the percentage of design capacity actually Actual production last week = 148,000 rolls/week
achieved Effective capacity = 175,000 rolls/week
Design capacity = 1,200 rolls per hour
Utilization = Actual output/Design capacity Bakery operates 7 days/week, three 8-hour shifts
Efficiency is the percentage of effective capacity actually Design capacity = (7 x 3 x 8) x (1,200) = 201,600 rolls/week
achieved
Efficiency = Actual output/Effective capacity

8 9
Bakery Example Utilization Bakery Example Efficiency
Actual production last week = 148,000 rolls Actual production last week = 148,000 rolls
Effective capacity = 175,000 rolls Effective capacity = 175,000 rolls
Design capacity = 1,200 rolls per hour Design capacity = 1,200 rolls per hour
Bakery operates 7 days/week, 3 - 8 hour shifts Bakery operates 7 days/week, 3 - 8 hour shifts
Design capacity = (7 x 3 x 8) x (1,200) = 201,600 rolls Design capacity = (7 x 3 x 8) x (1,200) = 201,600 rolls
Utilization = 148,000/201,600 = 73.4% Utilization = 148,000/201,600 = 73.4%
10
Efficiency = 148,000/175,000 = 84.6% 11
Design Effective
Bakery Example Capacity Bakery Example Capacity
Expansion Expansion
Design capacity = 201,600 rolls per line Design capacity = 201,600 rolls per line
Efficiency = 84.6% Efficiency = 84.6%
Expected output of new line = 130,000 rolls Expected output of new line = 130,000 rolls
Design capacity = 201,600 x 2 = 403,200 rolls Design capacity = 201,600 x 2 = 403,200 rolls
Effective capacity = 175,000 x 2 = 350,000 rolls
12 13
Actual Utilization
Bakery Example Output Bakery Example Efficiency
Expansion Expansion
Design capacity = 201,600 rolls per line Design capacity = 201,600 rolls per line
Efficiency = 84.6% Efficiency = 84.6%
Expected output of new line = 130,000 rolls Expected output of new line = 130,000 rolls
Design capacity = 201,600 x 2 = 403,200 rolls Design capacity = 201,600 x 2 = 403,200 rolls
Effective capacity = 175,000 x 2 = 350,000 rolls Effective capacity = 175,000 x 2 = 350,000 rolls
Actual output = 148,000 + 130,000 = 278,000 rolls Actual output = 148,000 + 130,000 = 278,000 rolls
Utilization = 278,000/403,200 = 68.95%
14 15
Efficiency = 278,000/350,000 = 79.43%
Capacity and Strategy 10 main operations decisions
► Capacity decisions impact all 10 decisions of • 1. Design of goods and services

operations management as well as other • 2. Managing quality
functional areas of the organization • 3. Process and capacity strategy
• 4. Location strategy
• 5. Layout strategy
► Capacity decisions must be integrated into the
• 6. Human resources and job design
organization’s mission and strategy
• 7. Supply-chain management
• 8. Inventory management
• 9. Scheduling
16 17
• 10. Maintenance
Economies and Diseconomies of
Capacity Considerations Scale
Figure S7.2
1. Forecast demand accurately
(sales per square foot)

2. Match technology increments and sales volume
Average unit cost

3. Find the optimum operating size
(volume) 1,300 sq 8,000 sq ft
ft store 2,600 sq store
4. Build for change ft store
Economies Diseconomies
of scale of scale
18 1,300 2,600 8,000 19
Number of square feet in store
Copyright © 2017 Pearson Education, Inc. S7 - 19
Managing Demand Complementary Demand

Combining the
► Demand exceeds capacity
Patterns
Figure S7.3 two demand
patterns
► Curtail demand by raising prices, scheduling longer 4,000 – reduces the
lead times, discouraging marginally profitable variation
business Sales in units
3,000 –
► Long-term solution is to increase capacity Snowmobile
motor sales
► Capacity exceeds demand 2,000 –
► Stimulate demand through price reductions or
aggressive marketing 1,000 – Jet ski
engine
► Product changes sales
► Adjusting to seasonal demands 20 JF MAM JJASO NDJF MAMJJASO ND J 21

Time (months)
► Produce products with complementary demand
patterns
Tactics for Matching Capacity to Service-Sector Demand and
Demand Capacity Management
1. Making staffing changes
► Demand management
2. Adjusting equipment ► Appointment, reservations, FCFS rule
► Purchasing additional machinery ► Capacity
► Selling or leasing out existing equipment management
3. Improving processes to increase throughput ► Full time,
temporary,
4. Redesigning products to facilitate more throughput part-time
staff
5. Adding process flexibility to meet changing product
preferences
6. Closing facilities
22 23
Responding to predictable
Managing Demand
variability
• Predictable variability is change in demand that can be forecasted
• Can cause increased costs and decreased responsiveness in the • Short-term price discounts
supply chain • Promotions
Management
• Two broad approaches

• Revenue Management
1. Manage demand using short-term price discounts and trade
promotions
2. Manage supply using capacity, inventory, subcontracting, and
backlogs
24 25
Managing Supply Inventory/Capacity tradeoff
• Managing capacity
Leveling capacity forces
• Time flexibility from workforce inventory to build up in
• Use of seasonal workforce anticipation of seasonal
variation in demand
• Use of subcontracting
• Use of dual facilities – specialized and flexible
• Designing product flexibility into production processes
• Managing inventory Carrying low levels of inventory
• Build inventory of high demand or predictable demand products requires capacity to vary with
seasonal variation in demand or
• Using common components across multiple products enough capacity to cover peak
• Exploiting substitution demand during season
26 27
• Postponing product differentiation (delayed differentiation)
Postponement Postponement
• Delay product differentiation or customization until closer to

the time the product is sold
• Have common components in the supply chain for most of the

push phase
• Move product differentiation as close to the pull phase of the
supply chain as possible
• Inventories in the supply chain are mostly aggregate
28 29
Examples of delayed
Risk pooling strategies
differentiation
• Idea: Pooling separate sources of demand
• Nokia reduces uncertainty
• Customers want different color phones
• Solution: design the product so that color plates can be added
quickly and locally • Four versions of risk pooling:
• location pooling
• Private label soup manufacturer • product pooling (or universal design)
• Many different private labels (Giant, Kroger, A&P, etc.) • lead time pooling (or delayed differentiation)
• Solution: hold inventory in cans without labels, add label only • capacity pooling
when demand is realized
30 31
Location pooling Location pooling (cont.)

At Medtronic
• The location pooling strategy:
• Current operations: • A single location stores inventory used by several sales reps.
• Each sales representative has her own inventory to serve • Sales reps no longer hold their own inventory, they must pull inventory
from the pooled location.
demand in her own territory.
• Inventory is automatically replenished at the pooled location as depleted
• Lead time is 1 day from Mounds View DC by demand.
• e.g., 3 territories, 3 stockpiles of inventory • Lead time to pooled location is still 1 day from Mounds View DC. (e.g., 3
pooled territories, 1 stockpile of inventory)
32 33
Location pooling pros, cons and
alternatives
Capacity pooling
• At Toyota
• Pros: • They have 10 production facilities
• Reduces demand uncertainty, which allows a firm to reduce • They have 10 vehicles to produce (Corolla, Camry, Prius, etc).
inventory, increase service, or a combination of both • Each plant is capable of producing 100 units
• Demand for each vehicle is Normally distributed with mean 100 and standard deviation 40
• Cons: • Each plant can be configured to produce up to 10 different vehicles
• Location pooling moves inventory away from customers: • But flexibility is expensive, i.e., the cost to construct a plant is increasing in the number of
• For Medtronic this creates an inconvenience for the sales reps vehicles it can produce
• But may reduce inbound transportation cost • Toyota must decide which plants can produce which vehicles before demand is realized
• After demand is realized, Toyota can allocate its capacity to satisfy demand
• Allocation issues
• If demand exceeds capacity, sales are lost
• Alternative:
• Virtual pooling:
• Each Medtronic rep keeps her own inventory, but shares inventory
with nearby reps only if needed.
34 35
Capacity pooling (cont.) Capacity pooling (cont.)

• Flexibility allows production shifts to high selling products to avoid
• Four possible capacity configurations: no flexibility to total flexibility lost sales.
• Consider a two plant, two product example and two configurations, no
flexibility and total flexibility:
• If demand turns out to be 75 for product A, 115 for product B then..
36 37
How do you think we should
Capacity pooling (cont.)
add flexibility?
• Adding flexibility increases capacity utilization and expected sales. If we could add 10 links, how do you think we
• Note: 20 links can provide nearly the same performance as total should add them?
flexibility!
38 39
Takeaway—Risk Pooling
• The main idea is to reduce demand uncertainty by aggregating

demands (of different location, products)
Session 13
Statistical Process Control
Fall 2021
40 UC3M
Planning October
Week 7
Monday
18
(Kirsten's Cookie case)

Tuesday
19
Wednesday
Process Analysis
20
Thursday
21
Friday
22
Video
Process Analysis
October 25 26 27 28 29
Week 8 Process Analysis Quiz 2

• Frito-Lay control chart video
Statistical Process
Control
November 1 2 3 4 5
MGMT660
MGMT660

Statistical Process
Week 9 Control
https://www.youtube.com/watch?v=bXmYAkz2tCw
November 8 9 10 11 12
Week 10 Lean Operations SCM

(Toyota case)
November 15 16 17 18 19
Week 11 Humanitarian Quiz 3

Logistics Industry 4.0
November 22 23 24 25 26
Week 12 Circular Economy Virtual speaker

(Monday 16:00-
(HP case) 17:00)
November 29 30 1 2 3
Week 13 Presentations Presentations
December Final Review
2 3
December 6 7 8 9 10

Types of variation
(SPC)
• Variation is inherent in any process:
• Random variation (common causes): Natural variations in output
of a process, created by countless minor factors;
Gemma Berenguer
Gemma Berenguer
• Assignable variation (special causes): A variation whose source
can be identified.
• Goal of SPC: monitor quality of manufacturing and service

processes to identify and eliminate assignable variation
4 5
Process Quality Control Control Charts
• In Control Process
Parameter
• No special causes (no assignable variation) • Track process
Upper Control Limit (UCL) parameter over sample
• Out of Control numbers;
Gemma Berenguer
Gemma Berenguer
• Special causes are present Center Line
• Find center line
(process average);
• Find control limits;
Lower Control Limit (LCL)
Tool to determine whether the process is in control: • Plot the samples;
CONTROL CHART • Analyze the chart for
“lack of control”.
Sample number
6 7
Control Charts
Statistics of samples: e.g. sample mean, sample range
Each point represents one sample, not one output
Types of control charts
Abnormal variation Out of • Variable Control Charts (Continuous metric)
due to assignable sources control
• E.g. length, weight, time
UCL
• X-bar chart
Gemma Berenguer
Gemma Berenguer
• R chart
Common (normal) variation
• Attribute Control Charts (Discrete metric)
Abnormal variation
LCL • E.g. Number of complaints, % of orders completed on time. Etc
Out of
due to assignable sources control
• p-chart
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 • c-chart
Sample number
• All control charts follow the same principle (center line, UCL, 9
LCL), but the formulas differ
8
X-bar chart X-bar chart
• Call Center – measure call time in minutes. Take 20 samples of size 5 each. • Calculate control limits
• Rule-of-thumb: 20-30 samples, at least 5 observations in each.
• Basic idea:
Sample #: 1 2 ... 19 20
• x is normally distributed (Central Limit Theorem)
1 1.7 2.7 ... 2.1 1.2
• Most observations should be within 3 std. dev. from the mean 
Gemma Berenguer
Gemma Berenguer
2 1.7 2.3 … 2.7 3.1
control limits at these points
3 3.7 1.8 … 4.5 7.5
4 3.6 3 … 3.5 6.1
5 2.8 2.1 … 2.9 3
X bar (Sample mean) 2.7 2.38 … 3.14 4.18 FLIP
Range (Max - min) 2 1.2 … 2.4 6.3
• Now we calculate the grand mean and average range to find the center line:
-3 -2 -1 0 1 2 3
2.7 + 2.38 + ... + 3.14 + 4.18

x=
20 10 11
2 + 1.2 + ... + 2.4 + 6.3
R=
20
Number of
X-bar chart X-bar chart

observations in a
sample
n A2 n A2
2 1.88 2 1.88
• Control chart: 3 1.023 • Back to Call Center Example: 3 1.023
4 0.729 4 0.729
5 0.577 5 0.577
6 0.483 CL = 3.81 6 0.483
UCL = 3.81 + 0.577 * 5.85 = 7.18

7 0.419 7 0.419
Gemma Berenguer
Gemma Berenguer
8 0.373 8 0.373
9
10
0.337
0.308
LCL = 3.81 − 0.577 * 5.85 = 0.43 9
10
0.337
0.308
11 0.285 11 0.285
12 0.266 X-bar Chart 12 0.266
13 0.249 12.00
13 0.249
14 0.235 14 0.235
CL = x
10.00
15 0.223 15 0.223
Sample Mean
8.00
16 0.212 16 0.212
UCL = x + A2 R 17 0.203
6.00
17 0.203
4.00
18 0.196 18 0.196
LCL = x − A2 R 19 0.187
12
2.00 19 0.187
13
20 0.18 0.00 20 0.18
1 8 15 22
Samples
R Chart R-Chart
• Distribution of R is not normal and we use different constants: • Back to Call Center Example:
CL = 5.85
Number of observations in a sample
UCL = 2.114 * 5.85 = 12.37
Gemma Berenguer
Gemma Berenguer
n 2 3 4 5 6 7 8 9 10
LCL = 0 * 5.85 = 0
D3 0 0 0 0 0 0.076 0.136 0.184 0.223
R-chart
D4 3.268 2.574 2.282 2.114 2.004 1.924 1.864 1.816 1.777 30.00
25.00
Sample Range
20.00
CL = R 15.00
UCL = D4 R 10.00
5.00
LCL = D3 R 14 0.00
15
1 8 15 22
Samples
Why do we need both charts? Why do we need both charts?

Gemma Berenguer
Gemma Berenguer
16 17
How to obtain stable control Exercise
limits? Sixteen samples of size 9 each were collected.
Gemma Berenguer
Gemma Berenguer
(1) Find
(2)
18
M&M’s chocolate Example

Exercise 20 samples of size 5 each
m:
1 2 3 4 … 17 18 19 20
1 30 55 19 36 … 27 29 30 30
2 30 51 25 36 … 27 32 34 31
3 34 47 21 37 … 26 34 34 29
4 33 33 44 35 … 29 35 35 29
Gemma Berenguer
Gemma Berenguer
n: 5 32 48 28 38 … 27 34 36 29
Now we calculate sample mean for 20 samples.

X bar 31.8 … 27.2 32.8 33.8 29.6
Now we calculate the grand mean:
X = (31.8 + 46.8 + … + 33.8 + 29.6) / 20 = 32.17

M&M’s chocolate Example: X-bar chart M&M’s chocolate Example: X-bar chart
Step 1: Calculation of Control Limits Step 2: Plot charts
X bar chart: trial control limits
Sample # 1 2 3 4 5 6 7 8 9 10 Points 2, 8, 9, 12,
14 are outside
Mean 31.8 46.8 27.4 36.4 34.8 34.6 32.6 42.8 26.6 31.2
limits.
42.55
Range 4 22 25 3 5 9 15 6 5 6 Find assignable
37.36
Gemma Berenguer
Gemma Berenguer
Sample # 11 12 13 14 15 16 17 18 19 20 cause and
32.17 eliminate these
Mean 33.6 20.8 32.6 25.8 28.2 34 27.2 32.8 33.8 29.6 26.98 points---taking
these samples out
Range 13 2 12 9 5 22 3 6 6 2 21.79 from the data
16.60
X = 32.17 R = 9.00 1 3 5 7 9 11 13 15 17 19
Then find new
For n = 5, A2 = 0.577 CL = 32.17
control limits.
Each dot is the sample mean
LCL =32.17 - 0.577*9 = 26.98
UCL = 32.17 + 0.577*9 = 37.36
M&M’s chocolate Example: X-bar chart M&M’s chocolate Example: R-chart

Step 3: Re-calculate limits 20 samples of size 5 each
m:
Revised values with m = 15 samples 1 2 3 4 … 17 18 19 20
X = 32.04 R = 9.07 CL = 32.04 1 30 55 19 36 … 27 29 30 30
UCL = 32.04 + 0.577 * 9.07 = 37.27 2 30 51 25 36 … 27 32 34 31

3 34 47 21 37 … 26 34 34 29
LCL = 32.04 - 0.577 * 9.07 = 26.81
4 33 33 44 35 … 29 35 35 29
Gemma Berenguer
Gemma Berenguer
Revised X bar chart n: 5 32 48 28 38 … 27 34 36 29
37.27
All points are within limits. Next, we calculate sample range for 20 samples.
32.04
Stable control limits found Range 4 22 25 3 … 3 6 6 2
(34 – 30 = 4)
26.81
1 3 5 7 9 11 13 15 17 19
The average range:
R = (4 + 22 + … + 6 + 2) / 20 = 9.00
M&M’s chocolate Example: R chart
M&M’s chocolate Example: R chart – Step 3
Step 1: Calculation of Control Limits
Revision 1: m = 17
X = 32.17 R = 9.00 n = 5, D3 = 0.0, D4 = 2.115
CL = 9.00, UCL = 2.115*9 = 19.04 , LCL = 0 R = 6.53 CL = 6.53, UCL = 13.81, LCL = 0
R chart: trial control limits R chart: revised control limits
Each dot is the sample range value
Gemma Berenguer
Gemma Berenguer
19.04 13.81
9.00 6.53
0.0
-1.04 0.0
-0.75
1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19
Points 2, 3 and 16 are outside limits- Point 7 is outside limits.

Find assignable causes; eliminate points--taking out these samples
Find assignable cause, eliminate point. Perform revision 2.
Find new control limits.
M&M’s chocolate Example: R chart – Step 3 (cont.) M&M’s chocolate Example: R chart – Step 3 (cont.)
Revision 2, m = 16 Revision 3, m = 15 Revision 4, m = 14
R = 6.00 R = 5.53 R = 5.07 CL = 5.07, UCL = 10.73, LCL = 0
CL = 6.00, UCL = 12.69, LCL = 0 CL = 5.53, UCL = 11.7, LCL = 0 R chart: revised control limits (revision 4)
R chart: control limits (revision 2) R chart: control limits (revision 3)

10.73
Gemma Berenguer
12.69 11.7
5.07
6.00
5.53
0.0
0.0 -0.59
-0.69 0 1 3 5 7 9 11 13 15 17 19
0 5 10 15 20 25 -0.64
1 3 5 7 9 11 13 15 17 19
Point 11 is outside limits. Point 13 is outside limits. All points are in control now! Stable control limits found.
Perform revision 3 Perform revision 4
How to tell if the process is not in How to tell if the process is not
control? in control?
• Nonrandom • Nonrandom patterns
Patterns • Trend, Cycles, Bias, Mean shift, Too much dispersion
• Rules :
Gemma Berenguer
Gemma Berenguer
• Point beyond control limits (special cause)
• Eight or more consecutive points on one side of the line
• Six or more consecutive points increasing or decreasing
• Fourteen or more consecutive points oscillating
30 31
Control Chart for Attributes Why do we use control charts?

• p-Chart - Control chart used to monitor the proportion of • They are:
non-conforming items in a process (Use when the data can be • Proven,
converted into proportions). • Effective at defective prevention,
• Effective at preventing unnecessary correction…
Gemma Berenguer
Gemma Berenguer
• Example: proportion of patients admitted in less than 4 hours X-bar Chart
8.00
7.00
• c-Chart - Control chart used to monitor the number of non-

6.00
Sample Mean
5.00
conformances per unit (Use when the number of occurrences 4.00
3.00
per unit of measure can be counted; non-occurrences cannot 2.00
be counted). 1.00
0.00
1 8 15 22
Observation
• Examples: Number of service tickets closed in a day 32 33

• Is the process in control? YES
• What percentage of parts meets specs? NO IDEA
Toyota Production System Philosophy
• “The most important objective of the Toyota system has been

to increase production efficiency by consistently and
thoroughly eliminating waste. This concept and equally
Gemma Berenguer
Session 14 important respect for humanity are the foundation for the
Toyota Production System.”
Lean Operations
-Taiichi Ohno, Founder of TPS
Fall 2021
UC3M 2
Toyota Production System Toyota Production System

Lean Production Lean Production
• Part I: Just-in-Time Production
• Doing it all for the customer
• Leveled production
• Running a Pull system
Gemma Berenguer
• Continuous flow processing
Gemma Berenguer
• Takt Time
• Multi-skilled operators
• Part II: Jidoka

• Building quality into the Production Process
• Part III: Standardized work and Kaizen

• Standardized work – Basis for kaizen
3 4
• Kaizen: The Lifeblood of Standardized Production.
www.lean.org
The Seven Deadly Wastes The Seven Deadly Wastes
(TIMWOOD) (TIMWOOD)
1. Transportation: moving materials between workstations 4. Waiting time: workers who are idle because they have
without adding value. completed their work or who watch machines but cannot
prevent problems.
Gemma Berenguer
Gemma Berenguer
2. Inventory: the costs associated with loss, obsolescence, and
damage to inventories, as well as the cost of excess inventory 5. Overproduction: producing more than is required by the
itself. market.
3. Motion: movement that is not related to adding value to the 6. Over-processing: workers spend more on a flow unit than
product. necessary
7. Defects: the costs of scrap and rework and, most important of
all, the costs associated with defective output reaching the
5
customer. 6
Forms of waste
• Transportation
• Inventory
• Motion
Gemma Berenguer
Gemma Berenguer
• Waiting time
• Overproduction
• Over-processing
• Defects
7 8
Reducing waste: Increase Problem Visibility Lower
Forms of waste the Water to Expose the Rocks
• Transportation
• Inventory
• Motion
Too Much
Gemma Berenguer
Gemma Berenguer
• Waiting time Missed Due Dates Space
• Overproduction Late
• Over-processing Deliveries
Inventory
Too much paperwork
• Defects Scrap & Engineering
Rework Change Orders
Poor 100% inspection Long queues

Machine
Quality
Downtime
9 10
Expose Problems
Remove
Lower Inventory  Increase Problem Visibility
the
Lower the Water to Expose the Rocks
obstacle
Production
High W.I.P. WIP reduced
Poor Poor
quality quality Poor
Long Poor Long
Bad 11 Bad maintenance 12
lead- maintenance lead-
material material
time time
Remove
the
obstacle
again!
WIP reduced further
Poor Poor
quality Poor quality
Poor
Long maintenance 13 maintenance 14
lead-
time
Reducing Waste: Cut Batch How to run Lean Operations:

Sizes Managing Variety
Example Process: A B C D
1 min/job 1 min/job 1 min/job 1 min/job
• Monthly Production Requirement:
Batch Shop (Lot size = 5) Flow Shop (Lot size = 1)
Space Space Model Quantity Required
Gemma Berenguer
Gemma Berenguer
0 5 0 1
1 1 1 A: Sedan 1,000
2 2 1
3 3 1 B: Coupe 1,000
4 4 1
5 5 5
A B C D A B C D
10 5 10
• How to schedule production for the month?
15 5 15 15
Time
20 5 20
Reducing Waste: Heijunka
Forms of waste
Leveled/Balanced Production
Unleveled Production Schedule Leveled Production Schedule • Transportation
(AAAAABBBBB..) (ABABAB...)
• Inventory
A A
• Motion
Gemma Berenguer
B B • Waiting time
• Overproduction
Inv Inv
• Over-processing
• Defects
Time Time
17 18
Tradeoff? Switchover costs vs Inventory costs.
Similar tradeoff to EOQ…
Overproduction: synchronization with Implementation of pull system:

demand Kanban
PUSH: Inputs availability triggers execution
• Card - typically carried in transparent covers.
• Various systems using one or more Kanban cards
Supplier Process Customer
inputs outputs • Example using two cards (come in matching sets):
Gemma Berenguer
Gemma Berenguer
• Move (Withdrawal ) Kanban (card)
PULL: Outputs need triggers execution • Production Kanban.
Customer demand pulls product
Move Kanban Production Kanban
Item no. : P447 Item no. : P447
Box Capacity : Process:
Supplier Process Customer
inputs outputs Preceding process: Xxx:
Subsequent process: Yyy:
Taiichi Ohno (1912-1990): idea based on US grocery stores 19 20
• Other systems: Kanban squares, Container system,
Colored golf balls.
Kanban card Video Kanban
• Kanban System example:
• https://www.youtube.com/watch?v=5hoTWwtYEQo
Gemma Berenguer
Gemma Berenguer
21 22
Forms of waste
• Transportation
• Inventory
• Motion
Gemma Berenguer
Gemma Berenguer
• Waiting time
• Overproduction
• Over-processing
• Defects
23 24
Reducing Waste Pays Off:
Forms of waste
Quality at the Source
• Transportation Defects Own Next End Final End User’s
Process Process of Line Inspection Hand
• Inventory Found at:
• Motion
Impact on the
Gemma Berenguer
Gemma Berenguer
• Waiting time company:
• Overproduction
Very Minor Rework Significant Warranty costs
• Over-processing Minor delay rework Administrative
Reschedule
• Defects of work
Delay in
delivery
costs
Reputation loss
Additional Loss of market
inspection share
25
Poka-Yoke (Fool-proof Design) JIT Fundamental principals

2. Reduce Inventory.
1. Eliminate Waste.
2.1 Expose hidden problems.
1.1 Over production.
3. Meet exact needs
Gemma Berenguer
1.2 Excess inventory.
1.3 Products waiting in at exact time.
queues.
3.1 Small lot size.
1.4 Unnecessary moves (poor
layout, operation 3.2 Pull system.
sequence).
3.2.1 High degree of planning.
1.5 Large set-up times.
3.2.2 Flexibility.
1.6 Waiting (due to shortage /
3.2.3 Level production.
scheduling).
27 3.2.4 Minimum difference in
1.7 Expediting.
products
28
Jidoka (autonomation)
• Andon cord
Standardize
Gemma Berenguer
Improved
Box A Box B design Box B
Using common handles on both boxes, you can achieve overall

savings, although cost of handle may be higher on for box B.
29 30
Continuous Improvement (Kaizen)

Root Cause Analysis - 5 Whys Self-Learning process
• “Many flat head screws break during the assembly process”; • “Produce a part and learn how to produce better next time”
WHY? • “Serve a client AND learn how to serve better next time”
Gemma Berenguer
Gemma Berenguer
• “Many screws rust before we use them on the assembly line”;
WHY? • “Teach a class and learn how to teach better next time”
• “Screws get rained and weathered”;
WHY?
• “Screws are stored outside”;
WHY?
• “Because all our indoor racks are full”;
WHY? 31 32
• “Because we have too much inventory.”
Summarize TPS Comparison of systems
• TPS is a production management system that aims for the “ideal” Factor Traditional JIT
through continuous improvement (kaizen);
Inventory Much to offset forecast Minimal necessary to
• Includes, but goes way beyond JIT;
errors, late deliveries operate
• Pillars:
Gemma Berenguer
Deliveries Few, large Many, small
• Synchronization: Heijunka
• Quality at Source: Jidoka, Poka-Yoke Lot sizes Large Small
• Continuous Improvement: Kaizen, Waste reduction
Setup; runs Few, long runs Many, short runs
Vendors Long-term relationships Partners
are unusual
Workers Necessary to do the Assets
work
....
Video: Noodles and Co. When does JIT not work?

• Requires an implementation throughout the supply chain. A
• This is an excellent example on how lean production is used in
unilateral implementation by a single vendor/manufacturer will not
a service company. work;
• Pay attention to how the company designs their process to • JIT requires a cultural change within an organization
Gemma Berenguer
Gemma Berenguer
offer fast and good quality service/food: Workers need to accept responsibility for making process changes,
managers should be willing to cede control;
https://mediaspace.itap.purdue.edu/media/MGMT460_Noodles%26Co_100824/1_jbp4v
8wx • Kanbans are reactive. Hence may not work when every job is unique,
requiring different components and some degree of customization;
• JIT is also based on aggregate forecasts for planning purposes. The

system may not work when end product demand is highly variable
35
Case discussion: TMM in 92 Seat Problems – Symptoms
• Production started in July 88 • Defective seat
• Producing Camrys
• 4000 people, $150 million in annual payroll • Growing number of cars sitting off the line
• Assembly plant: sheet-metal stamping, plastic molding, body • Line stoppage
Gemma Berenguer
Gemma Berenguer
welding, painting, assembly
• Assembly operations: • Line is stopped 40 times per shift
• 353 stations on conveyor line • Reduction in the run ratio from 95% to 85%
• 5 miles in length
• Trim lines, chassis lines and final assembly lines
• Cars found in the parking areas
• Several connected line segments
• 2 shifts with 7.5 hours of productive time each
Seat Problem –Description Seat Problem – Description

• Too many cars arrive at the overflow parking area with seat defects. • The causes and the sources of the seat defects are not clear.
• Also cars leave the overflow parking area too slowly.
• The specific seat quality defects (material flaws, missing parts,
• Failures in communication between seat bolster hook broken).
Gemma Berenguer
Gemma Berenguer
• TMM and KFS
• Actual defects are not clearly specified (handwritten, not
• Between groups in TMM
• Vertically in TMM specific).
• Seat defective cars are allowed to accumulate (only limited by
• Failure to act on the seat problem until it became critical
physical space) while the problem goes unsolved. In fact, this masks
• Group leaders should have detected and acted on this problem
sooner. the problem.
• Lack of mechanism to force managers to confront problems as soon

as they arise. Most important problems because they violate clear TPS principles!
Actions to be Taken
People Roles
• What are their responsibilities?
(Examples)
1. Jim Cremeens to fix the seat reorder form.
• Mike DaPrile – general manager of the assembly plant
2. Jim Cremeens to redesign the overflow area to highlight seat problems
• Doug Friesen – manager of assembly for Toyota’s Georgetown plant (8 cars in early (like andon pull)
Gemma Berenguer
3. KFS puts QC engineers at TMM
the clinic area)
4. Roger Lewis works with KFS to help discover quality problems in their
• Kevin Smith – manager of purchasing processes and inspection. Kevin smith should facilitate communication in
• Jim Cremeens – group leader of clinic and overflow area between.

5. Shirley Sargent to ensure that attention is drawn to seat problems on the
• Roger Lewis – assistant general manager of QC line.
• Shirley Sargent – group leader of Final 2 (who proposed engineering change of 6. Roger Lewis to investigate root causes. Roger Lewis and Shirley Sargent to
investigate whether the hook problem requires an engineering design
the hook)
change or better training on the assembly line.
Overflow Area – Jim Cremeens Assembly Line – Doug Friesen

• Should we eliminate overflow area?
• The bolster hook problem
• Streamline the order replacement process (why wrong replacement parts are
being sent?) • How significant is this problem?
• Better communication between TMM and KFS (a handwritten form that is faxed • Is this an assembly problem or an engineering design change problem?
– increases potential illegibility difficulties).
Gemma Berenguer
Gemma Berenguer
(Shirley Sargent)
• Redesign the form
• Computerize and training – process design (the right person gets the right • Stop the line when seat problems arise
information), equipment and software
• Correct the problem immediately by
• Organize this area so the repeat problems become evident quickly
• Designate an area for seat problems (learn quickly about the problem, limit • having backup seats available (is this a good idea?)
batch size) • stopping the line until a replacement seats arrive (probably a bad idea –
• Supervising team members more closely
this could take a long time)
• Draw attention to the problem and send to the overflow area for offline
repair (what are the pros and cons with this idea?)
KFS – Kevin Smith and Roger Lewis TMM and TPS
• Most of the defects are from KFS • Defects in the application of TPS
• They do 100% inspections so there is definitely something wrong here.
• Seat problem has gone unrecognized for so long.
• Review the seat transport system between KFS and TMM (wrong seats have
Gemma Berenguer
Gemma Berenguer
been sent). • “Batching” in the overflow area covers the problem.
• Ask KFS engineer to TMM • Failure to go after the root causes.

• Solve the seat problem on-site!
• Actions
• Send expert to KFS
• Change enablement – culture
• Examine quality issue in production
• Examine routing problem for reorder

• Train and motivate workforce
• Review them and perhaps their suppliers. • Examine management process!
Outline of the course

• Part 1 (Quiz 1)
• Forecasting
Gemma Berenguer
Session 15
• Part 2 (Quiz 2)
• Part 3 (Quiz 3)
Supply Chain Management 1 •
•
Lean Operations
• SCM
Fall 2021
Department of Business Administration • Part 4 (Final project)
• Circular Economy/Sustainability
UC3M 2
What is a supply chain? What is a supply chain?
• https://www.youtube.com/watch?v=Mi1QBxVjZAw
• Definition:
comprises all the parties involved in providing a product or service to a
customer, either directly or indirectly,
Gemma Berenguer
Gemma Berenguer
• What is supply chain management?
• a set of approaches utilized to efficiently integrate suppliers,
manufacturers, warehouses, and stores, so that merchandise is
produced and distributed
• at the right quantities,
• How much of a bottle of water is profit? • to the right locations, and
• at the right time,
• in order to minimize system-wide costs while satisfying service level
• What do hotels manufacture? requirements
3 4
Supply chain (slidefrom:UCBerkeleyprofessorMaxShen) Supply chain flows (1)
Information
Gemma Berenguer
Product
Firm Customer
Funds
5
Supply chain flows (2) Supply chain flows (2)
Upstream Downstream Upstream Downstream
Returns, repairs, raw materials,

Material Material servicing, recycling, intermediate products,
Gemma Berenguer
Gemma Berenguer
disposal finished goods
Sales, orders, capacity, promotion
Information inventory, quality, plans, delivery
Information
promotion plans schedules
Payments, Credits, consignment,
Finance consignments payment terms, invoice
Finance
Hau L. Lee, “Creating value through supply chain integration”, Supply Chain Management 7 Hau L. Lee, “Creating value through supply chain integration”, Supply Chain Management 8
Review, 2000. Review, 2000.
Decision phases of a supply

SCM decisions
chain
• Supply chain strategy or design
• What suppliers should we use and how many?
• How to structure the supply chain over the next
• How many factories and warehouses should we have?
several years
• What products should each factory produce?
Gemma Berenguer
Gemma Berenguer
• In what locations should our firm have factories and warehouses? • Supply chain planning
• How do we set capacity at each location? • Decisions over the next quarter or year
• Given locations and capacities, what quantities should we produce and
store at these locations? • Supply chain operation
• What quantities should move from location to location and at what time? • Daily or weekly operational decisions
9 10
Supply chain Strategy phase Supply chain Planning phase
• Decisions about the structure of the supply chain and what
processes each stage will perform
• Definition of a set of policies that govern short-term operations.
Fixed by the supply configuration from previous phase
• Strategic supply chain decisions
• Starts with a forecast of demand in the coming year
Gemma Berenguer
Gemma Berenguer
• Locations and capacities of facilities
• Planning decisions:
• Products to be made or stored at various locations
• Planned buildup of inventories
• Information systems type
• Subcontracting, backup locations
• Supply chain design must support strategic objectives • Inventory policies
• Supply chain design decisions are long-term and expensive to • Timing and size of market promotions
reverse – must take into account market uncertainty • Must consider in planning decisions demand uncertainty,
exchange rates, competition over the time horizon
11 12
Supply chain Operations

Examples of best practices
phase
• Time horizon is weekly or daily
• Decisions regarding individual customer orders
• Supply chain configuration is fixed and operating policies are • DELL
Gemma Berenguer
Gemma Berenguer
determined
• Goal is to implement the operating policies as effectively as
possible
• Operational decisions
• set order due dates • Zara
• generate pick lists at a warehouse
• allocate an order to a particular shipment
• set delivery schedules
• Much less uncertainty (short time horizon)
13 14
DELL DELL Global Supply Chain
• Direct business model

• No middlemen
• No warehouses
Gemma Berenguer
• Suppliers are close by Canada,
Video cards
U.S., CPUs,
France,
Windows,
Soundcards Korea, Japan
• Build-to-order strategy
China, Cases wt Power
Supply, Cables, RAM Chips,
Network Cards, Monitors
• Production doesn’t begin until order is placed Dell Assembly
Austin, Texas
Taiwan,
Mexico, Motherboard
• There is never more than 4 hours inventory on hand Keyboards, Mouse Thailand,
Hard Drives Malaysia,
• This leads to lower cost

CD-ROMs
15
Push/Pull view of a supply Push/Pull view of supply chain

chain process
• Supply chain processes fall into one of two
categories depending on the timing of their
Gemma Berenguer
execution relative to customer demand
• Pull: execution is initiated in response to a customer
order (reactive)
• Push: execution is initiated in anticipation of
customer orders (speculative)
• Push/pull boundary separates push processes from
pull processes
17 18
DELL customer driven supply
DELL’s supply chain
chain segmentation
• Before 2008
• Focus on responsiveness Customer Type Strategy Push or Pull?
• In 2008, Dell enters retail channel Online sales Build-to-order PULL
• The supply chain setup does not support low cost (online
Gemma Berenguer
Gemma Berenguer
customers are less price-sensitive than retail customers) Popular online Build-to-stock PUSH-PULL Procurement, production
sales and shipping to stocking points are all
• Need to re-engineer based on forecast; shipment to customer
• Segment supply chain using: locations is based on realized demand
Retail channels Build-to-Plan PUSH Strategy, procurement,
• Demand uncertainty production and shipment decisions are all
• Cost drivers based on forecast
• Relationships with customers Enterprise clients Build-to-Spec PUSH-PULL Products are assembled to
• Customer value proposition order using components ordered well
ahead of time, based on forecast
• Technology clock-speed 19 20
“When One Size Does Not Fit All”, MIT Sloan Management Review, 2012, Simchi-Levi, Clayton
“When One Size Does Not Fit All”, MIT Sloan Management Review, 2012, Simchi-Levi, Clayton
and Raven
and Raven
DELL’s four supply chains Cost-Responsiveness Efficient Frontier

Responsiveness
High DELL’s
Gemma Berenguer
position before
2008
Low
21
Cost
“When One Size Does Not Fit All”, MIT Sloan Management Review, 2012, Simchi-Levi, Clayton High Low
and Raven
22
Zara “Fast Fashion” supply
What is behind Zara’s success?
chain
• VIDEO: • Stockout often
• https://www.youtube.com/watch?v=qhCM0F81vEg • Produce in the most expensive part of the world
• Redesigns apparel all time, which lead to demand
Gemma Berenguer
• Answer the following questions while watching the video: uncertainty
• How often the assortment is modified in stores? • Does not advertise
• No product innovation
• How long are clothes warehoused? • No technology innovation
• Not a new niche
• Zara innovates through better supply chain
management!
23 24
What is beyond Zara’s success? Achieving Strategic Fit

Responsive
• It transferred Dell’s build to order model to the apparel industry. supply chain
It produces fashion to order
• Close production location (Spain and Mexico), short lead
times
Gemma Berenguer
Gemma Berenguer
• Quantity decisions made days before sales
Responsiveness
• Detailed product differentiation decision is postponed Extra reading:
spectrum
(design team develops platform models but holds off What is the right
finalizing detailed designs) supply chain for
• Few days of inventory your product?
• 21-29 day cycle (from design to sales/discount) (6 months

for others)
Efficient
• It pays with more expensive capacity, but the trade-off is supply chain
favorable 25
Certain Implied Uncertain
demand uncertainty demand
spectrum
26
Matching supply chain with Matching supply chain with
products products
Functional Products Innovative Products Functional Products Innovative Products
Gemma Berenguer
Gemma Berenguer
Efficient Efficient
Match Mismatch
supply chain supply chain
Responsive Responsive
Mismatch Match
supply chain supply chain
Extra reading: What is the right supply chain for your product? Extra reading: What is the right supply chain for your product?
27 28
Facilities Inventory
Efficient Responsive
Minimize inventory Inventory close to the customer
Small number of large facilities Larger number of smaller facilities
Gemma Berenguer
Gemma Berenguer
Large inventory
Operate with high utilization Flexible facilities
High variety
Cheap location Location close to markets
Cost of under-stocking is high
Assembly line Job shop
Use of automation
29 30
Transportation/Shipping Sourcing
Efficient Responsive Efficient Responsive

Focus on cost Flexible suppliers
Gemma Berenguer
Transportation cost Transportation need (expediting,
(truck/rail/grounds) air, express carriers)
Single sourcing Dual sourcing (multiple)
Full truckloads Less than truckload (LTL)
Outsource In-house
Small number of large shipment Large number of small shipments
Purchase in bulk Small shipments
31 32
Efficient/responsive supply
Pricing
chains
Efficient supply chain Responsive supply chain
Supply demand at the lowest Respond quickly to demand

Primary goal
cost
Maximize performance at a Create modularity to allow
Gemma Berenguer
Low High Product design

minimum product cost postponement of product differentiation
strategy
EDLP Loyal customer Lower margins because price is Higher margins because price is not a
Pricing strategy a prime customer driver prime customer driver
Manufacturing Lower costs through high Maintain capacity flexibility to buffer

strategy utilization against demand/supply uncertainty
Inventory Minimize inventory to lower cost Maintain buffer inventory to deal with
strategy demand/supply uncertainty
Lead-time Reduce, but not at the expense Reduce aggressively, even if the costs
33 strategy of costs are significant 34
Select based on cost and quality Select based on speed, flexibility,
Supplier strategy
reliability, and quality
Hierarchy of Network Design
Decisions
Network Design
Strategy
Gemma Berenguer
Session 15 Regional Facility
Configuration
Supply Chain Management 2

Selecting Desirable
Sites
Fall 2021
UC3M Location Choices 2
Framework for Network Design

Stage I: Network Design Strategy
Decisions
Goal: identify the broad structure of the supply chain network
based on business strategy, customer segments served, and
product characteristics
- in-house or outsourced
Gemma Berenguer
Gemma Berenguer
- how many layers/tiers?
- What products? What markets?
…
• What factors influence network design decisions?
• What are alternative structures for a distribution network?
How do they compare in terms of supply chain cost and
3
capabilities? Which customers and products are they suited 4
for?
Relationship between Service and Number of
Costs as a Function of Number of Facilities
Facilities
Inventory
High
More facilities allow for a
quicker response time
Gemma Berenguer
Gemma Berenguer
Costs Facility costs
Response Time
Transportation
Low
Number of Facilities 5 Number of facilities 6

(Inverse of distance from customer)
Stages III and IV: Facility Location, Capacity

Stage II: Facility Configuration and Demand Allocation
• Which regions should have facilities?
• Key Factors: • Typical problem when rationalizing supply chain network after a
• Regional demand configuration merger
Gemma Berenguer
Gemma Berenguer
• Taxes and tariffs
• Fixed facility cost (economies of scale) • Which plants to keep open?
• Transportation cost
• What capacity to assign to each plant?
• Production cost
• Inventory cost • Which markets to assign to each plant?
• Coordination cost
• Figure out demand per region, capacity of each facility,…

7 8
Multi-Stage Location & Allocation Decisions In-Session Example
• Consider Home Store, a major retailer of home improvement
Plants (i) Markets (j)
goods. They would like to build new distribution centers (DCs) to
1 1 serve Southeastern United States, namely the states Florida,
Georgia, North Carolina, South Carolina and Alabama. Candidate
Gemma Berenguer
Gemma Berenguer
2 2 sites are in Jacksonville, FL, Atlanta, GA, Columbia, SC or
Birmingham, AL.
• Where and how many DCs should Home Store build?
• Which DC should serve which market?
n
m
Which plants to establish? Which plants distribute to which 9 10
markets? How to configure the network?
In-Class Example What type of data is needed?

• Parameters: Costs, capacities, demands
• Demand and capacity data go into the constraints:
Demand regions Annualized • Locations of customers and annual demand for each product at
Transportation, handling, and inventory fixed each location
related costs per 1000 ton cost($)
• Capacities of potential sites
Gemma Berenguer
Gemma Berenguer
DC NC SC GA FL AL
• Variable costs go into the objective function
Jacksonville 95 90 89 88 93 700
• Transportation costs between potential sites and customer
Atlanta 90 85 80 85 88 800 locations
Columbia 88 82 88 100 98 650
• Facility, labor and material costs (i.e. production, sourcing,
Birmingham 95 93 90 105 78 630 handling costs) by site
Demand 13 10 19 21 15 • Inventory costs by site
• Sale price of each product at each location (if the objective is
Each DC can handle up to 30,000 tons of flow. profit maximization)
Costs are given in terms of 1000 $s. 11 12
Demands are given in terms of 1000 tons.
What type of data is needed?
(Cont.)
Quantitative Approach
• Fixed costs go into the objective function • Home Store’s objective is to determine
• Fixed overhead (independent of volume) at each potential site • the location and number of DCs
(e.g. some utilities, management personnel, depreciation, etc.) • which DC will serve which market
• Others: Can go in either the obj. function or constraints
Gemma Berenguer
Gemma Berenguer
at minimum total cost
• Taxes and tariffs • One approach: enumeration. But it is time consuming!!
• Desired response time and other service constraints
• Excel and Solver.
Note: The total cost we are concerned about is also known as total landed cost.
Total landed cost is the total cost of making the products and delivering them to
the customers.
13 14
Optimal solution Software Vendors

Decision Variables
Demand regions Transportation, handling,
DC NC
and inventory related costs per 1000 ton
SC GA FL AL
DC
(open =1)
• BlueYonder
Jacksonville
Atlanta
0
0
0
0
0
9 21
0 0
0
0
1
• SAP
Gemma Berenguer
Gemma Berenguer
Columbia
Birmingham
13
0
10
0
7
3
0
0
0
15
1
1
• IBM-ILOG
• Infor
Constraints
DC Excess capacity
Jacksonville 0
Atlanta 0
Columbia -1.776E-15
Birmingham 12
NC SC GA FL AL
Unmet demand 0 0 0 0 0
15 16
Objective function
8605
Capacity Location Capacity Location:
Globalization and operations Four types of analyses
• Geographical or spatial analysis and total landed cost (TLC)
Market Demand Factors • Foreign demand growth • Quantitative models
• Foreign competition
Gemma Berenguer
Gemma Berenguer
• Competitive threat/priorities
• State-of-the-art markets
• Network analysis
Globalization • Timing and location decisions interact
Technological Factors Supply Factors
Process • Need to take a dynamic rather than myopic view on location
• Access to knowledge & skills • Direct labor v. total
costs decisions
• Infrastructure
• Capital investments
• IT: Network planning and coordination & economies of scale
Macroeconomic and Non-Market • Taxes and incentives • Competitive analysis
Factors
• Tariffs, quotas and other protectionism
• Trade and global institutional agreements
• Exchange rates • Facility role or strategic focus analysis
• Political stability
Kouvelis and Niederhoff (2007)
17 18
Capacity Location: Network analysis Capacity Location: Competitive Analysis

Myopic vs. Dynamic View on Location Decisions • Competitive considerations in location decisions:
• Hotelling Model
• Competing firms tend to locate at a distance from each other to split the
• market (e.g., airline hubs, large retailers, etc.)
Gemma Berenguer
Gemma Berenguer
B
A
C
• Positive externality between firms

• Competing firms sometimes concentrate in certain regions to take advantage
20
• of easy access to skilled labor, technology, or customers (e.g., high-tech firms,
http://nationalatlas.gov/ Source: Courtesy of Jan A. Van Mieghem
financial service firms, pharmaceutical firms, etc.)
19
Capacity Location: Strategic analysis
Strategic role of a foreign location
General Rationale for Offshoring
Broad & high
What is the scope of its current
Source plant Lead plant Contributor plant Pro Con

activities and competencies?
• Produces at low cost but with same • Global hub for product and process • Serves specific regional market but
ability as best plant in the network knowledge and innovation. competes with other plants in network
• Authority over procurement, production • Full authority over all activities. Taps for new processes or products.
• Cost reductions • Transportation costs
Gemma Berenguer
planning, logistics, process & product into local skills and knowledge to • Authority over product and process
initiate company - wide innovation. development as well as supplier
customization
choice. • Proximity to local or new markets • Lead times
Offshore plant Outpost plant Server plant • Domestic labor market • Risks
constraints • Quality, including health,
• Produces specific items at low cost • Primary role is to collect information • Serves specific regional market to
and exports for further work or for from advanced suppliers, competitors, overcome tariffs, taxes, logistics, or • Operational hedge environmental and CSR
sale. research labs, or customers. foreign- exchange risk.
• Minimal authority and investment.
• Secondary role as offshore or
• Limited authority to make minor • Foreign trade barriers • Currency, IP, political,
modifications to fit local conditions.
server. competitive
Narrow & low Access to low -cost production Access to skills and knowledge Access to market • Domestic trade barriers
• Global operations’ complexity
and social implications
What is the primary strategic reason for the location?
Adapted from Ferdows (1997) “Making the most of foreign factories. HBR, Mar-Apr, 73-88.
21
Key issues in procurement 1. How many suppliers?

• Advantages of having a large supplier base
1. How many suppliers for each product category? • More buying power
• Broadens the range of choices (prices, etc.)
• Risk diversification
Gemma Berenguer
Gemma Berenguer
2. How to select a supplier based on multiple attributes
• Examples (2002 data)

3. Purchasing function in-house or outsourced?
• Ford had 40,000 suppliers
• GM had 30,000 suppliers
We’ll see how by designing contracts we can improve
relationships with suppliers • IBM had 15,000 suppliers
• Lockheed Martin had 37,000 suppliers
23 24
Gemma Berenguer/MGMT 664 Gemma Berenguer/MGMT 664

1. How many suppliers? (cont.) 1. How many suppliers? (cont.)
• Single-sourcing
• Advantages of having a small supplier base
Contracts:
• Less transaction costs - Buyback contract (allows to return unsold
inventory at an agreed-upon price)
Gemma Berenguer
Gemma Berenguer
• More non-contractible investments such as quality, information
sharing, design help and innovation - Revenue-sharing contract (low wholesale price
and sharing fraction of retailer’s revenue)
- Quantity-flexibility (quantity ordered can be

changed)
25 26
Gemma Berenguer/MGMT 664 Gemma Berenguer/MGMT 664
2. Supplier selection 2. Supplier selection (cont.)

• Attributes Method: Multi attribute simple
• Product quality weighting
• Lead time
• Delivery reliability Criteria Weight Supplier 1 Supplier 2
Gemma Berenguer
• Product costs Cost 9 8 4
• Cost transparency Quality 3 9 2
• Technical capabilities of the supplier Weighted score 99 42

• Financial strength of the supplier The
• Location of the supplier winner!!
• Development capabilities More sophisticated method: Analytical Hierarchy Process
… (AHP)
27 28
Gemma Berenguer/MGMT 664

Humanitarian supply chains
Development Emergency
Gemma Berenguer
Session 16
Operations Operations
Humanitarian operations
Fall 2021
UC3M 2
Types of disasters Are disasters increasing?

Gemma Berenguer
Gemma Berenguer
Geophysical disasters have remained fairly stable since
1970s, while the number of climate related disasters
3 4
have greatly increased
Source: http://www.accuweather.com/en/weather-blogs/climatechange/steady-increase-in-
climate-rel/19974069
Disaster timeline Supply Sources
Gemma Berenguer
Gemma Berenguer
Source: Pinar Keskinocak: School of Industrial and Systems Engineering, Georgia
Tech Source: Pinar Keskinocak: School of Industrial and Systems Engineering, Georgia
Tech
Inter-organizational
Players/Stakeholders
collaboration
Gemma Berenguer
Gemma Berenguer
Source: Pinar Keskinocak: School of Industrial and Systems Engineering, Georgia Source: Pinar Keskinocak: School of Industrial and Systems Engineering, Georgia
Tech Tech
an NGO example UN Humanitarian Response Depots
• https://www.youtube.com/watch?v=XrnoSc5NyyQ&list=PL527
47A584C38DEB9
• The World Food Program (WFP)
• Answer these questions while watching the video:
• How many tons of aid WFP delivers per year?
• How success is measured by WFP?
9 10
http://www.wfp.org/logistics/humanitarian-response-depot
Gemma Berenguer/MGMT 664
Current state of SCM in most

Compare with SCM in Walmart
NGOs
Source: Pinar Keskinocak: School of Industrial and Systems Engineering, Georgia Source: Pinar Keskinocak: School of Industrial and Systems Engineering, Georgia
Tech Tech
Outline of the course
• Part 1 (Quiz 1)
• Forecasting
Gemma Berenguer
Session 17
• Part 2 (Quiz 2)
• Part 3 (Quiz 3)
Circular Economy •
•
Lean Operations
• SCM
Fall 2021
Department of Business Administration • Part 4 (Final project)
• Circular Economy/Sustainability 2
UC3M
Today Inkjet Printer
• HP case discussion
• Some environmental legislations
• Sustainable Operations
3 4
Returned products process 2 basic issues
• Acquisition of used products 1.How to reduce returns

• Logistics process
• Disposition and remanufacturing process 2.How to leverage value from unavoidable returns.
• Remarketing and reselling
5 6
The product returns problem in a global The product returns problem in a global
context context
• Have you tried to return products? • Returns are likely to _________ in the future
• globalization
• Pressures for companies to manage all product returns in a
responsible manner • increasing likelihood of environmental legislation
• increment of online sales
7 8
Why do customers return printers? What can
we do to reduce returns? All-in-one printer-copier-fax
What is the % of
defective products?
9 10
Problem 1:What can we do to

What can we do to reduce returns?
reduce returns?
• Key points
• Product designers could
• improve compatibility with other products • An important part of returned products are in perfect condition
• facilitate ease of installation • Lack of information and incentive alignment between all
functions of the company
• Packaging designers could improve product information and
avoid misleading the customer
• Sales people could make an extra effort to provide the • product returns should be considered as an integrated business process (C-level
customer with the right information and thereby avoid wrong attention)
expectations
• Modify the return policy for resellers
11 12
Problem 2: How to leverage value from
Consumer and channel returns unavoidable returns?
• Do you think consumer returns can be eliminated completely? Closed-loop supply chain
• What about returns at the reseller level?
13 Source: “Reverse Supply Chain for Commercial Returns”, , J. D. Blackburn, V. D. R. Guide, G. C. Souza, L. N. Van Wassenhove 14
HP’s Return process Marginal value of time

(Figure 1)
PCs devaluate
by 1% per week!
15 Source: “Reverse Supply Chain for Commercial Returns”, , J. D. Blackburn, V. D. R. Guide, G. C. Souza, L. N. Van Wassenhove 16
Efficient versus responsive
chain
• What is the overall loss of product value?
• We can assume that the lead time (time between product is

returned to went it is resold): 4 months
• 4 months (16 weeks) x 1% value loss per week = 16%
17 18
Does it make sense to engage in

Returns lifecycle remanufacturing activities?
• Or leave it to a third party to come up with a streamlined, fast
process to recover value from your products
Pros of outsourcing Cons of outsourcing

Faster process
Cheaper
Remanufacturing reveals a lot
about how the product is
done
third parties might do a poor
job
Source: “Time Value of Commercial Product Returns” Gilvan C. Souza, V. Daniel R. Guide Jr., Luk N. Van Wassenhove, Joseph D. Blackburn 19 20
What can be done to maximize
New market creation
value recovery?
• How would you measure cannibalization?
• Improve velocity of reverse flow by
• express shipments • When would cannibalization be likely and when not?
• faster processing at the return depot and T&R facility
• … • Markets for remanufactured goods for HP
• Outsource part of the process
• Standarization of product design
21 22
Remanufactured stock has a

Closed-loop supply chains
unique business problem
• Attempts to integrate the forward flow and the different
return flows into a single service process to the customer
• There is no control over the rate at which the returns flow,
because this rate depends on the rate of returned products
- product acquisition
- reverse logistics
• We cannot buy as much as we want of returned stock as we - testing and sorting
- refurbishing
do with raw material ☺ - redistribution
- remarketing
23 24
Decentralized returns supply Centralized returns supply
chain chain
25 26
Source: “Reverse Supply Chain for Commercial Returns”, , J. D. Blackburn, V. D. R. Guide, G. C. Souza, L. N. Van Wassenhove Source: “Reverse Supply Chain for Commercial Returns”, , J. D. Blackburn, V. D. R. Guide, G. C. Souza, L. N. Van Wassenhove
Conclusions of the case Conclusions (cont.)

• The product returns problem in a global context • Product returns should no longer be considered as a burden, a
• Why do customers return printers? What can we do to reduce nuisance, something unavoidable, the only goal of which
returns? should be to minimize cost impact.
• The time value of product returns
• What can we do with the returned items? • Should instead be to support the business and to recover
• New market creation maximum value from returned products (i.e. business model
on product returns)
• Closed-loop supply chains
27 28
Take-Back Legislation in
Legislation Electronics
• Who should be responsible of the handling of waste?
• Take-Back Legislation in Electronics • In Europe (WEEE Directive 2003)
• Legislation of Greenhouse gas emissions

• EU ETS Category Recycling target Recovery target
• USA Large household 80% 85%
appliances
IT & 70% 80%
telecommunications
Electric and electronic 55% 75%
tools, medical
equipment,etc.
Lamps 85% 85%
29 30
Take-Back Legislation in
Electronics Take-Back Legislation (cont.)
• In Japan (SHAR)
• Characteristics of the product
• In US
• California is the only state where the consumers pay an ARF
(advanced recycling fee) • short life-cycle
• expensive materials
Dimension New York Washington Indiana
• expensive repairing
Collection and • State-wide goal: By convenience: 2011: 60% amount
recycling targets 3lbs/person in manufacturer must sold in previous year
2011 and have collection sites After 2011: 80% of
5lbs/person in in every county plus amount sold
2013 every city of 10,000 or
• Producers assign more
their own goals
based on market
share
Who gets free All except large Consumers, charities, Consumers, public
recycling businesses (>50 schools, small schools, small
employees) businesses businesses (<100 31 32
employees)
Legislation of Greenhouse gas Command and control versus
emissions market-based instruments
• Cap-and-trade legislation
• for CO2 emissions in Europe and in California (2011)
• SO2 emissions in the US (acid rain)
• Other legislation can be established besides cap-and-trade…

• emissions tax
• command-and-control legislation
• (certificates: ISO, EMS)
33
Source: http://www.treasury.govt.nz/publications/research-policy/wp/2003/03-02/09.htm
Sustainable Development Goals GRI standards and SDGs

• Website: https://sdgs.un.org/goals • SDGs and business:
• https://www.youtube.com/watch?v=Rht4d4GdYbM
• GRI standards
• Website: https://www.globalreporting.org/standards/
• Video: https://www.youtube.com/watch?v=-1K7z96i5gs
35 36
Corporate Social Responsibility Sustainability
▶ Meeting the needs
of the present without
▶ How products and services affect people and the compromising the
environment ability of future
▶ Stakeholders have strong opinions about environmental, generations to meet
social, and ethical issues their needs
▶ Doing what’s right can be beneficial to all stakeholders ▶ More than “going green”
▶ Corporate social responsibility (CSR) ▶ Includes employees, customers, community, and
company reputation
37 38
Commons Triple Bottom Line

• Many inputs to a production system held by the public ▶ Consider the systems necessary to support the three Ps:
• Common resources often misallocated people, planet, and profit
• Possible solutions include
1) Moving some of the common to private property
2) Allocation of rights
3) Allocation of yield
39 Figure S5.1 40
Triple Bottom Line Triple Bottom Line
People Planet
• Decisions affect people • The planet’s environment
• Globalization and outsourcing complicate the task • Look for ways to reduce the environmental impact of
• Supplier selection and performance criteria are important operations
• Materials must be safe for people • Overarching objective is to conserve scarce resources
• Carbon footprint and greenhouse gas (GHG) emissions
41 42
The 3 Rs: Reduce, Reuse &

Carbon Footprint Recycle
Figure S5.2
34.5-gram Bag of Frito-Lay Chips
43 44
Triple Bottom Line
Profit
• Social and environmental sustainability do not exist
without economic sustainability
• Staying in business requires making a profit
• Alternate measures of success include risk profile,
intellectual property, employee morale, and company
valuation
• Social accounting can supplement financial accounting
to support economic sustainability
45

Lectures Op

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lectures Op

Uploaded by

Copyright:

Available Formats

Today’s class

About me About you?

• Jay Heizer, Barry Render and Chuck Munson "Operations Management:

• Class attendance and participation (5%)

o 3 in-class quizzes (10% each)

• Related to theoretical sessions (magistrales) (35%)

o 3 Case deliverables (10% each, but only counting best 2, so 20%)

o Final project (15%)

Final Examination (30%): December 20th! 7 8

• In short: Matching supply with demand

Computer repair, restaurant meal

Automobile Repair, fast food

Home remodeling, retail sales

Operations Management Supply Chain Management

Delivered value of process =

Quality Q: Time T (Response time):

• DOES NOT CHANGE WITH QUANTITY

• CHANGES WITH QUANTITY

Understanding tradeoffs Caterpillar Inc.

Profit $63M $67M

High Caterpillar move: High Caterpillar move:

Shifting the frontier

Type of decisions in OM Outline of the course

• Part 4 (Final project)

Session 2 • Product/Process Matrix

American Connector Company • Start with Project Management…

Competitive position Compare Manufacturing Costs

Introduction to Operations Management

• Customization and • Efficiency in

Dynamic Costs Comparison Operations Strategies

What should ACC do? Production strategies

Slower Response Time Faster

Manufacturing Cell / Batch

Forester Power Body Engine Paint

A Continuum of Process Types Types of Processes

Examples of projects Management of projects

2. Scheduling - relate people, money, and supplies to

3. Controlling - monitor resources, costs, quality, and

1) Ensure that all activities are planned for

Ethical issues Work Breakdown Structure

Software Cost Management System

A Simple Gantt Chart Service For a Delta Jet

Variability in Activity Times Six Steps PERT and CPM

► CPM assumes we know a fixed time estimate for each

6. Use the network to help plan, schedule,

Questions PERT and CPM A Comparison of AON and

Copyright © 2020, 2017, 2014 Pearson Education, Inc. 3 - 26

AON Network for Milwaukee AON Network for Milwaukee

Determining the Project Schedule Determining the Project Schedule

► Slack is the length of time an activity can be B 0 3 1 4 1 No

Activities with zero slack are on the critical path. It:

► Includes only critical activities

Milwaukee Paper Milwaukee Paper EF = ES + Activity time

A Build internal A Build internal

Variability in Activity Times

Session 3 ► PERT uses a probability distribution for activity times to

Variability in Activity Times Computing Variance

Optimistic Most Likely Pessimistic 5 6

Probability of Project Completion Probability of Project Completion

► Total project completion times follow a

► Activity times are statistically independent

Z = Due – Expected date /sp .5 = due .69497

= 0.57 Where Z is the number of = 0.57 Where Z is the number of

Probability of Project Completion Variability of Completion Time for