ESD.

36 System & Project Management

+

Lecture 2

Critical Path Method (CPM)

Instructor(s)

Prof. Olivier de Weck

Todays Agenda

Overview of PM methods and tools

CPM 101
Critical Paths, Slack
Task Crashing and Cost
Conclusions and Class Discussions
Introduce HW3

- ESD.36 SPM

History of Project Management

Big Projects since antiquity

Pyramids (Egypt), Great Wall (China)

Enormous workforce, but little documented evidence of
formal project management

Formal Project Management

Henry Gantt (1861-1919) bar chart 1910

1957 Sputnik Crisis revival of scientific management
Polaris (1958) Project Evaluation and Review
Technique (PERT)
DuPont Company (1960) Critical Path Method (CPM)
1960s NASA projects: Mercury, Gemini, Apollo

Work Breakdown Structures (WBS)

Cost and Schedule Tracking, Configuration Management

- ESD.36 SPM

Comments about early PM

Project decomposition necessary due to complexity

Resource allocation and workload smoothing
Schedule urgency ..before the decade is out JFK
Circumstances

Other Innovations

Complex Relations between Government and Contractors

Shielded from Society, Competition, Regulations
Cold War Pressures for Nuclear Power, Space Race ..
Project Manager as a central figure
Beginnings of Matrix Organization
Earned Value adopted by USAF (1963)

Professionalization since 1969

Diffusion into other industries: computers, automotive

Project Management Institute (PMI) founded PMBOK
ISO 10006:2003 Quality in Project Management
Recent criticism about PM standards as bureaucratic

- ESD.36 SPM

Fundamental Approaches

How to represent task relationships?

Network-based (graph theory) methods

Matrix-based methods

CPM, PERT, .
Task is a node or an arc

DSM - Tasks are columns and rows

Interrelationships are off-diagonal entries

Feedback loops, causal relationships

Stocks and flows simulation
Tasks that are done or waiting to be done
are stocks amount of work
Doing project work causes a flow

WORK
TO BE
DONE

- ESD.36 SPM

x x
xx
x x
x x
x x

System Dynamics

PEOPLE

PRODUCTIVITY

WORK
BEING DONE

WORK
DONE

Gantt Charts

Attributed to Henry Gantt most popular PM tool (80%)

Used to plan big shipbuilding projects (cargo ships WWI)
Graphical way of showing task durations, project schedule
Does not explicitly show relationships between tasks
Limited use for project tracking
Easy to understand
calendar
today
completion

Gantt Chart Builder System (Excel) 1.6

08 Sep'03

Bus Unit
Mkt
Sys
Eng
Sys
Mfg
Sys

Project
Project "XYZ"
Customer Clinic
Requirements Definition
Parts Design
Design Review
Manufacturing
Product Release

- ESD.36 SPM

tasks

%
30
100
100
50
0
0
0

Start
9-Sep-03
09-Sep-2003
11-Sep-2003
15-Sep-2003
23-Sep-2003
24-Sep-2003
06-Oct-2003

Finish
6-Oct-03
12-Sep-2003
15-Sep-2003
22-Sep-2003
23-Sep-2003
05-Oct-2003
06-Oct-2003

08/09

15/09

actual

22 Sep'03
22/09

29/09

06 Oct'03
06/10

13/10

milestone

planned
6

CPM 101

Represent a project (set of task) as a

network using graph theory

Capture task durations

Capture task logic (dependencies)
Expected
Duration

Task ID (Series)
A,5

B,8

B can only start after A

is completed
- ESD.36 SPM

(Parallel)
B,8
A,5

D,3
C,2

B and C do not depend

on each other
7

CPM Assumptions
Project consists of a collection of well defined
tasks (jobs)
Project ends when all jobs completed
Jobs may be started and stopped
independently of each other within a given
sequence (no continuous-flow processes)
Jobs are ordered technological sequence

- ESD.36 SPM

Task Representations

Tasks as Nodes of a Graph

Circles
Boxes

A,5

Tasks as Arcs of a Graph

ES
ID:A
LS

EF
Dur:5
LF

Tasks are uni-directional arrows

Nodes now represent states of a project
Kelley-Walker form
broken

- ESD.36 SPM

A, 5

fixed
9

Concept Question 1

What in project management is named after

Henry Gantt (1861-1919)?

A new method for managing budgets in shipyards

A matrix representation of the task relationships
known as the Gantt-Matrix
A horizontal bar chart known as the Gantt-Chart
Gantt-teams, a new form of organizational
structure
Gantt was a mechanical engineer and did not care
about projects

- ESD.36 SPM

10

Work Breakdown Structure

Used to create the task (job) list
Tree-decomposition of project tasks
WBS identifies terminal elements
The key starting point for project planning
Required by US Govt as part of SOW
Can be activity-oriented or deliverable- oriented
Use sticky-notes method early on
Carl L. Pritchard. Nuts and Bolts Series 1: How to
Build a Work Breakdown Structure. ISBN
1890367125
Job A
- ESD.36 SPM

Job B

Job X

Job G
11

WBS Painting a Room

1 Prepare materials

2. Prepare room

1.1 Buy paint

1.2 Buy brushes/rollers
1.3 Buy wallpaper remover
2.1
2.2
2.3
2.4
2.5

Remove old wallpaper

Remove detachable decorations
Cover floor with old newspapers
Cover electrical outlets/switches with tape
Cover furniture with sheets

4.1
4.2
4.3
4.4

Dispose or store left over paint

Clean brushes/rollers
Dispose of old newspapers
Remove covers

3. Paint the room

4. Clean up the room

- ESD.36 SPM

Source: [Link]

12

WBS of MIT/NASA Space

Logistics Project

IP S C M & L A

1 . 0 P r o je ct
M a n a g e m e nt

1 . 1 M I T Pr o je ct
M a n a g e m e nt

1 . 1 . 1 M I T Pr o je ct
M a n a ge r / PI

1 . 1 . 2 M IT T ra v el

2 . 0 T e rr e st r i al
A n a l o g i es
( T a s k 1)

2 . 1 L o w D e n si t y E q ui p m e nt
S C M L e s s on s L e a r n ed

3 . 0 S pa c e L og i s t i cs
N e t w o r k M o d el
( T a s k 2)

3 . 1 M o d e l Fo r m ul a t i on

4 . 0 E x pl o r a t i on
S u p p l y /D e m a nd
M o d e l i ng
( T a s k 3)

5 . 0 In t e rp l a ne t a ry
S C A r ch T r a de
S t u d i es
( T a s k 4)

4 . 1 S up p l y / D e m a nd
D a t a M o de l i ng

2 . 2 Lo n g -R a n ge M i l i t a ry
S C M L e s s on s L e a r n ed

3 . 1 . 1 M I T S e gm e nt

4 . 1 . 1 M I T S e gm e nt

5 . 1 . 1 M I T S e gm e nt

2 . 3 R em o t e En v i r on m e n ts
S C M L e s s on s L e a r n ed

3 . 1 . 2 J P L S e g m e nt

4 . 1 . 2 J P L S e g m e nt

5 . 1 . 2 J P L S e g m e nt

3 . 2 M o d el
D e v e l o p m e nt

4 . 1 . 3 U S A S e g m e nt

1 . 1 . 3 M I T R e so u r ce
A d m i n i s t ra t i on

1 . 1 . 4 R e s e r v es

2 . 4 I SS S C M
L e s s o n s L e a r n ed

1 . 2 J P L P r o je ct
M a n a g e m e nt

1 . 2 . 1 J PL P r o je ct
M a n a g er / C o - PI

6 . 3 In t e r na l I nt e g ra t i on

2 . 5 Sh u t t le / I SS
L e s s o n s L e a r n ed

4 . 2 D a ta
C o l l ec t i on

3 . 3 M o d el
V a l i da t i on

3 . 3 . 1 M I T S e gm e nt

4 . 2 . 2 U S A S e g m e nt

4 . 3 D a ta
V a l i da t i on

1 . 3 . 3 U S A R es o u r ce
A d m i n i s t ra t i on

- ESD.36 SPM

5 . 3 T r a de S t u d i es
R e p or t i ng

6 . 3 . 2 J P L S e g m e nt

6 . 4 P ub l i c E du c a t i on
a n d O ut r e a ch

6 . 4 . 1 M I T S e gm e nt

6 . 4 . 2 J P L S e g m e nt
3 . 3 . 2 J P L S e g m e nt

4 . 3 . 1 J P L S e g m e nt

3 . 3 . 3 U S A S e g m e nt

4 . 3 . 2 U S A S e g m e nt

6 . 4 . 3 U S A S e g m e nt

1 . 3 U S A P r o je ct
M a n a g e m e nt

1 . 3 . 2 U S A T r a v el

6 . 3 . 1 M I T S e gm e nt

4 . 2 . 1 J P L S e g m e nt

1 . 2 . 4 R e s e r v es

1 . 3 . 1 U SA P r o je ct
M a n a g er / C o - PI

5 . 2 T r a de S t u d i es
E x e c u tio n

6 . 3 . 3 U S A S e g m e nt
3 . 2 . 2 J P L S e g m e nt

1 . 2 . 2 J P L T r a v el

1 . 2 . 3 J P L R e s o u r ce
A d m i n i s t ra t i on

6 . 1 N A SA W o rk s h o p I

6 . 2 N A SA W o rk s h o p II

3 . 2 . 1 M I T S e gm e nt

1 . 3 . 4 R e s e r v es

5 . 1 T ra d e St u d i es
P l a n n i n g a n d S e t up

6 . 0 P r o je ct
I n t e gr a t i on a nd
D i s se m i na t i on

6 . 5 S p ac e S C M
E x e c u t i v e Sh o r t C o u r se

MIT (incl. PSI):

JPL:
USA:

cyan
gray
green

Color code WBS by

Organization in charge
13

Discussion Point 1

Why is it difficult to come up with a

good WBS (task list, task structure) in a
complex project?

Not all tasks known ahead of time if

completely new product/system
Others?

- ESD.36 SPM

14

WBS Guidelines

No more than 100-200 terminal elements, if

more use subprojects
Can be up to 3-4 Levels deep
Not more than 5-9 jobs at one level

Human cognitive bandwidth only 3 bits=23=8

Short term memory for most people 5-9 items
Poorer planning if too-fine grained dilution of
attention
The more tasks there are, the more intricate
dependencies there will be to keep track of

Jobs should be of similar size/complexity

Manageable chunks sense of progress
Level of graininess very difficult to answer
- ESD.36 SPM

15

Task List

List all tasks in a table with

Identifying symbol (tag, ID number)

Task description
Immediate prerequisite jobs
Expected task duration

Arrange jobs in technological order

No job appears in the list until all its predecessors

have been listed
Iterations are NOT allowed cycle error
Job a precedes b precedes c precedes a
We will discuss iterations a lot in this class !!!

- ESD.36 SPM

16

Simple Example: Job List

Two Parts X and Y: Manufacture and Assembly

Job #

Description

Start

Get materials for X

10

Get materials for Y

20

Turn X on lathe

B,C

30

Turn Y on lathe

B,C

20

Polish Y

40

Assemble X and Y

D,F

20

Finish

- ESD.36 SPM

Immediate
Time
Predecessors [min]
0

17

Project Graph
Each job is drawn on a graph as a circle*
Connect each job with immediate predecessor(s)
unidirectional arrows
Jobs with no predecessor connect to Start
Jobs with no successors connect to Finish
Start and Finish are pseudo-jobs of length 0
A finite number of arrow paths from Start to
Finish will be the result
Total time of each path is the sum of job times
Path with the longest total time critical path
There can be multiple critical paths minimum time to
complete project
* or other symbol, see before
- ESD.36 SPM

18

Project Graph
F,40

E,20

C,20

critical path

G,20

A,0

H,0
Finish

Start

B,10

D,30

4 unique paths: A,C,E,F,G,H; A,C,D,G,H; A,B,D,G,H; A,B,E,F,G,H

- ESD.36 SPM

100

70

60

90

19

Critical Path

CP is the bottleneck route

Shortening or lengthening tasks on the critical path
directly affects project finish
Duration of non-critical tasks is irrelevant
Crashing all jobs is ineffective, focus on the few % of
jobs that are on the CP
Crashing tasks can shift the CP to a different task
Shortening tasks technical and economical challenge

How can it be done?

Previously non-critical tasks can become critical

Lengthening of non-critical tasks can also shift the
critical path (see HW1).
- ESD.36 SPM

20

Discussion Point 2

What is the usefulness of knowing the

CP in a project?

Tells which task to shorten to finish project

earlier.
Others ?

- ESD.36 SPM

21

Critical Path Algorithm

For large projects there are many paths

Need a algorithm to identify the CP efficiently
Develop information about each task in context of the
overall project
Times

Start time (S)

For each job: Earliest Start (ES)

Earliest start time of a job if all its predecessors start at ES

Job duration: t
Earliest Finish (EF)=(ES)+t

Finish time (F) earliest finish time of the overall project

Show algorithm using project graph
- ESD.36 SPM

22

1.
2.

3.

4.

CP Algorithm
Mark the value of S to left and right of Start
Consider any new unmarked job, all of whose
predecessors have been marked. Mark to the
left of the new job the largest number to the
right of its immediate predecessors: (ES)
Add to ES the job time t and mark result to
the right (EF)
Stop when Finish has been reached

- ESD.36 SPM

23

CP Algorithm - Graphical

critical path
0

C,20

20

20

E,20

40

F,40

40

S=0 A,0 0

80

80

Start

0 B,10 10

- ESD.36 SPM

Finish

G,20

H,0

100 100 F=100

20 D,30 50

24

Concept Question 2

A project starts with (A,5). Task (B,10)

can start after A is completed. This is
also true for task (E,5). Task (C,8)
depends only on (B,10), while task (F,10)
depends on both (B,10) and (E,5). Task
(D,5) is the last task in the project and it
can start once (C,8) and (F,10) have
been finished.
The Earliest Finish (EF) time for the
whole project is:

- ESD.36 SPM

Possible
Answers
20
22
25
27
30
35
40
25

Latest Start and Finish Times

Set target finish time for project: T >= F
Usually target is a specific calendar date,
e.g. October 1, 2007
When is the latest date the project can be
started?
Late Finish (LF) - latest time a job can be
finished, without delaying the project
beyond its target time (T)
Late Start: LS = LF-t
- ESD.36 SPM

26

Determine LF and LS

Work from the end of the project: T

1. Mark value of T to left and right of Finish
2. Consider any new unmarked job, all of
whose successors have been marked - mark
to the right the smallest LS time marked to
the left of any of its immediate successors
3. Subtract from this number, LF, the job time t
and mark result to the left of the job: LS
4. Continue upstream until Start has been
reached, then stop

- ESD.36 SPM

27

LS and LF : Project Graph

0 20
0 20

20 40
20 40

40 80
40 80

C,20

E,20

F,40
Finish
G,20

S=0 A,0

80 100 100 100

80 100 100 100

Start

0 0
0 0

H,0

0 B,10 10

D,30

0 10
10 20

20 50
50 80

- ESD.36 SPM

Legend
early ES EF
late LS LF
28

Slack

Some tasks have ES=LS --> no slack

Total Slack of a task TS=LS-ES
Maximum amount of time a task may be delayed
beyond its early start without delaying project
completion
Slack time is precious managerial freedom, dont
squander it unnecessarily

e.g. resource, work load smoothing

When T=F then all critical tasks have TS=0

At least one path from Start->Finish with critical jobs
only
When T>F, then all critical jobs have TS=T-F
- ESD.36 SPM

29

Project Graph - Slack

0 20
0 20

20 40
20 40

40 80
40 80

C,20

E,20

F,40

TS=0

TS=0

TS=0

0 0
0 0

TS=0 TS=0
G,20

S=0 A,0
Start

Finish

TS=10

TS=30

0 B,10 10

D,30

0 10
10 20

20 50
50 80

- ESD.36 SPM

H,0

80 100 100 100

80 100 100 100
Legend
early ES EF
late LS LF
30

Task Times Detail - Task i

LS(i)
ES(i)
Duration t(i)
Total Slack
TS(i)
FS(i)

LF(i)

EF(i)

Duration t(i)

j is the immediate
successor of i with
the smallest ES

ES(j)

j>i

Free Slack

Free Slack (FS) is the amount a job can be

delayed without delaying the Early Start (ES)
of any other job.
FS<=TS always

- ESD.36 SPM

31

Main CPM Errors

Estimated job times are wrong
Predecessor relationships may contain cycles cycle
error
List of prerequisites contains more than the immediate
predecessors, e.g. ab, bc and a,bc
Overlooked some predecessor relationships
Some predecessor relationships may be listed that are
spurious
and . Some tasks/jobs may be missing !!!

- ESD.36 SPM

32

Gradual Refinement of CPM

Job Times

Predecessor Relationships

Given rough time estimates construct CPM chart

Re-estimate times for CP and those with very small TS
Iterate until the critical path is stable
Focus attention on a subset of tasks

Check algorithmically for cycle errors and pre-predecessor

errors
Cancel all except immediate predecessor relationships

Wrong or Missing Facts

Cannot be detected by computers!

- ESD.36 SPM

33

Sample CPM (2007)

A&D High Tech Case, Online Store Project, 60 tasks, HW2, 2007
- ESD.36 SPM

34

Crashing Tasks

What is we want to speed up project

completion?
Options

Work overtime
Put more resources on the critical tasks
Parallelize tasks that are really serial (later in class)

Cost of speedup?
Is there a net savings resulting from
reduction in overall project time?

- ESD.36 SPM

35

Cost Calculations
Can compute project costs if cost of each job is
included in the task data
(Potentially) shorten crew jobs by adding personnel or
working overtime with existing personnel
Speedup carries price tag: normal time, crash time
Assign some critical jobs to their crash time
Direct costs will increase as we crash critical tasks
Indirect (fixed, overhead) costs will decrease as the
overall project duration decreases standing army
phenomenon
Minimize the sum of fixed and direct costs
- ESD.36 SPM

36

Typical Cost Pattern

Project
Costs [$]

Total Costs
natural pace
savings

Cum Fixed Costs

Direct Costs

A
- ESD.36 SPM

crash

Total Project
Time [days]
37

CPM Judgment

Focuses attention on a subset of critical tasks

Determine effect of shortening/lengthening tasks
Evaluate costs of a crash program

Doesnt capture task iterations, in fact

Prohibits iterations = cycle error
Treats task durations as deterministic

- ESD.36 SPM

38

Summary

CPM is useful, despite criticism, to identify the

critical path - focus on a subset of the project
Slack (TS and FS) is precious

PERT treats task times as probabilistic

apply flexibility to smooth resource/schedules

Next lecture

Selective crashing of critical tasks can

reduce (or increase) total project cost
CPM does not allow for task iterations
- ESD.36 SPM

39

Class Frustrations

Poor examples set by project managers

Perception of PMs as bureaucratic boxcheckers and ankle-biters
Why?

Traditional Project Management

Doesnt acknowledge the existence of iterations

Is inflexible, changing the plan considered a failure
Does not think of projects in a probabilistic sense
Hostage to existing project management software
In a reactive mode no early warning systems
- ESD.36 SPM

40

HW1 Introduction

You are Project Manager for the new

CityCar development project
Plan the project

Task list
Create project graph
Critical path
Slack times
Re-planning after change
managerial-type questions
- ESD.36 SPM

41

MIT OpenCourseWare
[Link]

ESD. 6\VWHP3URMHFW0DQDJHPHQW
Fall 2012

For information about citing these materials or our Terms of Use, visit: [Link]