Course Outline

 Linear Programming problems – Formulation, Graphical

method - Sensitivity Analysis and Duality
 Transportation, Transshipment and Assignment
problems
 Decision trees
 Network optimization models
 Integer programming problems
 Project management
 Simulation – A brief overview
 Multi-objective decision problems
 Dynamic Programming
 Queuing models
Course Outline
 Linear Programming problems – Formulation, Graphical
method - Sensitivity Analysis and Duality
 Transportation, Transshipment and Assignment
problems
 Decision trees
 Network optimization models
 Integer programming problems
 Project management
 Simulation – A brief overview
 Multi-objective decision problems
 Dynamic Programming
 Queuing models
Definition of a Project ??
 A project is a complex, non-routine, one-time effort
limited by time, budget, resources, and performance
specifications designed to meet customer needs
 A temporary endeavor to create a unique product,
service or result - Project Management Institute
 A project is an agreed upon set of resources and
activities that have been designed to produce
predetermined outputs within a given budget and a
time limit
Characteristics of a Project
 An established objective
 A defined lifespan with a beginning and an end
 Usually, the involvement of several departments and
professionals (interdependencies)
 Typically, doing something that has never been done
before (unique)
 Specific time, cost, and performance requirements
Scope of Project Management
 Project approach: Information Technology,
Construction, Defense, Hollywood, Port expansion,
Automobiles....
The Triple Constraint

Time
Time

Scope
Scope Cost
Cost

Iron Triangle or Project Management Triangle

Project scope or Project
Charter
Project scope or Project
Charter
 A Famous Example: “Put a man on the moon and
return him safely to Earth by the end of the decade
at a cost of $9 billion.”
◦ Scope – “Put a man on the moon and return him
safely to Earth.”
◦ Schedule – “By the end of the decade.”
◦ Resources –”At a cost of $9 billion.”
 Routine work Vs Project
Routine, Repetitive Work Projects
•Taking class notes •Writing a term paper
•Daily entering sales receipts •Setting up a sales kiosk for a
into the accounting ledger professional accounting
•Responding to a supply-chain meeting
request •Developing a supply-chain
•Practicing scales on the piano information system
•Writing a new piano piece
•Routine manufacture of an
Apple iPod •Designing an iPod that is
approximately 2 X 4 inches,
interfaces with PC, and stores
•Attaching tags on a 10,000 songs
manufactured product
•Wire-tag projects for GE and
Wal-Mart
Project Network
 Tool used for planning, scheduling and monitoring
project progress
 Information collected from WBS
 Network depicts
- Activities to be completed
- Logical sequences
- Interdependencies
- Start and Finish times
- Critical path
Work Breakdown Structure
(WBS)
Work Breakdown Structure
(WBS)
Example Problem
Job Predecessors Normal
time
(days)
A - 15
B - 10
C A, B 10
D A, B 10
E B 5
F D, E 5
G C, F 20
H D, E 10
I G, H 15
Network Techniques: PERT and
CPM
 The Program Evaluation and Review Technique (PERT)
was developed by the U.S. Navy in 1958.
 It is a network model that allows for UNCERTAINITY in
activity completion times.
 Helps in determining the probabilities of completing
various stages of the project by specified deadlines,
including the expected time to complete the project.
 Mostly used in research and development projects
Network Techniques: PERT and
CPM
 CPM was developed independently by E.I. du Pont de
Nemours Company
 Major difference between the two techniques is that CPM
does not incorporate uncertainties in job times
 Basic assumption is that the activity times are proportional
to the level of resources allocated to them
 Assigning additional resources (capital, people, materials
and machines) to an activity reduces its duration to a
certain extent
 Shortening the duration of an activity is known as crashing
in the CPM terminology. Additional cost is called Crashing
cost
Terminology
 Activity - A specific task or set of tasks that are
required by the project, use up resources, and
take time to complete
 Event - The result of completing one or more
activities. An identifiable end state occurring at a
particular time. Events use no resources.
 Network - The combination of all activities and
events define the project and the activity
precedence relationships
Terminology
Contd…

 Path - The series of connected activities (or

intermediate events) between any two events in a
network
 Critical - Activities, events, or paths which, if
delayed, will delay the completion of the project. A
project’s critical path is understood to mean that
sequence of critical activities that connect the
project’s start event to its finish event
Drawing Networks
 Activity-on-Arrow (AOA) networks use arrows to

represent activities while nodes stand for events

 Activity-on-Node (AON) networks use nodes to

represent activities with arrows to show precedence

relationships

 The choice between AOA and AON representation is

largely a matter of personal preference

 ACTIVITY ON ARROW NETWORK

A B C
1 2 4 5

E
D 3

ACTIVITY ON NODE NETWORK

B
A
C e
n
D d
E
Comparison of AON and AOA
NETWORK DIAGRAMS

Activity-on-arrow
NETWORK DIAGRAMS
 Some basic rules for Activity –on-arrow:
◦ Tasks are represented as arrows
◦ Nodes represent the start and finish points of tasks
◦ There is only one overall start node
◦ There is only one overall finish node
◦ Two tasks cannot share the same start and end node.
B

A D
2 3

C
Tasks B & C share
the same start and
end node
Dummy Activities

 Sometimes it is necessary to insert dummy activities (duration

zero) in order to maintain the clarity of the diagram and the
precedence relationships between activities.
 In activity-on-arrow diagrams, each activity must be uniquely
identifiable by its start and end nodes.
 However, sometimes multiple tasks have the same
predecessors and successors.
Dummy Activities
Contd…
 Dummy Activities
Contd…

 Inserting a dummy activity can ensure that

multiple tasks have different successors.

A B D E
1 2 3 5 6
C

A new node is A dummy task is

4
inserted to give inserted to
C a different preserve the
finish node to B immediate 25
predecessors of
D
 Dummy Activities
Contd…

 Inserting a dummy activity can ensure that

multiple tasks have different successors.

A C D E
1 2 3 5 6
B

A new node is A dummy task is

4
inserted to give inserted to
B a different preserve the
finish node to C immediate 26
predecessors of D
Dummy Activities
Contd…
Dummy Activities
Contd…

 The solution is to insert a dummy task so that the precedence

of E is preserved and activity C remains uniquely identifiable.
Example 1
Job Predecessors Normal
time
(days)
A - 15
B - 10
C A, B 10
D A, B 10
E B 5
F D, E 5
G C, F 20
H D, E 10
I G, H 15
Example 2
Job Predecessors Duration
(weeks)

A - 4

B - 7

C - 8

D A 5

E C 4

F B ,E 4

G C 11

H G, F 4
Solution to Example 2

3
A(4) D(5)

B(7)
F(4) H(4)
1 4 5 6

E(4)
C(8) G(11)
2
Finding the Critical Path
 Earliest start time
 Earliest Finish time
A D (12) E
 Latest Start time (10) (10)

C (8)
 Latest Finish time B
(8)

 Slack
 Critical job / Non-critical job
Find the critical path
Job Predecessors Duration
(days)

A - 15
B - 10
C A, B 10
D A, B 10
E B 5
F D, E 5
G C, F 20
H D, E 10
I G, H 15
Find the critical path
Job Predecessors Duration
(weeks)

A - 4

B - 7

C - 8

D A 5

E C 4

F B ,E 4

G C 11

H G, F 4
Program Evaluation and
Review Technique (PERT)
 Assumes each activity duration has a range that
statistically follows a beta distribution

 PERT uses three time estimates for each activity:

optimistic (a), pessimistic (b), and most probable
(m) to account for uncertainties

 Average time of an activity = (a + 4m + b) / 6

 Standard deviation of an activity= (b – a )/ 6

Program Evaluation and Review
Technique (PERT) Contd…
Some interesting stuff…
 Central Limit Theorem
The distribution of sum of large number of independent and
identically distributed (i.i.d) random variables tend to follow a
normal distribution
 Law of Large Numbers
 In probability theory, the law of large numbers (LLN) is a
theorem that describes the result of performing the same
experiment a large number of times.
 The average of the results obtained from a large number of
trials should be close to the expected value, and will tend to
become closer as more trials are performed
Some interesting stuff…
 Gambler’s Fallacy

The Gambler's fallacy, also known as the Monte Carlo

fallacy is the belief that if deviations from expected
behavior are observed in repeated independent trials of
some random process, future deviations in the opposite
direction are then more likely.
Some interesting stuff…
 Monte Carlo Casino (Source: Wikipedia)
 The most famous example happened in a
Monte Carlo Casino in the summer of 1913, when the
ball fell in black 26 times in a row, an extremely
uncommon occurrence.
 Gamblers lost millions of francs betting against black
after the black streak happened.
 Gamblers reasoned incorrectly that the streak was
causing an "imbalance" in the randomness of the
wheel, and that it had to be followed by a long streak
Consider a project consisting of nine jobs (A,B,C,D,E,F,G, H
and i). About each job, we know the following:

Job Predecessors Optimistic Most Pessimistic

time (days) Probable time
time
A - 2 5 8
B A 6 9 12
C A 6 7 8
D B,C 1 4 7
E A 8 8 8
F D,E 5 14 17
G C 3 12 21
H F, G 3 6 9
I H 5 8 11
Job Average Standard Variance
time Deviation

A 5 1 1

B 9 1 1

C 7 1/3 1/9

D 4 1 1

E 8 0 0

F 13 2 4

G 12 3 9

H 6 1 1
I 8 1 1
 The critical jobs are A, B, D, F, H and I
 Let T denote the project duration. Expected length of
the project is
E (T) = Sum of the expected times of critical jobs
= 5+9+4+13+6+8 = 45 days
 Variance of the project duration is
V(T) = Sum of the variances of critical jobs
=1+1+1+4+1+1=9
 Standard Deviation of the project duration = 3 days
 As per Central Limit Theorem, T is normally distributed
with mean 45 and standard deviation 3
 There is a 68% chance that the project duration
will be between 42 and 48 days (one sigma)
 Similarly, there is a 99.7% chance that T will lie
within three standard deviations (b/n 36 and 54)

Prob (T≤ 41) ???

Prob (T ≤ 50)???
Normal Variate
Z= x – mu / sigma
Z=(50 – 45) / 3
Z=1.666
Reducing Project
Duration -
Crashing
Why to reduce Project
Duration?
Reasons for imposed project duration dates:

 Customer requirements and contract commitments

 Time-to-market pressures

 Incentive contracts (bonuses for early completion)

 Unforeseen delays

 Overhead and goodwill costs

 Pressure to move resources to other projects

Project Costs
 Direct Costs
- Normal costs that can be assigned directly to a specific work
package or project activity
 Labor, materials, equipment, and subcontractors
 Crashing activities increases direct costs

 Project Indirect Costs

- Costs that cannot be associated with any particular work
package or project activity. Indirect costs increase with time.
 Supervision, administration, consultants, advertising and interest

 Reducing project time directly reduces indirect costs

Project Cost-Duration graph
Example
Job Predecessor Normal Normal Crash Crash
s time Cost time Cost
(days) (days)
A - 3 50 2 70
B A 6 80 4 160
C A 10 60 9 90
D A 11 50 7 150
E B 8 100 6 160
F C, D 5 40 4 70
G E, F 6 70 6 70

Indirect costs for 25 days:400

Savings in indirect costs per day = $50
Example: Network
4
E(8)
B(6)

A(3) C(10) F(5) G(6)

1 2 5 6 7

D(11)
3
 Example: Critical Path
9
11
4
E(8)
3 B(6)
3
0 A(3) C(10) F(5) G(6)
0 1 2 5 6 7

14 19 25
D(11) 14 19 25
3

14 Project Duration: 25 days

14 Total Cost = Direct cost + Indirect cost
Direct Cost = Normal cost + Crashing
cost
Direct Cost= 450+0 = 450
Indirect Cost = 400
Total Cost = 450 + 400 = 850
 Example: Crashing-Iteration 1
8
10
4
E(8)
2 B(6)
2
0 A(2) C(10) F(5) G(6)
0 1 2 5 6 7

13 18 24
D(11) 13 18 24
3

Action: Activity A 13 Project Duration: 24 days

crashed by one 13 Total Cost = Direct cost + Indirect cost
day Direct Cost = Normal cost + Crashing
cost
Direct Cost= 450+20 = 470
Indirect Cost = 350
Total Cost = 470 + 350 = 820
 Example: Crashing-Iteration 2
8
9
4
E(8)
2 B(6)
2
0 A(2) C(10) F(5) G(6)
0 1 2 5 6 7

12 17 23
Action: D(10) 12 17 23
Activity D
3
crashed by
one day 12 Project Duration: 23 days
IMPORTANT: Activity ‘C’ is 12 Direct Cost = Normal cost + Crashing
part of critical path now cost
Direct Cost= 450+20+25 = 495
Indirect Cost = 300
Total Cost = 495 + 300 = 795
 Example: Crashing-Iteration 3
8
8
4
E(8)
2 B(6)
2
0 A(2) C(10) F(4) G(6)
0 1 2 5 6 7

12 16 22
D(10) 12 16 22
3

Action: Activity F 12 Project Duration: 22 days

crashed by one 12 Direct Cost= 450+20+25+30 = 525
day Indirect Cost = 250
Total Cost = 525 + 250 = 775
IMPORTANT: Activities ‘B’ & ‘E’ are part of critical
path now
 Example: Crashing-Iteration 4
8
8
4
E(7)
2 B(6)
2
0 A(2) C(9) F(4) G(6)
0 1 2 5 6 7

11 15 21
D(9) 11 15 21
3

Action: Activities 11 Project Duration: 21 days

C, D and E 11 Direct Cost = Normal cost + Crashing
crashed by one Cost
day Direct Cost= 450+20+25+30
+30+30+25= 610
Indirect Cost = 200
Total Cost = 610 + 200 = 810
Example
Contd…
Example
Contd…
Linear Programming for
Project Management
Example
Consider five jobs A, B, C, D, and E with the
following job sequence:
Job A precedes C and D
Job B precedes D
Jobs C and D precede E
A C E
1 2 4 5
3 4 5
B
D
1 2
3
Solution by Linear
Programming
Minimize: Z = t 5 – t1 Minimize: Z = tn – t1
Subject to: tj – ti ≥ tij
Subject to: t2 – t1 ≥ 3
t3 – t 1 ≥ 1 tj ≥ 0
t1=0 ti – time at which event
t3 – t 2 ≥ 0
t2=3 t4 – t 2 ≥ 4 ‘i’ occurs
A C E
t3=3 t4 – t 3 ≥ 2 1 2 4 5
3 4 5
t4=7 t5 – t 4 ≥ 5 B
D
1 3 2
t5=12
Solution by Linear
Programming
Minimize: Z = t 5 – t1 A C E
1 2 4 5
3 4 5
Subject to: t2 – t1 ≥ 3 B
D
t3 – t 1 ≥ 1 1
3 2

t3 – t 2 ≥ 0
t4 – t 2 ≥ 4
t1=0 t2=3 t3=3 t4=7
t4 – t 3 ≥ 2
t5 – t 4 ≥ 5 t5=12
LP for Project Management –
Model 1
 Given that the project must be completed by time T, we
want to determine how the project activities are to be
expedited such that the total cost of crashing is
minimized
 Kij – Normal completion time of activity (i, j)
 Lij – Crash completion time with maximum resources
 Cij – Cost of shortening the duration of job(i, j) by one
time unit
 tij- Actual completion time
 (Lij ≤ tij ≤ Kij)
LP for Project Management –
Model 1
Minimize: Z = C
(i , j )
ij (kij  tij )

Subject to: tj – ti ≥ tij

lij ≤ tij ≤ kij
tn – t1 ≤ T
tj ≥ 0
LP for Project Management –
Model 2
 Suppose an additional budget of B dollars is available

for crashing the project activities. We want to

determine how these additional resources may be

allocated in the best possible manner so as to

minimize the project completion time

LP for Project Management –
Model 2
Minimize: Z = tn – t1

Subject to: tj – ti ≥ tij

lij ≤ tij ≤ kij
C
(i , j )
ij (kij  tij )
≤B

tj ≥ 0
LP for Project Management –
Model 3
Minimize: Z = F (tn – t1) +  Cij (kij  tij )
(i , j )

Subject to: tj – ti ≥ tij

lij ≤ tij ≤ kij
tj ≥ 0
F – Indirect costs per time unit
Example
Job Predecessor Normal Crash Cost of
s time Time crashin
(days) (days) g per
day Overhead
costs - $5
A - 10 7 4 per day

B - 5 4 2
C B 3 2 2
D A, C 4 3 3
E A, C 5 3 3
F D 6 3 5
G E 5 2 1
H F, G 5 4 4
 Example
Minimize Z = 5(t7-t1)+4(10-t13)+2(5-t12)+2(3-t23)+3(4-t34)+
3(5-t35)+5(6-t46)+1(5-t56)+4(5-t67)

Subject to
t3 – t1 ≥ t13 3≤ t35 ≤ 5
t2 – t1 ≥ t12 3≤ t46 ≤ 6
t3 – t2 ≥ t23 2≤ t56 ≤ 5
t4 – t3 ≥ t34 4≤ t67 ≤ 5
t5 – t3 ≥ t35
t6 – t4 ≥ t46 t1, t2, ……t7 ≥ 0
t6 – t5 ≥ t56
t7 – t6 ≥ t67
7≤ t13 ≤ 10
4≤ t12 ≤ 5
2≤ t23 ≤ 3
3≤ t34 ≤ 4