Chapter 16: Program Evaluation

and Review Techniques (PERT)

PERT

• PERT is a method for determining the

length of a construction project and the
probability of project completion by a
specified date

• PERT is based on probabilistic activity

durations

1
• Recall that AON diagrams were based
on deterministic activity durations

• When we assume that the duration of

activity “rebar columns” is 10 days, what
does that really mean?

– will “rebar columns” take exactly 10 days to

complete?
– or will the actual duration vary from the
estimated duration?

• It could mean that, on average, the

duration is 10 days

• To accommodate the uncertainty

associated with activity duration
estimates, PERT is based on
probabilistic activity durations

2
 • Since construction companies engage
in work that they have done in the past,
this results in multiple occurrences of
the same activity and a historical record
of durations or productivities

• PERT relies on activity durations that

are established either by an analysis of
historical data or through estimates of
the range of probable activity durations

• Such data can be shown as a frequency

histogram like the one shown below

Histogram of Duration (days), with Normal

Curve for a construction activity

4
Frequency

8 9 10 11 12 13 14 15 16

Duration (days)

Source: Weber (2005, p.226)

3
 • No matter of the actual distribution, there are
three measures of central tendency: mean,
mode, and median

Histogram of Duration (days), with Normal

Curve for a construction activity

4
Frequency

8 9 10 11 12 13 14 15 16

Duration (days)

Source: Weber (2005, p.226)

Histogram of Duration (days), with Normal

Curve for a construction activity

4
Frequency

8 9 10 11 12 13 14 15 16

Duration (days)

• Mean = 11.48
• Mode = 10 (most frequent occurrence)
• Median = 11 (equal number of observations above it
and equal number of observations below it)
• Note also that the range of observations = 16 – 8 = 8

4
• If all activities have been performed
multiple times in the past enough times
to generate a frequency histogram, a
sample can be taken from each
distribution that will give a duration for
each activity

• Activity durations in PERT are based on

three time estimates:
– Optimistic duration
– Most likely duration
– Pessimistic duration

Histogram of Duration (days), with Normal

Curve for a construction activity

4
Frequency

8 9 10 11 12 13 14 15 16

Duration (days)

• Optimistic duration: assumes maximum productivity

– How many days in this example?

• Pessimistic duration: assumes the worst productivity

– How many days in this example?

5
 Histogram of Duration (days), with Normal
Curve for a construction activity

4
Frequency

8 9 10 11 12 13 14 15 16

Duration (days)

• Most likely: most often based on historical

performance
– How many days in this example?

Calculating the mean estimate of

duration
• The mean estimate of the activity duration is
computed as follows

to + 4tm + tp
te =
6

6
 to + 4tm + tp
te =
6
te = mean or expected activity duration
to :optimistic activity duration
tm : most likely activity duration
tp : pessimistic activity duration

Network calculations

• In PERT, project duration is called

“project mean duration” (Te)

• Te is calculated based on the regular

forward pass using the activity mean
durations te for every activity

7
 Example

Calculating the Standard deviation

• Note that the mean value of the activity

duration does not convey any information
about the degree of uncertainty

• It would be helpful to have a measure to

describe the extent to which the duration is
expected to vary from the derived mean value

• Such a measure is known as the Standard

Deviation (S)

8
• We can use S to describe the extent to which the
duration is expected to vary from the derived
mean
Standard deviation (S) = Range of activity durations
6

tp − to
S=
6
• Note that 6 in the equation refers to ±3 standard
deviations from the mean of a normal distribution, which
contains 99.73% of all population values

The variance

tp − to 2
Variance(V ) = S = ( 2
)
6
S= V
• Note that S Pr oject = SCP =  V CP

9
 Example

Slack

• In PERT, what we used to know as

“float” is called “slack”

• Activity Total Slack = ATS

• Activity Free Slack = AFS

10
Calculating the probability of meeting
deadline dates
• Based on the normal distribution, we can
calculate the probability of project completion
within certain duration

• The probabilities of occurrence of a specific

duration can be determined by simply
knowing the number of standard deviations
that the value in question is away from the
mean

• The “standard normal curve areas” table

is set up to give information of the
probability that a particular duration will
be less than some specified value that
is given in terms of the number of
standard deviations that the value
extends beyond the mean

11
 Te = 24 days Ts = 27 days

This is the normal distribution

The probability to complete the project in
24 days (mean duration) or less = 50%,
which is the area under the curve

Te = 24 days Ts = 27 days

Now to find the probability of completing

the project in 27 days, we need to find out
the number of standard deviations that Ts
(specified date) is away from Te
T s − Te
Z=
SCP

12
 T s − Te
Z=
SCP
Ts = 27 days

27 − 24
Te = 24 days

Z= = 1.43
2.095

From the table, Z=1.43; probability = 0.9236

Therefore the probability of project completion in 27

days or less = 92.4%

PERT

13
 PERT
• The Program Evaluation and Review Technique, commonly
abbreviated PERT, is a statistical tool, used in project
management, that is designed to analyze and represent the tasks
involved in completing a given project.

• Example
– For the following project determine the
following !
– The Critical Path ( C.P)
– What is The probability of finishing the project before or on day number 21?
– What is the finish time for the project with a probability of 95 % ?

14
Activity Pre. Duration
To Tm Tp
A - 2 3 5
B A 4 6 7
C A 1 2 3
D A 3 4 5
B
E 2 3 5
C
C
F 2 3 4
D
E
G 1 2 4
F
B
H 2 4 5
F
D
I 1 2 4
F
G
N H 2 3 4
I

To Tm Tp

Activity

Te

15
 4 6 7 2 4 5

B H
5.83 3
3.83
2 5

E
3.17 2
2 3 5 1 2 3 1 4 2 3 4

A C G N
3.17 2 3
2.17 3
2 4

4
3 2
3 5 1 4

D I
4 2.17

Determination of Critical Path :

- Each Path Give us a specific duration, and we will take the
longest :
15.83 Days 15.34 Days

17.34 Days 15.34 Days

19 Days 12.34 Days

13.34 Days

13.51 Days
C.P

16
Critical Path Activities Are :

A,B,E,H,N Te (project) = 19 days

2) What is The probability of finishing the project before or on day number 21?

Z = (Ts – Te ) / Scp

Standard Deviation For The Project (Scp) = ∑ √S² , where S : standard deviation for each Critical activity

S for each activity = (Tp – To ) / Activity S

6 A 0.50
B 0.50
E 0.50
H 0.50
N 0.33

17
 So, :

Scp = ∑ √ (.5²) + (.5²) + (.5²) + (.5²) + (.33²) = 1.053

And : Z = (21 – 19 ) / 1.053 = 1.9

So, the probability of completing the

project
In 21 days is 97.13%

3) What is the finish time for the project with a probability of 95 % ?

From the Z-Table we find that the Z value

that has a probability of 95% equals to =
1.65

So, Z = (Ts – Te) / Scp

1.65 = ( Ts – 19 ) / 1.053

We Get Ts = 20.74 Days

18