Unit 4. p .

• ro1ect M
anagement By PERT-CPM
Job 4.325
E Optimum Duration (days) : (1, 2) (1, 3)
(1,4) (2, 4)
and xarnple 11. In th . • 9
(3, 4) (4, 5)
the act· • • e 8 14
iv,t,es by letterspro1
5
Thect network shown m
cost . 6 1
s are shown below : • e normal and crash the figure given below , the node s are denoted •
durations of the various activities along with theby numbers
associated

Fig..
7 34
Activity
Normal duration Normal cost
(days) Crash duration Crash cost
(Rs.) (days)
A 8
(Rs.) 1
1800 6 2200
B 16 1500 11 2200
C 14 1800 9 2400
D 12 2400 9 3000
E 25 800 14 '2000
F 10 2000 8 4000
Determine the least cost 36 days schedule.
Solution. First assume that all activities occur .
at nonnal times. Then the following network
critical path computations under nonnal cond shows the
itions. The critical path is A B C F. The
project is 48 days and its associated nonnal cost schedule of the -
becomes
= Rs.( 1800 + 1500 + 1800 + 2400 + 800 + 2000) =
Rs. I0,300.
The different minimum cost schedule that can
oeeur between nonnal and crash times, which
dependent on the cost time slopes for different are mainly
activities. The cost time slopes can be computed
by the formula :
. SI
Cost-time ope Crash cost- .
Nonnal cost
.
Normal time -Crash time

E6 =48
~=4 8

Fig. 7.35
• ·t'
o·es for the act1v1, 1es of the above netwprk are obtained as follows :
These Sl0 ~ . A B C O E F
Activity • 200 i 40 120 200 1200 1ooo
Slope
 Operations n
8'esea
• "c~
• 4.32 6
• be reduced by crashi
Now proceed step-by-step as follows : • on the critical pa~g :hllle of'"-
Step 1. Since the present schedule dura consumes more time, the sched_ul~.ca~ying • ' e durati"'
tion is controlled by the acuv iues
activities. Since the project
ced. . . st Acti vity C with m • • °"
ities on the critical path is redu to •~•m um
which inv~lves !'1 mmum cod ·from
of som e activ 1 14 days 9 s)°lle
First reduce the duration of that activity of activity C is compresse nds to 43 da . ays With
give s the min imu m cost . So the dura tion ys WJth a cost:
efore, new schedule correspo
additional cost Rs.5 x 120 = Rs.600. Ther
. more time and als 0 not a))•"-
Rs.(10300 + 600) =Rs.1 0900.
rved that the pres ent sche dule stdl ,con sum es . d .
on cant'be. red Uce<f -.,
Step 2. Now, it can hbe obse th · d s Hen ce the proJ ect uratJ
th th b
· · path are at e1r eras ura ion • h t • .. 1
. · · on t e cnt1
activities cal on the cnti ca pa • de ac IVJly B has 0,?
remaining activ ities
crashing some other activity. Out of the B ~rom l 6 days 38 to
11 ays_ at a cost c
redu ce the dura tion of acti vity th a cost 0:
le~ t slope. So ect dura tion beco mes day s wi 0
· proj
Rs. = 5 x 140 = Rs. 700. Thus the new
Rs.(10900 + 700) = Rs.1 ) 600.more than the required duration. o~ _36 days. So sele ct so1!1_e other activj~
This project duration is still cntical path can be •
Step 3. iously, only the act1v1t1es A an~ Fon tbe
lyin g O.Q the critical path for crashing. Obv be compresSed.
A has the smaller ~Jope, the duration of A6 days). Because, the
can
considered for crashing. _Since activity to
can be compressed by 2 dafs (from 8
Compress A by only one day although it as A 1s com pres sed by one day. Tht1s a
path 1 4 5 6 becomes a parallel critical path as soon
a cost of Rs,( l I 600 + 200) = Rs •.11800.
new schedule corresponds to 37 days with one day· To do so compress one
is required, compress some activity by
Step 4. Since only 36 days schedule
s. So there are three choices :
day in each of the two parallel critical path
( i) . Acti vity F can be compressed by
one day at a cost of Rs. 1000. e alre ady at their crash
Acti vitie s A and D can be com pres sed by one day each (since Ban d Car
(iz)
(200 + 200) = Rs.400.
points). This give s the total cost of Rs. pressed by one day each at a total cost of
can . be com
(iii) Acti vitie s A and F
. =
Rs. (200 + 1200) Rs.. 1400.
sche
.
dule and hence it should be sele cted . This invo
lves a 36 days
But, th_e seco nd cho ice give s the ]eas t cost •
• -
a cost of Rs.( 118 00 + 400 ) = Rs.( 12,2 00). •
sche dule with technique.
dule form the following project usig CPM
Example 12. Determine the least cost sche
,,

_
Overhead cost per day is Rs. 6. . H.I.Sc. (M.Sc., Appl. Maths) 77]
_ have their usual meaning
The numbers above and below the activities . .

Fig. 7.36.
found as
. Tak ing into acco unt the norm al duration of the project, the criticaJ path is
Solution
1 .
ities. Since
t cost schedule, com ~ress the duration by crashing some of the activI • on the
In .o~~ er to ~ete rmin e th~ !eas ·
· s y1ng
, the duration of some ac u·v1tte
on the cntJcaJ path control the proJ• ect duration
th~ _act1vit1es- lyinbeg shor
cnticaJ path can tened.
crash cost, and total cost.
' The fqllowing ta~Je g~ves the normal length,

d
Unit4. p .
• ro1ect M,
anage,n
ent By PERT-CPM 4.327

Es= 18
Ls= 18

5 4
E, = O
L, = O
E4 = 15
L4 = 15

Flg.737
.
Normal project length
Crashing cost (Rs.) ' Overhead cost Total cost (Rs.)
(days)
@Rs. 6 per day
18
17 - 18 x6 108
1x2=2. 17 x6 104
16 100
~4 16x6
15 97
2x2+1x3=7 15 x6
14 /'
2x2+2x3=10 14x6 94
13 2x2+2x3+1x4=14 92
13~6
12 2x2+2x3+2x4=18 90
12 x6
11 2 X 2 + 2 X 3 + 3 X 4 = 22 88
11 X 6
10 2x2+2x3+3x4+1x6=28 10x6 88
9 2 X 2 + 2 X 3 +3 X 4 + 2 X 6 - 34 9x6 88
Si~ce there ar~ two parallel critical paths ~hen the pr~ject length is 9 days an;so th~ctivities are at their
crash time, the optimum length of the schedule 1s 9 days with total cost of Rs. 88. .,,~--:-'-: . ,,.·
EXAMINATION PROBLEMS '
1. A dircted graph has the lines (a, b),, (a, c); (c, a), (a, d) and (d, c).
Draw the graph and say from whicti node there is a path to each
other node and from which node there is no path.
[Dibrugarh (M. Stat) 76]
[Ans. There is no path from node b. The possible paths from other
nodes are:

2. Draw a network for the following project and number the events
according to Fulkerson's rule :
(,) A is the start event and K is the end event.
(it) J is the successor event to F.
(iii) c and o are successor events to 8.
(iV) D is the preceding event in G.
( V) E and F occur after C •
( vi) E precedes F. Fig. 7.38.
. C trains the occurrence of G and G precedes II.
(VII) res
(viii) II precedes J.
iX) F restrains the occurrence of II. .
( K succeds J. . . . . . [Bombay (B.Sc. Stat.) 74]
(X) b the nodes of the network such tha_t their ascending order indicates the direction of progress in the
[Hint. Num er . .
network.] (PERT) diagram from the following list of Activities.
3. . Draw network
 Operations R
4.328 esea
"c~

,.

A 8

Fig 7.39
Job Name lmmediat Predecessors
Job Name Immediate
Predecessors
a - I
m
k
k
b a
n k
C b
d C 0 d
p 0
8 b
f e q b
e r n
0:
h . C s I, n
. i C, f t s
J g, ~. I u p,q
k J V u
[Ans. the network is drawn as below.]
4. Draw network for jobbing produ~tion and ind{cate the critical path from the foUwing :

Fig. 7.40.
Activity Description Time (Weeks)
A Market research Preceded by
15 -
B Make dr,-wings
15 -
C Decide production policy
3 A
D Prepare sales program
5 A
E Prepare operation sheets
8 B,C
F Buy materials
G Plan labour force
12 B,C
Make tools
1 E
H
I Schedule production 14 E
J Produce product 3
D,G
14 F, H, I
Unit 4 : Project Man
agement B
(Ans. The y PERT-CpM
network is d
rawn as sh 4.329
, own below.]

J
14

5. The f0 II .
owing infon Fig. 7.41
activities and numbnati. onb IS
• •
known for a proiect. Draw the network
er in racket denote aCIVI r ·ty r1mes. and find the critical path. Capital lette
rs denote
This must be completed
Before this can start This must be completed
A(30) Before this can start
8(7) C F(7) H
B D F I
B G F •. L
C(10) K G(21) I .
D F
C L
G H(7) J(15)
0(14)
E l(12) J
E(10)
F K(30) L(15)

(Ans. The network is drawn as below. By [C.A. May 80]

enumerating the durations of the various
the critical (longest) path is C path s it can be ascertained that
D E F Jwith 98 days duration.]

10 7
E F
I
DUMMY t
I

10 G
C 21

~DUMMY
I
K
7 30

Fig. 7.42
.
ects state the problem \n terms of even
ts and draw the
•

For the following two pro . '. event oriented networks.

6. cting an exam
1

1nat1
(a) Con u d on
. .
oesi gn questionnaire 7 days
ti pape 2 days
Print ques on . usr centres
Distribute to vano . 4 days
e'tionna1re 1 days
Answer qu ooks at main office
colle ct answer b 4 days
[Bombay (B.S c. Stat.) 75]
 Operations Res
4.330 earch
(b) Holding a conference 6days
By mail ask members for suitable dates 2days
Inform date to members 3 days
Prepare agenda 7days
Send agenda and relevent materials to members by mail
2days
Arrange conferece room 1 days
Arrange refreshments
(Ans. (a) Examination (b) Conference .
1. Questionnaire prepared. 1. Member's choice of date obtained.
2. Question paper printed. 2. Members informed of date.
3. Papers distributed to various centres. 3. Agenda prepared.
4. Questionnaire answered. 4. Agenda and relevant materials sent by mail to members.
5. Answer-books collected at main office. 5. Conference room arranged.
6. Refreshments arranged.

_Fig. 7.43
Event 7 is the actual start of the conference Activity 4-7 might represent the time for arrival of members from distant
places.

--- --- --

Fig. 7.44
7. Draw the network diagram from the following activities and find critical path and total float of activities.
Job Job-time (days) Immediate Predecessors
"
A 13 -
B 8 A
C 10 B
D 9 C
E 11 --
B
F. 10 E
G 8 D,F
H 6 E
I 7 H
J 14 G, I
K 18
J
[C.A. (May) 83]
{Ans. Critical path 1 2 3 5 6 8 9 1Oas shown in the following network diagram.
 4.331
Unit 4 : Project Management By PERT-CPM

Es =32
Ls =32

E4 = 31
L4 =33
Fig. 7.45. K
G H I J
1'CtiVity A R C D E F 0 0
0 0 2
Total Float (Tt - Tc) 0 0 0 0 5 0
• • formation plot a network.
8. .
Define , event and dummy constra nits in a PERT networ k. Using the following in
an activity
Detennine the critical path and compute slack for all events •
-- Activity duration
Activity Activity duration Activity
(in weeks)
(in weeks)
7
0-1 5 3-5
11
0-2 10 3-6
4-6 8
1-2 4
9
1-3 8 4-7
5-7 9
1-4 3
• 5-6 . 4
2-3 6
6-7 1
2-5 8
is reduced, what will be the new critical path ?
If the duration of activity 5-6 is increased to 6 and activity 3-6
[C.A. (Man. Inf. Con. Sys.) May 79]
led within a resource limit of 12 men. All the men are
The project represented in the table below is to be schedu lar day, a man is idle but still draws pay. Each job
capable of working on any of the jobs.If not assigne d on a particu
the three possible crew sizes listed in the table. No. in
must be assigned to a crew of men corresponding to one of
fixed for a job until it is finished. Job duration equals
between assignement~ may be made, the crew size must remain le the project so as to minimize idl~ man-days over
man-days divided by crew size for any crew sfze chosen . Schedu
its active span.
Resource Crew size (men)
Job i-j requirem ent Minimum Normal Maximum
(Man-dayJ)
32 2 4 8
1-2
48 4 6 8
1-3
40 4 5 8
2-3
2-4 12 2 3 4
30 3 5 6
4-5
54 3 6 9
3-5
'--
[Madurai B.E. (Mech ) 76· D lh 1• (M •B•A.) 75]
3 5.] • • e
[Ans. Critical path is to do jobs 1 2
a sea side resort The d r f .
A contractor has received order for constructing a cottage on • e ivery o materials must be
10. planned and the comple te job finished in 13 weeks.
r of days):
The work involves the following (the numbers represent numbe
and cemen t (8, 10, 14) Wiring for_electricals (16, 20, 26)
Buying brickets
Constr ucting roof (8, 8, 1O)
Roof titles (20, 24'. 30) 4 1 6 12, 18)
1 : _) Plaster ing (12,
Repairing foundation (12,
La~~s caping (4, 4, 6)
Erecting s~ell structure1~f building (18, 20, 24)
Painting and cleanin g (10, 12, 14)
Laying drains (12, 14, )
 4.332
___/perations Research I
l-
( o 24 30) •Laying pathway (4, 4~) (4 4 4) · ·
II' doors and fittings • •
Plumbing 2 , • lnSta mg d d indicating activity with a
Flooring (8, 10, 12) . d ark ~n It the critical path circling the no es i~ombay (Dip. Oper. Man~)~j

!:S
Construct a logical PEAT _d,ag~ai a~h ~ompletion target be met ? . •. arketing project that has a sales
What is total critical path time . an ean has an opportunity to partic1~ate m received Friday aftemo~:m. Both
11.
The chairman of ABC tons~lg~gc~~Fete~ within s weeks. The letter and completed the appropriate tirne
price of Rs. 90,000 ~u ;;us rtment and the Cost Accountant came on a u er at 8.30 A.M. on Monday ( start of the
the Head of Marketmg epa • s· ce the chairman needs an answ . n a week basis. An answer at
and cost for you based upon ::'!!~~~b::; d~7ermlne the profitability '!' th ~ pr~Je~!,,':',,!thin the 8 weeks demanded by

::rtt~: ~~ii;~!;:E:~ ~~:11,,:~r~~dT~;i~

the customer.
basis. What answer
0
~,:~M~i ·~:syr
e , shouId the chairman give the customer on on a
hEi1~~ ~~~;:,/: :,a:::i: 1
~~~;ts ,~-;~:~
below:
Normal Crash
Event Preceding Event Time Cost T ime Cost )
(in weeks) (in weeks) (in Rs.
(in Rs.) 1 13,000
4 1 2 R,000 1 13,000
2 1 3 7,000
3 1 6 11,000 5 13•500
4 2 4 6,000 3 10,000
3 2 2 9,000 1 10,000
5 2 7 8,500 6 11,500
'5 4 4 10,500 S 16•000
5 3 3 5,000 2 7 ,00'J
•• [Delhi (Dip. MKT. & Sales Man.) 77]
12. (a) Discuss the benefits of Netw~rk Techniques in Project Plan~ing and. Control. . .
(b) The basic cost-time data for jobs in a project are as given below: . f.

Job Normal Time · Crash Time Cost of Crashing

' • Days Cost (Rs.) ' Days Cost (Rs.) per day
A 3 140 2· 210 70
B 6 215 5 275 60
C 2 160, 1 240 80
.. D 4 130 3 180 50
E 2 170 - 1. 250 80
F 7 165 4 285 40
G 4 210 3 290 80
H 3 110 2 160 50
Total 1,500 1 890
The activity (Job) dependencies are ~s below:
(/) A, B, Care starting activities. . . · ·
(i1) Activities D, E and F can start when once A Is completed.
(Ii,, Activity G can start after Band Dare completed.
(/11) Activity H can start after C and E are completed.
(1,') • Activities G, F and Hare the final activities.
(1) Draw the network and indicate the critical path.
(2) What is the total time required to complete the project? (based on normal times). ,,
(3) If the project fs to be completed in a days, what is the minimum cost to be incurred ? Indicate this cheapest cost
13. schedule. • [Bombay (Dip. Bus. Man.) 74; D.I.M. 75]
Explain network problems. Describe the method of drawing network diagrams. _ [Kerala (Stat.) 83]

7.10 PROJECT EVA~U_ATION AND REVIEW TECHNIQUE (PERT) •

In the network analysis discussed so far, it is implicitly assumed that the time values are deterministic or
variations in time are insignificant. This assumption is valid in regular jobs such as maintenance of a machine,
etc. construction of a building or road, planning for production, as these are done from time to time and various
Unit 4 : Project Management B p
Y ERT-CPM · 4.333
activities c~ld be timed .
• • •• very Well Howev . b f a new machme,
vanous achvtt1es are based • • er, m research projects or design of a gear ox O ·
changing repidly. Time val on Judgen:ient. A r~liable time estimate is difficult to get because the techno1ogy is
The mam • ues are sub1ect t h . .
objective in h .i o c ance vanatJons. . within
specifi~d ~ate. If yes, whatt: a~alys~s through PERT is to find out the completion for a parakucul~tev;;~unt the
uncertainties. In this a r e e c a_nces of completing the job ? The PERT approach t ~s .10. 0 Lue the
pessimistic value andPfhoach, three. time values are associated with each activity : the optim,st v~certainty
associated with that activ'te most. likely value. These three time values provide a measure O u
1
1 y.
Def• 1• The optim • f • •
that everyting goes fl · l d It assumes
rs icll tim~ 1~ the shortest possible time in which the activity can be ,ms ,e •
Df very We • This is denoted by t0 • .
e • 2• The most likelv ,;_ • h .. .. ld take This assumes
normal dela s 1 ..., nme t e es1t1mate of the normal time the activity wou .• ·n that time
period, then {h~ lf a gr~ph ,s. plot~ed m the time of completion and the frequencey of ~omplet,on I t
D t
moS likely tine will represent the highest frequency of occurrence. This ,s denoted by m·
A • ef•.3•. !he p~ssimistic time represents the longest time the activity could take if everything g~es 7rong.
s in °h~timi stic es~mz:1te, this value may be such that only one in hundred or one in twenty will take time onger
than t is value. T?is is denoted by tp.
These three llme values are shown in Fig. 7.46.
In ord~r to obtain these values, one could use time >-
val~es available for similar jobs, but most of the time the (.)
z
estunator may not be so fortunate to have this data. w
:,
Secondly, values are the functions of manpower machines 0
w
an~ ~upporting facility. A better approach woula' be to seek a:
u.
opm10n of 'experts in the field' keeping in view the
resources available.
This estimate does not take into account such natural QL...__....i...._ _....__ _ _~-----=-~

c~tastrophes as fire, etc. OPTIMISTIC MOST

LIKELY
PESSIMISTIC

In PERT calculation, all values are used to obtain the Fig. 7.46. Time distribution curve.
per ce~t expected value.
Def. 4. Expected time is the average time an activity will take if it were to be repeated on large number of
times and is based on the assumption that tha activity time follows Beta distribution*. This is given by the
for.mu/a:
t, = (t0 + 4tm + tp)/6
Def. S. The variance for the activity is given by the formula :
cr2 =[(tp - t0 )!6]2
where 10 is the optimistic time, tP is the pessimistic time, tm is the most-likely time, t, is the expected time and cl
is the variance.
PERT computations are essentially the same ~s used earlier.
EXAMINATION QUESTION
1. Explain the following terms in PE~T : .. . . . . . .
(t) Optimistic time, (it) Normal time, (111) Pess1m1st1c time, (tV, Expected time, (V) Variance in relation to activities.
[Shivaji 85; C.A. (May) 82]

- • The Beta distri~uti_on ~as chosen possibily because it is a

(1) Unimodal d1stnbution .
2) Finite non-negative end points . . .
( N n-symmetrtic or symmetric Beta density Is given by
(3) O _ 1 ,f _ X p {0 < X < 1 .
f(x,a,P)-B(cx+1,P+1) _< 1 ), cx>-1,P>-1"
values between limits a and b, a new vanable Y= (x- a)l(b- a) can be defined d th"
If x takes on Beta function is defined by an Is takes values
zero and one. rcx rp
between . B(cx, P) = r(cx + {l) , where rcx is wgamma a".
 4.334 Operations Res
'

earch
The main differnce is that instead of activity duration, expected time le for the acti~ity is _consi~ered. With
each node, variance is associated. Thus, the duration of the project is the mean expected time ~•~h vanance .
. Conside_r the network of Fig 7.20 again. Table 1.5 give three time estim~tes for each activity, the expected
value and the variance also. ••
Table 7.6.
t
Activity la Im Ip - · I = (In +41m +lp)/6 ci- = [( lp-111)16] 2
1-2 1.0 2.00 3.0 2 I ... 4/36
1-3 1.5 2.00
. 2.5
. '
2
'
1/36
1-4 1.5 2.75 3.5 3 4/36
2-5 3.0 3.00 7.0 4 16/36
3-6 4.0 4.50 8.0 5 16/36
3-7 6.0 8.25 9.0 8 9/36 .
4-7 3.0 3.50 7.0 4 16/36
'
5-8 2.0 2.00 2.0 2 0
.-
6-8 2.0 •'
4.00 6.0 4 16/36
7-9 2.0 4.50 8.0 I \
5 36/36
8-9 2.0 3.00 4.0 3 4/36
9-10 2.5 4.25 4.5 · 4 4/36
• I

' Once, expected values have been calculated, .these are used in finding the critical path. In this particular
example, three estimates are so chosen that mean values are same as before and hence critical path ·calculations
are same as ~efore. However, the interpretation of the critical path is now different. In this case, the expected
duration of job taken time less than 9 days or n1ore than 19 days too. Then, meaning of the expected duration is
that-if the same job is performed again under similar conditions, the average duration will be 19 days. If .the
job takes 1 days, then there is· probability value a~sociated with it which can be calculated under some
assumptions. Since duration of each activity is a random variable, the duration of a path which consists of a set
of activities will also be a random variable. To calculate the exact distribution of _the duration of a path will be
difficult and for management decisions it is enough to know the mean and the variance. Mean value has been
calculated using the method discussed earlier. The same approach is used to find the variance.
Rules for finding variance of events.
. (r) Variance for the initial event is zero. Set V1 = 0. ,
•(ii) Yj, the variance for succeeding event j in question is obtained by adding activities variance to the
. . ,,, 2 . .
. variance of prede.,cesser event except at merge points, i.e. Vi= V; + aii •
(iii) At merge points, the variance is computed along the longest path. In the case of two paths having the
same length, the larger of the two-variance is chose~ as the variance for that event.
, V5 =V2 + a 2,.s =20/36 V9 =V1 + a1,9 = 46/36*
2
V1 =0 ' 2

V2 = 4/36 •t V6 =¼·+ af6 =i7/36 V10 =V9 +·01 10 =50/36*

V1 =V3 + at1 = 13/36*
\

V3 =-1/36.
V4 ;::'4/36 Vs= V6 + ~.s =33/36*
ltis iq;iportant to note that variances cannot be added as easily as is done, unless two random variables are
1ilM11endent of each other. In this.case, it is assumed that two activities are independent of each other and hence
~J.: :v@ripces can be added. • . • . . •
~x.~t ed duration of the project is 19 days and the variance •of this path is 50/36. If the ex~ct
06 • distribqton Qf the path is known, it would have been easy to find out the probability of completing
·vefi ~e. Since the varin_ace of the :,ath is known, the Chebychev inequality** could be used to
g the longest path. ,
as follows: Prob (I x- µ I > kcr] 1/~.
variable exceeding the mean ~alue by k standard deviation Is less than 1Ji. The
2
1/(1.Slc) . , • . •

I
 t w:;;os

lJ/14 : Project M
anagement 8 PE 4.335
rt an · Y RT-CPM
·ou}d estimate of probab·1·1 .
t be assumed C tty for a given duflltton (if unimodality
mequal
us ''ty Will give' beamp•N~1'de1 inodificsuon
• to Chebychev 's
of~ central limit th tter estimate). PERT~sers, however, have
time for each ev eorem to claim that probability distribution
,
1
str~ng assumption
e! ~ould be considered as nonnal. This .is a
easi1y understood by hich greatly simplifies calculations and
nonnality, the prob b~~st of the users 0{ PERT. Assuming the
c~rtain date can be e~ ; tty of the project b~ing completed by a
. Prob [p . a uated easily .
ro1ect duration 20 d 1·
ays :::: ?
Prob [D s 20] == Prob [Q.:J
S 20 - µ-J . 17 18 19 20
DURATION DAYS_,.
21
where µ is the m o a •
deviation. ean of the distribution and a is the standard Fig. 7.47
Now, the value (D
- µ)/a is a normalized value and usually written as Z. Then, .
• •. Prob [U ~]==Prob [zs _!_]=Prob [; s 0.85] = o.~O (from N;rmal Distribution Table).
Th . 6) 7.01;6 r-
e p~obability of finishing
statement ts • If th. . • b . the job in less· than or equal to 20 days is 0.80. The physical meaning of th is
j b •u h • . h h'
i°
onger than 20 days.
is JO as done hundred times under same conditions then there will be 80 occasions
1 w en t ts
Wt • ave taken 20 days or less to complete it. In other words onl y 20 times, the job would have taken time
'
One of the main advantages of PERT approach to the management. of a large scale project is in binding for
contractual dates to finish the project. A designer would like to know the duration of the project that 'wilJ have
95% chances of being completed. Let T.v b~ the scheduled
duration such that

Prob [t ST,]: 0.95, Prob['~µ S 7'.,; ~] = 0.95, Prob[Z:, ,7'.•; µ] •0.95

From nonnal table Z.95 = J.64. Therefore,

• f
(Ts-µ)/O= 1.64, and T.v= 19~ 7·~7 X 1.64=20.90=21 days.
That is, if the designer says that he will need 21 days to complete this project, he has better than 95% chances of '
meeting the committed date. In larger projects, failure to meet the ·commited date could result in heavy penalties
and great loss of good-will.

•7.10-1 illustrative Examples On PERT

Example 13. For the project represented by the network diagram, find the earliest and latest times to reach
each node, given the following data :

..

Fig. 7.48
a uces
------

'v

Operations Research
4.336
C .D E F • G H I J K
Task A B. 8 5 3 5
a 2 4 6 ·, 6
Least time 5 6 9 7 11 1
Greatesttimo : 10 12 7 10 15 • 2
1
1 6 5 8
3
11 3 7 9 •\ ' 9
Most likely time 5 7 , , [Shlva jl 85; Delhi (Math
76
41m + fp)/6 as follows: •l 1
Solution. First calculate the expected time t, bj the formaula t, ""(I,,+ I J_ K
Task ' : A 8 C • D E F Gt H
to : 4 5 8 2 . 4 6 8\ 5 3 5 6
7 10 15 16 '\ 9 7 11 13 .
fp : 8 1O 12
3 7 9 12 l 6 5 8 9
• fm : 5 7 11
10.7 10.7 7 9.5 12 .\ _6.3 5 8 9.1
fs : 5.3 • 7.2
~mg the sun of the expected times for
tjow, the earliest expected times E; for each node are obtained by
to a node i ; the maximum of E; is selected.
aH the activities leading to node i, when more than one activity leads ''

• ,·~
Therfore,
E1 = 0, E2 = 0 + 5.3 =_5}, E3 = 0 + 3.5 = 3.5, E4 =
max [~J 7.2, O-i-,~0.7)-: 12 "i,
. E5 = max [12.5 + 7.0, .:,.:: ·.5] = 19.5, ·E6 = 19.5+ 6.3 = 25.8,
l
E1 =.max [19.5 + 5, 3.5 +12] = 24.5, Es= 25.8 +9.1 34.9,. 9
E = 24.5 \.~.O 32.5. . =
1$f node as equal to E;. Now moving •
To find the latest expected times to s~art with the latest time Ti for the
y link to have
backwards for each path, subtracting the expected _time 'te' for each activit
•
lg:.... 32.5, Lg= 34.9, L, = 32.5 - 8 = 24.5, 4, = 34.9 - 9.1 = 25.8,
Ls= min [25.8 - 6.3, 24.5 - 5] = 19.5, L4 = 19.5 - 7 = 12.5
· Lj = min [19.5 - 9.5, 24.5 - 12] • 10, Li= 12.5 - 7.2 = 5.3,
L1 = min [5.3 - 5.3, 12.~ - 10.7, 10- 3.5] = 0.
£; L; Stack
Node tt'

2 5.3 g/
, §i, 0
3.5 ~-5 10.0 6.5
3
7.2 12.5 12:5 0
4
19.5 '!9.5 0
5
6
7.0
6.3 25.8 25.8 0 -
5.0 24.5 24.5 • 0
7
9.1 34.~ "' 34.9 0
8
8.0 32.5 32.5 0
9
Example 14. For the project

' J
Fig. 7.49
•
Task : A B C D E F G H I
5 8· 2 4 6 . 8 J K ,·
Leasttlrne : 4
5 6
. 10 5 3
Greatest time . 8 10 12 7 15
-
16
9 7 11 13
Abstlkely time : •
Find . • s 7 . 11 3 •1 9 12 6 5 8 3
ne~or k. [C.A. (Nov.) 771
in the
i earliest and lat~st ~pect ed tim~s to each event and a!so critical path.
.I

i
Unit4 • p • 4.337
• roiect Management By PERT-CPM
Solution.
Expected Time Computation•
T~k Expected time
Lcat time Greatest time Most likely time
·ca+ b + 4m)/6
a b m
A 5 5½ 1:,\ b~
4 8
B
7 71/s
5 10
C 11 10%
8 12
D 3½
2 7 3
E 7
4 10 7
F 9½
6 15 9
G 12
8 16 12
H 6½
5 9 6
3 5 5
J 7
5 8 8
K 11
6 13 9 91/e

..
51
1
E7 =293
1
L1 = 293

1
E4 =3½ Ea=382
1
L. = 13½ la= 382
Fig. 7.50

Task Expected Start t...J Finish

time (te) Total float
Earliest Latest Earliest Latest
A 51,'J (t
I
0. 5½ 5½ 0
B 7¼ 0 ' 8%,. 71/s 1) 8%
' I
C
D
10%
3½
5½
0
. 5½ /
10'
16
3½
16
131h
0
16'-
' I..,
E 7 16 j§' 23 -i- 23 0 .
I
F 91/i 31J 13½, 13 ·2_3 10
G 12 311.z . 18½ 15½ 301h 15
H 6½ 23 23 291,'J 29½ 0
I 5 23 25½ 28 30½ 2½
J 8 1~8 30½· 36 38½ 21h
K 9¼ 2911'3 29½ 31½ 381h 0
Critical path is A C E H K.

WI ¥
 Operations~
ese
4.338 Q~c~

tics. . istic Time Most likely Time

Example 15. A project has the following characteriS
. . Most Pess1m ( )
• . .
Activity • Most Opt1m1st1c Time (b) m
(a) 1.5
5
1-2 1 3 2
2-3 1 5 3
2-4 1 5 4
3-5 3 3
4
4-5 2 5
7
4-6 3 5
6
5-7 4 B 7
6-7 6 4
6 6
7-8 2 8
5 2
8-10 1 ; 3
1 3
9- 0 . .. d • fl each event. Find the project du,-,a,·ion QJ
Construct a PERT network. Fmd crlllcal path an vanance or
95% probability. .
•
• vanances computed by the followmg formaulae :
-~ Acttv1ty
Soluuon.
·
· · expected times and the1r are
-
2
-a /b •

. a+h+4m
Expected time, le= 6 , v.- 6
Ir , V -
Activity (a) (b) (4m)
6 2 4/9
1-2 1 5 1/9
8 2
2-3 1 3 4/9
5 • 12 3
2-4 1 1/9
3-5 3 5 16 4
12 ' 3 1/9
4-5 2· 4
7 20 .. 5 4/9
4-6 3
4 6 20 5 1/9
5-7
6 28 7 1/9
6-7 8
7-8 2 6 16 4 4/9
7-9 5 8 24 6½· 1/4
8-10 1 3 8 2 1/9
9-10 3 7 20 5 4/9
The network is constructed as below. the earliest and latest times of each ev~nt have been computed and
indicated on.the network. With the help of latest times, the longest path 2 4 6 7 9 IO can be.
traced.
E3 =4 E5 =8 E8 = 19
la=B ls= 12 Ls= 26 1/6

E10 = 281/5
L10 =:281/5

E6 = 10 E9 = 231/5
Ls= 10 Lg= 23 1/5
Fig. 7.51

\ '

\
I
Unit 4. p .
• roJect M
1
Exam 1
anagement By PERT-CPM
4.339
·~
P e 16 Ob • • PERT network. [C.A. May 82]
• tam the crit~cal path and project duration for the following .

Fig. 7.52
Solution. E's amd L's for all events are computed below oq the network diagram by -forward and
backward passes respectively. Thi~ provides us the project duration as 51 days ..
The network analysis table is construced as below.

10

E10 = 41
L10 = 41

E9 =36
Lg_= 43
-------

Fig. 7.53.
 Operations Research
4.340
Finish Total float ...
Duration Stmt
Activity
Earliest Latest
Earliest Latest
10 10 0
10 0 0
1-2 12 12 0
2 10 10
2-3 16 32 16
6 10 26
2-5 24 24 0
12 12 12
3-4 36 15
9 12 27 21
3-7 32 0
8 24 24 32
4-5 36 7
5 24 31 29
4-6 2
10 24 26 34 36
4-8 0
32 32 36 36
5-8 4
36 29 36 7
6--7 0 29
36 36 43 7
7-9 7 29
36 41 41 0
8-10 5 36
43 44 51 7
9-10 8 36
41 41 51 • 51 0
10-11 10
The critical path is traced along zero total float activities as 1 2 3 4 5 8 10 11 and is
shown by double li~es on the above network.
Example 17. In the network shown below, the three time estimates for the activiries are indicated. Number
the events ac~ording to f ulk_erson 's rule and cal~ulate the variance and expected time for each uuivity.
[Bombay B.Sc. (Stat) 75]

Solution. Using the fonnula le= (10 + 4tm + tp)/6, variance er= (Ip -10 /116, compute the following table.

Activity lo Im Ip It a2
1-2 3 6 10 6.2 1.36
1-3 6 -j 12 • 7.7 1.00
1-4 7 9 12 9.2 0.69
2-3 0 0 0 0.0 0.00
2-5• 8 12 17 12.2 2.25
3-6 .10 12 15 I 12.2 0.69
4-7 8 ,13 19 \
I
13.2 3.36
5-8 12
8
14
9
15 ./ 13.9 0.25
10 9.0
6-5
6-9
8-9
13
4
16
7
19
10
(
I
16.0
7.0
0.11
1.00
1.00
7-10 10 13 17 \ 13.2 1.36
9-11 6 8 12 8.4 1.00
10-11 10 12 14 12.p,' 0.66
-
..

. '
13-16-19
•

8-13-18 10-13-17

Fig. 7.54.
Unit 4. p .
• ro1ect M
E anagement By PERT-CPM
xarnpte 18 . . 4.341
• A pro1ect is re
presented by the network shown below and has
the Jo llowin
• d t ·
g aa•

--+0 --NO DE

Fig. 7.55
Task
Leat time A B C o· E F G H
5 18 26 16 6
Greatest time 15 7 7 3
10 22 40 20 12 12
Most likely time 25 9 5
8 20 33 18 20 9 10
Dete rmin e the following : 8 4

(i) expe cted task time and their variance

(ii) the earliest and latest expected times to 1
(Marathwada (Appl. Maths) 82]
reach each node, •
(iii) the critical path, and (Delhi (M.Sc. Maths) 83)
(iv) the probability of node occuring at
the proposed comptetion date if the original
completing the project is 41.5 weeks contract time of
. • [Shlvajl 85; Andhra (Prod. Coun .) 78]
Solution. (i) Proceeding as in above example
obtain the following tabl e:
Activity t" tp tm
1-2 5 . 10
tr a2
8 7.8 0.69
1-3 18 22 20 20.0 0.44
1-4· 26 40 33 33.0 5.43
2-5 16 20 18 18.0 0.44
2-6 15 25 20 20.0 2.78
3-6 6 12 9 9.0 1.00
4-7 7 12· 10 9.8 0.69
5-7 7 9 8 8.0 0.11
6-7 3 5 4 4.0 0.11
(ii) Proc eedi ng exactly as in above example, find
carhest times m usual notations.
Ei= 0,£ =0+ 7.8, £ 3 =0+ 20= 20,£ 4=0
2 +33 =33 , Es= ?.8+ 18= 25.8 ,
E6=max [7 •8 + ' 20 20 + 9] = 29, £ 7 = max [33 + 9.8, 25.8 + 8, 29 +4] =
. . 42.8.
k ds c·alculate the latest times as before,
Mov ing bac w ar 'L _ '-
L-,= 42 •8 I 6- 42•8 _ 4 = 38.8, Ls= •
42.8 - 8 = 34.3, L4 = 42.8 - 9.8 = 33, L = 38.8
3 - 9 = 29.8
.
= min [34 _8 _ 18, 38.8 -20 ] = 16.8,. Li =1min .[16. •
8,.- 7.8, 29.8 -20 , 33 - 33] = o.
Li . . 1 path calculate slack time by taking difference between
( iii) To fi~d th~ cnUcCa lcul~tions are given in the the earliest expected times
wable umes. a following table and critical path is shown by
11 double line
f; d late st a 0
the follo wing figure.
.

m
 4.342 Stack Var. o;
Li 9.0 0.69
Node (i) ,,, E1
16.8 9.8 0.44
7.8 29.8
2 7.8 0.0 5.42
20.0
3 20.0 33.0 9.0 • 1.13
33.0
33.0 34.8 _f4a
-
4 25.8 . 9.8
5 18.0 38.8 6.12
29.0 0.0
6 9.0 42.8
9.8 42.8
7
E3 ?20
L 3 = 29.8

E 7 = 42.8
L7 = 42.8

E1 = 0
L1 = 0

• Fig. 7.56
th t in standani
sche dule d time of comp letin g the proje ct is i,l.~ weeks. Ther efor e, e diS ance
(iv) The
expected times E;, is given by
devi ation s, that sche dule time from earli est
D ST; - E; _ 41.5 - 42.8 =_ 0.5 2
;- - ~{6.12)
\

whe re ST; deno tes the sche dule time. , ch is the area under
] = J - 0.70 = 0.30 (from Norm al Tabl e) _whi
- Ther efor~ . P(Z - 0.52 ) = J - P[Z 0.52
ates at x 0, and x =0.52. = .
the stan dard nortn aJ curv e boun ded by ordin .. 100 time s unde r the sam e cond ition s, there will be
is perfo nned
From this it is conc lude d that if the proje ct
weeks or less to com plete it.
30 chan ces whe n this job wou ld take 41.5
ofa network with their estimates.
Exa mpl e 19. The follo wing table lists'the jobs
Duration ldavs)
Job (I-JJ Pess imist ic ( fo)
Optimistic (fo) Most likely (fm)
6 15
1-2 3
. 2 5 14
1-6 30
6 12 . .
.. 2-3
2 5 8
2-4
11 17
·3-5 ,5
3 6 15
4-5
3 9 27
6-7
1 4 7
fr8
7-8 4 19 28 -
_leng~h and variance ofthe criti cal path , and
Dra w _the proj ect n~tworl,, (ii) ~a_lculate the plete d in 41 days.
(111) wha t 1s the appr oxim ate pro/,ab1l1
ty that the Jobs on the critical path will be com 75
bay (D.C .M.) 85, B.Sc. (Stat.) ]
[~oc hlne (O.R. & Comp. Appl.) 86; Bom
/6 and ci = (IP -1") /36, calc ulate le and 0
2 2_
Solu tion . Usin g the form ula le= (10 + 4tm +lp) 7-8 •
2-4 3-5 4-5 6-7 5-8
• Activity ~-72 1-6 2-3
18
4 11 7 11 4
t. • 6 14 •
4 4 16 1 16 ·
a2 . ,4 16 1 4
 4.343
Unit 4 : Project Managemen t By PERT-CPM
The earliest expected times are cal culae t d as
E _0 E
1
-
7 7 +
' 2 =O + = • £3 =7 + 14 =21, E4'= 7 5 = 12 E:-; = max [21 +1 I, 12 + 1==
7 32•

£ 6 =O + = • E7 =6 + 11 .:.. 17, Es =max [32 + 4, 17 18] =36. - •

6 6
The latest expected times are calculated as . .
21
Lg _ ~• L, = -
3 36 18 7,
.:_ 18, 4 = 18 - 11 = L5 =36 - 4 =32, L4 =32 - 7 = 25, =
== 32 - I I '
Li- mm [21 - 14, 25 -5] =7, Li= min [7 - 7 ' 7 -6] =0 • •
.
ted
.
times
F th • • .
d l or e cnti~al ~ath, calculate the slack time by taking the difference between the earliest expec
t
an ateS allowable.times. Critical path is shown by double line in the following diagram.
E2 = 7 \o E3 = 21 •
=
L2 7 -=
L3 21

E5 =32
L5 =32

E7 = 17
L 7 = 18
Fig. 7.57

The optimum length of the critical path is 36 days and variance of the critical path is 25.
Now the scheduled time of completing the jobs is given 41 days. Therefore, the distance m standard
deviations, that ~chedule time from earliest expected time is given by -
St; - E; 41 - 36 5
D; = = = 5 = l.
where ST; denotes the scheduled time.
=
Hence P(Z D;) 0.84, which 1s area under standard normal curve bounded by the ordinates
I

x=Oandx = 1.
This concludes that only 16 times the job would take time longer than 41 days.
Example 20. Consider the network s~own in the figure given below. - The estimates of t"' tm and tp are
shown in this order for each of the activities on the top of the arcs denoting the respective activities. Find the
probability of completing the project in 25 days.
3,5, 7

1; 2, 9

Start

Fig. 7.58
--.....
I

Operations R
4.344 .esearch
• . d the variance a2, these values are °blain~
Solution. Using the fonnula 'for expected activity duration te an
1
as shown below : t, a2
lo 1,,, Ip
Activity 9 3 1.78
2 4 1.00
1-2 1 7
1 4 5 2.78
2-3 12
2 4 0 0 0.00
2-4
O 4 3 0.11
3-4 O
3-5 2 3 9 2.78
16
3-7 6 8
8 6 ~«
4-6 4 6 0.45
3 5 7 5
4-6
1 % 1 0.03
5-6 ½
·7 15 8 2.78
S-7 5
5 13 6 2.78
6-7 3
ble time m~n g use of all ~e's. For criticai
• Now calculate tf:ie earliest expected times and the latest ~llowa
en the earliest expected times and latest
path, detennine the slack time by computing the difference betwe •
ing figure..
allowable times. The critical path is shown by double line in the follow l path is 7. 78.
the varian ce of the critica
The optimum length of critical path is 22 days and
fore,
But, the scheduled time of completing all the activities is 25 day·s. There
E· 25 - 22 ed I •
• D· = ST.·' - ' = _,__ •= l •08 • where ST; ·
denotes sch u e tune.
' V[Var.(1)]'
From standard nonnal tables, P(Z D;) = 0.86

. 3
1
E1 =0
'-1 =0

• E3 =7
L3 =8
Fig. 7.59
.
'
pessimistic times are respectively shown
Example 21. In· the following table optimistic, most-likely andcritica
the l path by constructing a network.
against each connected activityfrom 10 to 100 in a project. FindCalcu late the probability offinishing the project
The scheduled completion time for the project is 48 days. ion of+ J.25). • .
l 4eviat
within this time (given that 89.5% probability corresponds to a norma
.
Activftv Times Activity Times
1~20 4,8, 12 ' 20-30 1, 4, 7
8, 12,16 30-50 3, 5, 7
2()-4() r
40-50 - 0, 0, 0 40-60 3, 6,. 9
-- • 50-S0 4, 8, 6
- 50-70 \ 3~ 6, 9
60-100 4, 6, 8 70-90 4, 8, 12
· SM0 t
2, 5, 8 ./" 90-100
.,
4, 10, 16
7
(Calcutta (Appl. Math~) &J
Unit4: Project Manag
ement By PERT-CPM
Solution. The ex 4.345
- pected activity ti• ' I
tne tt and the varia cr1- . .
Activity
to nee are calculated as given in the followm g ta6re.
'
10-20 1,,.
20-40
4
8
Ip I,= (lo+ 41111 +tp)/6 c/- - [(tp- fo)/6]2 ',.
8. 12 ,

40-so
50-70
0
12
0
16
0
8
12 ~:~··:
-::::e>,.
(r I (
3 0 0.00 '
60-100 6 9
(
4 6
6 LlN. I
(
80-90 2 8 6 0.44 (
-20-30 5 8
1 5 1.00
30-50 4 7
3 4 1.00
40-60 5 7 0.44
' 3 ' 5
6
50-80 4
9 6 1.00
70-90 8 6 7 0.11
4 8
90-100 12 8 1.78
.. 4 10
.
16 10 4.00 .
To fin d the cntical path I
allowable time and then obtam
the differe nee between th ' ca culat~ the earbest expected time and the latest
wo•. Vanances for each node (i) are calculated by rules given after Table 1.5. This
infonnation is given in the~eltl
e ,o owing table. ·-
Node E1 T6
20
.L, TL Slack Ts
8 8 0
30 12
40
15 3 2.78
' 20 20 o· 3.56-'
50 20 20 0 3.22
60 36 36 10 4.56
70 26 26 0
80 29 29 2
90 34 34 0
100 44 44 0

90=34
=34

E1o=O
L1o=O

Fig. 7.60

The crjtical path is-shown in the following figure by double line.

But, the scheduled completion time for the project is 48 days, therefore
48-44
D; = =1.27.
P z < D·) =og9k\ the area under the standard nonnal curve b~unded by the ord'mates
-~"-"X -x....... :~~- , •
. . • . . · -
x=0an dx=l. 27.
that the expected times are normally distributed' find the probabi•ti·ry oif meeting the
.
Examp 'Assuming
• k
as given for the netwol • ,
schedu le date

iG fi
 t
!
!
Operations Res • t
earch f
4.346
'
.. Days .. Pessimistic (b)
Activity (i - J) Most likely (m)
Optimistic (ll) 14
5 ''
2 15
1-2 ' . •. 12
17
1-3. 9 .
• 14
5 8
2-4 5
.
'
2 ,. 12
3-4
I

6
4 5 6 20
17
3-5 8 project manager ca,,
the date on which the . (I.I.Sc. Dip.. Ind. I.Ian,
.date
comple11011is 30 days . Also find
Scheduled pro;•ect - . .

rianc e <12 are calculated Jn the followmg table

. 761
ab1llty of 0·90. . . .
complete the project with a prob the act1v1ty va -
Solution. The expected activity time'.fe- and , a2
(b)
t,
Activity (ll) (m)
6 4
1-2 2 5 14 ... 1
15 12
I 1-3 \ 9 12 4
17 13
2-4 5 14 1
8 5
. 3-4 .,
.2 5 ,
1
' 12 7
4-5 6 6
3-5 - 8 17 20 . 16 4
.
• •est expecte t1m~ a~d the·- Jatest allow
d .· able time and then find the
, e the• earh
- critical path, detennin
:For the
• The optim
• e b,Y•dou ble .r•~e.
. figur
_. wmg um length of the critical path
e betw
renccritic
diffeThe een theistwo. :foJlo . ·
al path shown in the .
J 2 4 5 or J 3 5 is 28 days and
the variance of the cnt1cal path ,_s .9or 5. . · .

...

E5 =28
L5 =28

. Fig. 7.61

pleting all the activities is iive n 30 days. Therefore, the distance in standard
But, the scheduled time of c~m cted time, is give n by
• deviation that schedule time from earliest expe .
' ST.- -E· •.

D; = V~r (i)J , where ST; denotes schedule time

• =30"9 rs - . .
- 28 or 30 - 28 ·- O65 ,or O.89 ,
nonnal tables].
HenceP(ZSD1)=0·71 or0.81 [fro~ standard
'1~1 I. UPDATING
time is not
updating or review of the netWork from time to
l,f'any benefits of network plann!ng may be lost if the remaining
the progress that has been made and its impact on
• carried out The updating imp l~ the study of ates of time on various activities, re-drawing of the r~maining
jobs or activities. 111is also includes revising estim es were the
two steps are necessary because original valu
netwoifc and calculation of cridcal paths. Latter