753

17 d. The probability ofnot fiDi.shing before week IS is the complement of the probability olr
-u t.ained in part c: 1 - .1459 = .8541.
I
I 15
I weeks
I Path I
I I
I I
I I
I
: wem aIH:
I
Weeks

16.0
d-e I

g-hi
Weeks

Turning to Appendix B, Table Bwith z = +1.00, you will find that the area under the curve
to the left of z is .8413. The compulalions are sulnrtw:ized in the following table. Note: If
the value of z etcceds +3.00. treat the probability of completion as being equal to I .000.
SIMULATION
17 - bplde4,... dmliell ~ol c-pldioa We have examined a method for computing the probability tllat a projecl would be completed
l a 1'101 Slalllln l6ftilliot i111WMb
Padl in a specified length of time.That discussion assumed that the paths of the project were inde-
17-10 : + 7.22
pendenr, that is, the same activities arc not on more than one jJath.lf :m activity were on more
aIK 1.00
0.97 than one path and it happened that the completion time for that activity far exceeded its et
pcctcd time, all paths that included that activity would be aftbctcd and:. hence, their times
d+l
11 - 16 = + 1.00 .am
1.00 would not be independent.Where activities arc on multiple paths. one must consider iftbe pre-
17 - l3..5 1.00
ceding approach can be used. For instance, if only a few activities arc on multiple paths, par-
gIH -- ~ +117
I.a! ticularly if the paths arc muclt shorter than the critical path, that approach may still be
ieasooable. Moreover, for purposes of illustration, as in the text problems and examples, the
P(finish by week 17) = P(paih a-b-c fini~h) X P(palh d-e-ffinish) X P(path g-h-i finish) paths are treated as being independent when. in fact, they may not be.
1.00 X .8413 X 1.00 = .8413 In practice, when dtpendenJ cnses occur, project plannerS bften use simulation. It amounts
to a form ofrepeated sampling wherein many passes are made through the projec1 network. In
c. For a specified time of 15 weeks. the z values are
each pass, a randomly selected value for each activity time is made based on the characteris-
15 - EqeciM pd tbrltioa ~of~ tics of the activity's probability distribution (e.g., its mean, stnbdanl deviation, and distribution
z~
hill l'llb.an!WriaWa iltSw.tb type). After each pass, the expected projec1 duration is determined by adding the times along
eacb path and designating the time of the longest path as the project duration. After a l&~ge
aIre ~s +S.IS 1.00 number of such passes (e.g., several hundred), there is enough information to prepare a fre-
0.97
quency distribution of the project dw-ation times. Planners cah use this distribution to make a
d-1
15 - 16.00 . - too .1!>87 probabilistic assessment ofthe actual projec1 duration, allowing for some activities lhat arc oo
1.00 .
15 -taso
more tbm one path. Problem 19 in the supplement to Chapter 18 illustrntes this.
g-hi 1:072 + 140
. .9192

Paths d-e-fand g-h-i have z values that are less than +3.00.
From Appendix B, Table B. the area to the left of z = -1.00 is .1587. and the tow:a.
theleftofz = +1.40 is .9192.Thepathdlstributions are illustlatedin tbe figure.TheJOIDl
probability of all finishing before week IS is the product of their probabilities:
1.00(.1587)(.9192) = .1459.
TIME-COST TRADE-OFFS: CRASHING
Estimates ofactivity times for projects usually arc made for sbme given level of resources. In
many situations. it is possible to reduce the length of a project by injecting additional re-
sources.The impetUs to sborten projects may reflec1 efforts to 8\'0id late penalties. to take ad-
vantage of monetary incentives for timely or early competition of a project, or to free
754
755

Pro}tcts sutlr as btidgt

. FIGURE 17.11 .'
crJIISITUCflon f1rwJiw huge
btidgels. ltfQ/Iy supplltrS and Crashing activities
tmpl~m. tmd btdup piau
Mllltriols,~pll,prous.ws.
olld UllcbntrofJdbf4
cimtmsuliiCts, suc1t 4S wttztlter.
GVI1l CIIIU defoyJ f1ltll adJitiol!a/
cosu.

Projact ie~gth
~ Crashing activities reduce Indirect project costs and
lllCte&se direct costs; the optimum amount of craahing
results In minimizing the aum of these two types or costs.

resources for U$t on other projects. In new product development, shortening may lead to a
stiategic benefit: beating the competition to the muket. In some cases, however, the desire to Using the following information, develop the optimal lilllH:ost solution. Indirect project costs
shortm the length of a project merely reflects 1111 attempt to reduce the indirect costs associ-
are $1,000 per day. EXAMPLE 7
ated with running the project, SIICh as facilities ahd equipment costs, supervision, and labor Nomal Ctaa Cas!Ptr D.y
and personnel costs. Managers often have various options at their disposal that will allow them Actirily llln n.. tatmll
crash Shortening activity to shorten. or crash. certain activities. Among the most obvious options are the use ofaddi-
dUilltions. tional funds to support additional personnel or more efficient equipment, and the relaxing of
10 $500
some work specifications. Hence, a project manager may be able to shorten a project by in-
:Q)
creasing dirl!t:t ~penses to speed up the project, thereby realizing savings on indirect project
700
costs. The goal in evaluating time-cost lrade-offs il; to identify activities that will reduce the
600
sum of the indirect and direct project costs.
800
In order to make a rational decision on which actiVities, ifany, to cmsh and on the extent of
cmshing desirable, a manager needs certain infonnation:
L Regulat time and crash time estimates fur each IICtivity.
2. Regular cost and crash cost estimates for each activity.
3. Alist ofactivities that are on the critical patti.
Activities on the critical path are potential candidates for crashing, because shortening
noncritical activities would not have an impact an total project dU111tion. From an economic
standpoint, activities should be cmhed according to cmsbing costs: Crash those with
the lowest crash costs first. Moreover, crashing should continue as long as lbe cost to crash
is less than the benefits derived from cmshing. Figure 17.11 illustmtes the basic cost
relationships.
Cmshing analysis requires estimates of regular and crash times and costs for each activity, SOLUTION - - -
path lengths, and identificatioo of critical activities. The general procedure for aashing is a. Determine which activities are on tbe critical path, its lenglh, and the length of the other
path:
I. Crash the project one period at a time.
hill leqlb
2. Crash the least expensive activity that is on the critical pa!b.
ibl . 18
3. When there are multiple critical paths, find tbe sum of crashing lite least expensive activ-
<-del. .... ~0 fcrilicai pWIJ
ity on each critical path. Lf two or more critical paths share common activities, compare
the least expensive cost of crashing a common activity shared by critical paths with the b. Rank the critical path activities in order oflowcst crashing cost, and determine the num-
sum for the separate critical paths. ber of days each can be cmsbed.
756

I. Use of these techniques forces1be manager to otganize and quantify available informa-
tion and to recognize wbere additional infonnarioo is neetled.
c. . 0 0 . . . 0 $300 2. The techniques provide a graphic display of the project lind its major activities.
... .... ..... 600 3. They identify (a) activities that should be closely watched because of the potential for de-
d........ ... 700 laying the project and (b) other activities that have ~lack lime and so can be delayed with-
f........... 600 out affecting project completion time. This raises the possibility of reallocating resources
to shorten the project.
c. Begin shortening the project, one day at a time. 111d check after each reduction to see
which path is critical {After a certain point, another path may equal the length of the No analytical technique is without potential errors. Among the more important sources of
shortened criti~ path.) Thus: errors are
(I) Shorten activity c one day at a cost of SJOO. The length of the critical path now be-
comes 19 days. J. When developing the project networlc, managers may WJwittingly omit one or mom im-
portant activities.
{2) Activity c C8IJil()t be shortened any mom. Shorten activity e one day at a cost ofS600.
The teng1h of path c-d-e-f now beco~ 18 days, which is the same as the length of 2. Precedence relationships may not all be correct as shown.
palha-b-f. 3. Time estimates may include a fudge factor; managers may feel uncomfortable about mak-
(3) The paths are now both critical; further iniprovemeoiS wm necessitate shortening ing time estimates because they appear 10 commit themselves to completion within a cer-
bolhpaths. tain time period
The remaining activities for crashing and their COSill are: 4. There may be a tendency to focus solely on activities that are on the critical path. As the
project progresses, other paths may become critical. Furtllertnore, major risk events may
not be on 1be critical path.
1-llf No11du~~
S500
600
No ful1lllt rtdue1ion P!ISSible
TECHNOLOGY FOR MANAGING PROJECTS
$700 Technology has had a nwnber of benefits for project management Among those benefits are
600 the use of computer-aided design (CAD) to produce updated prototypes on construction aod
600 product-development projects; software such as Lotus Notes tb l<eep team members who are
in separate locations in close contact; the ability for remote vieWing of projects, allowing those
At first glance, it \WUld seem that crashing f wollld nol be advantageous, because it has in different locations a firsthand view of progress and problems; and computer software.
the higbest ~hing cost However, f is oo both paths, so shortening fby one day would Project management software is often used to facilitate project lnanagement Here are a few
shorten both paths (and hence, the project) by one day fora cost ofS800. The option of of the many project management software packages that are aviilable: CA-Super Project, Har-
shortening the least expensive activity on each patb would cost SSOO forb and $600 fore, vard Total Project Manager, Microsoft Project for Wiodows, and Time Line. The particular
or Sl,IOO. Thus shorten fby one day. The project duration is now 17 days. version, e.g., 4.0, is not included because new versions are released frequently.
(4) At 1bis point, no additional improvement is feasible. The cost to crash b is SSOO and There are many advantages to using a project management software package. Among them
the cost to crash e is S600, for a total ofSI,lOO, and that would exceed the indirect are the following:
project cosiS of$1,000 per day.
It imposes a methodology and a common project managerhent terminology.
(S) The crashing sequence is sumroarized below:
It provides a logical planning structure.
1B1611t AfTBI CIIASIUlfQ DAYS: It can enhance communication among team members.
PG = t It can flag the occurreoce of consuaint violations.
lbf 11 18 18 11 It automatically formats reports.

-
c-d+f 20 19 18 17
It can generate multiple levels of swnmary reports and detailed reports.
ActMty crlslled r
Co$l $300 SUI It enables "what-ir scenarios.
It can generate various chan types, including basic Gaott charts.
An important benefit of the sequential crashing procedure just descn'bed is that it provides
the ability to quote different budget costs for differem project times. Certain requirements am associated with using project management software. One is the
need to keep the system current by entering cb1111ges as SOOil as possible. Also, as with any
ADVANTAGES OF USING PERT AND software, as new versions are released, the vendor may drop support for older versions, re-
quiring users to move to the latest version, and that can mean learning to use that version.
POTENTIAL SOURCES OF ERROR One thing to keep in mind is that project mana,.oement is more than choosing the rigbt soft-
PERT and sinn1ar project scheduling techniques can provide important services for the project ware. There is much that a project manager must do. Recall the key decisions that were dis-
mannger. Among the most useful features are these: cussed early in the chapter.
764 765

T-50weelcs a. (I) Compute palh ltngtlu and identify the aitical path: SolUtion
llanlica
l 50- 51 - -.82 (wttb)
1.22
e-b..... .... .. . 24(triticlt path)
c-4............ 19
l.f ............ 23

(2) Rank critical activities acrording Ill msh costs:

tost,.w...
Amily IIICIJM
l 50- 49 1.67 b..... ..... .. $ 3,lllO
.60
... .. .. . .. ... 11.000

Activity b should bt $hortened one week siru:e it has the Jawor cruhihg cost This ~ re-
duce indirett costs by Sl2,000 at a cost of$3,000, for and sav~ of$9,000. At this point.
paths a-band e-fwould both have a length of23 weeks, so both would bt aitia.l.
(3) RJnlt ~itics by msbing costs on the two critical paths:
50
l 50- <41 10,59
E Cost'"' Wotlt
.85 et.s~~

1-b.........

..r.........
0 10.59 zsc:ale
41 50 Weeks Choose ooe activity (the least costly) oo each patb 10 crdsh: bona-band f on e-f. for a total
=
costof$<4,000 + $2,000 $6.000 and a ne~-mgs of$12,000- S6,000 =$6,000.
(4) Check Ill set wbich path(s)migbt bt critical: a-band o-fv.wld bt 22 weeks in leng1h, and c-
Problem 4 IDdi~ costs fora project tre $12,000 per week for as long as !he project lasts. The projec1 managct d would still bt 19 ~ks.
has !Upplied the cost and time information shown.. Usc lilt inAlrmation Ill (S) Rank activities on the critical paths:
a. Determine an optimum crashing plan. tost,..Wni
b. G13ph the to!al costs for the plan. toCmll

,...,..
Cmliatl
tw.bl
a-11.. ....... s 4,lm
n;m
..!......... 6.1100
Sll,lm (no huther mshilg pouib!a)
3.11) fintM8. SUI! ollws
Crash boo path a-band eon e-rror a costof$4,000 + S6,000 = $10,000, foranetsavingsof
&,lm $12.000 - $10,000 =$2,000.
t,lm
(6) At this point, no further irnprovtment is possible: paths a-band e-fwould bt 21 weeks in
6,000
length. and one activity from each p3th would e 111 bt sbortened. This would mean activity
2JDJ aat$11,000 and e at$6,000 for a total of$17,000, whidl cxcteds the Sl2,000 poteutia! SIIV
inp in indim:t costs.
b. The followins table summarizes the results, showing the ltogth of the project after crashing n
...ms:
hill = 1 l
..b ...................... 24 23 21 21
td ...................... 19 19 19 19
.......................... 23 23 21 !1
AciMty tl'l~d ... ........ b bJ b,e
Cmhing costs 1$11.101 .. .... 10
766
767
II. Wby mighl a ptnOD wish to be iuvolved with a critical path activity? What are SOIIlC of the reasons
A sumnmy of cost,s for the ~ing sdledule Wf)U)d look Hke this: one might have for aot wanting Ibis association?
~ r:.olaiM 12. Wbatare some of the pocmtial benefits of working oo a special project in one~ firm? What are
50lDt ofthe risks?
Project Wab ~ Wlittt Talll
l..ollga SllaMid CciiiiSIIIII Coils111011 C.l-1 13. What are someupects oftbeprojectmanagenjob that mmit~ demmding !han tbejob of a
24(12) 718 managtt WOiking in amore routine orpniD1iooaJ ~?
24 188
23 23(121 71& 279 14. What is the main benefd or. project organization rYVr:t mon1 traditional forms of operations man-
3+6 = 9 221121 c 2&4 agemenl for project worl(l
22 Zl3
21 9+ 10 = 19 21ft2l "' 252 271
21) 19+17=35 ~121 2411 71& I. What trade-offs are associated wi1h time and cost estim..tts fof a proposed project?
TAKING
The gmph oftocal C05IS isu follows. 2. Wbo needs to be involved in assessing the cost ofa project?
STOCK
3. Name and ~ briefty tMI diffrreot 'M1Y3 that leclmology has had an impact 1lll project managancrrt.

290 288
Project ITWiagemCnt techniques have been \lied suec:mfully for a wide variety ofefforts. including the
279 many NASA space missioos, hU&C OOIIStnlction projects, implementldoo of1111jor ~ SliCh as ERP,
CRITICAL
280 276
273 produaioo of movies. de\-elopmem of new productt and services, th<atrical productions, and much THINKING
8u 271
more. Why not usc them for managing the operations function of auy business? EXERCISE
270 ~

~
{! Crash
200 r
I. For each of the following network diagrams, determiue bolh ille tritical path and the expec1ed proj-
cct dwalion. The liUIIIben 1lll the IITOWS ~expected lictivhy times.
PROBLEMS

oL
a. AOAdiagram

J
20 21 22 23 24
Project duration ~ksl

DISCUSSION L A project llliiJ1I8Cr may nted I"MIIkiU sets-diose of a ~ and !bose of a lca&r. Explain.
AND REVIEW 2. Explain the tennproject dunrrpion and list some vltyS to keep acb3mpion invoMd with the project
QUESTIONS 3. Us! the step$ in risk management.
4. Give some c:wnp!cs of ed!ical issues that may arise oo projects. What can a project manager do to
miDimizc ~ issues?
S. What are tilt key cdvaDt3ges ofusing poject managtment soflwm?
6. Will! is I wed: breakdown struciUle, and bow is it ujjjliiJ. for project planning? b. AON~
7. Identify the tenn being dcsaibed for each of the ~
a. A sequence ofactivities in a project 3
b. The ~oa&est time sequence oractivities in a prtljeclt.
t. Used when tMI activities have the same startilli illld finUbingpoinls.
d. The dilf=w:e in time length oflny pal!! and tllt !!ritial palb.
t. The mtistical distribution wed to describe variability of311 activity time.
J The llalistical distnOution \lied to descrme palil V1riability.
~ Sbtening an activity by alkxating additionalle!1ources.
8. liJt the main advantages ofPERT. Ust the maio liallruioas.
9. Why might a prob.1bilisticcstimareofa project'sconlpletioo time based solely oo the \"Uiaoce oflhc
critia11 pat1t be misleading? Under what t~ 'MlUld it be ~7
10. DefiDe each of these terms, l!1d iDdiate IIOf ead! Is cletennined.
a. ~activity time.
b. VoriaDce oflnactivity 1inle.
c. Stmdard cleviatiao ofa p;llh~ time.