Professional Documents
Culture Documents
Dec 02
Dec 02
Management
Journal
The Professional Journal of the Project Management Institute
Volume 33, Number 4 | December 2002
PAPERS
71 Calendar
71 Correspondence
72 Advertisers Index
Project Management Journal
PROJECT MANAGEMENT Book Development Editor PRODUCTION SERVICES PROVIDED
JOURNAL EDITOR Richard Schwartz; richard.schwartz@pmi.org BY IMAGINATION PUBLISHING,
Permissions Coordinator CHICAGO, IL, USA
Parviz F. Rad, PhD, PMP
Stevens Institute of Technology Insuk Choe; insuk.choe@pmi.org Senior Editor
Bookstore Administrator Ross Foti; rfoti@imaginepub.com
Book Review Editor
Kenneth H. Rose, PMP Regina Madonna; regina.madonna@pmi.org Assistant Managing Editor
Book Publishing Planner Lauren Strandquist; lstrandquist@imaginepub.com
PMI PUBLISHING STAFF Danielle Moore; danielle.moore@pmi.org Art Director
Publisher Doug Kelly; dkelly@imaginepub.com
Administrative Assistant
Linda Cherry; linda.cherry@pmi.org
Dotti Bobst; dotti.bobst@pmi.org Associate Art Director
Editor in Chief Tonya Weiland; tweiland@imaginepub.com
General e-mail:
Dan Goldfischer; dan.goldfischer@pmi.org
pmipub@pmi.org Associate Art Director
Publishing Support Specialist Theresa Rogulic; trogulic@imaginepub.com
Natasha Pollard; natasha.pollard@pmi.org
Publications Services Manager
Angela Kramer; akramer@imaginepub.com
EDITORIAL REVIEW BOARD PUBLICATION & MEMBERSHIP Project Management Institute, Publishing
John R. Adams, Western Carolina University; Frank T. Department, Four Campus Boulevard, Newtown Square,
The Project Management Journal (USPS 8756-9728/
Anbari, PMP, The George Washington University; Bud Pennsylvania 19073-3299 USA. Phone +610-356-4600,
02) is published quarterly (March, June, September,
Baker, Wright State University; Rick Bilbro, The Innova fax +610-356-4647; e-mail: pmipub@pmi.org
December) by the Project Management Institute. PMJ
Group, Inc.; David Christensen, Cedar City, UT; David Unless otherwise specified, all letters and articles
is printed in the USA by Cadmus Magazine, Richmond,
Cleland, University of Pittsburgh; Helen S. Cooke, sent to the PMI Publishing Department are assumed
VA. Periodical postage paid at Newtown Square, PA
Cooke & Cooke; Dan H. Cooprider, Creative People for publication and become the copyright property
19073 USA and at additional mailing offices.
Management; Jeffrey Covin, Indiana University; Steven of PMI if published. PMI is not responsible for loss,
Canadian agreement #40030957. Postmaster: Send
V. DelGrosso, IBM; Deborah Fisher, University of New damage, or injury to unsolicited manuscripts or other
address changes to Project Management Institute,
Mexico; Vipul Gupta, Saint Joseph’s University; Kenneth material.
Headquarters, Four Campus Boulevard, Newtown
O. Hartley, PMP, Parsons Brinckerhoff; Francis T. Square, Pennsylvania 19073-3299 USA. Phone +610-
Hartman, The University of Calgary; Gary C. 356-4600, fax +610-356-4647. READER SERVICES
Humphreys, Humphreys & Associates; Lewis Ireland, The mission of the PMJ is to provide information Photocopies. PMJ has been registered with the
PMP, Project Technologies Corp.; Peter Kapsales, advancing the state of the art of the knowledge of pro- Copyright Clearance Center, Inc. Consent is given for
Bellcore; Young Hoon Kwak, The George Washington ject management. PMJ is devoted to both theory and copying of articles for personal or internal use, or for
University; Lee R. Lambert, PMP, Lambert Consulting practice in the field of project management. Authors the personal use of specific clients. This consent is
Group, Inc.; Alexander Laufer, Technion-Israel Inst. are encouraged to submit original manuscripts that given on the condition that the copier pays through the
of Technology; Bill Leban, Keller Graduate School are derived from research-oriented studies as well as Center the per copy fee stated in the code on the first
of Management; Robert Loo, The University of practitioner case studies. All articles in PMJ are the page of each article for copying beyond that permitted
Lethbridge; Kenneth G. Merkel, PMP, University of views of the authors and are not necessarily those of by Sections 107 or 108 of the U.S. Copyright Law. The
Nebraska; Michael D. Okrent, Agilent Technologies PMI. Subscription rate for members is $14 per year appropriate fee should be forwarded with a copy of the
Inc.; Peggy C. Poindexter, Great Falls, VA; Tzvi Raz, Tel and is included in the annual dues. first page of the article to the Copyright Clearance
Aviv Unversity; Paul B. Robinson, The Eastern Claims for undelivered copies must be made no Center, Inc., 222 Rosewood Drive, Danvers, MA 01923
Management Group; Arnold M. Ruskin, PMP, Claremont later than two months following month of publication. USA (tel: +508-750-8400; fax: +508-750-4744). The
Consulting Group; Avraham Shtub, Tel Aviv University; The publisher will supply missing copies when losses fee indicated on the first page of an article in this issue
Richard L. Sinatra, PMP, Potomac, MD; Dwight Smith- have been sustained in transit and when the reserve will apply retroactively to all articles published in the
Daniels, Arizona State University; James Snyder, stock will permit. journal, regardless of the year of publication. This con-
Springfield, PA; Paul Solomon, PMP, B-2 Earned Value PMI is a nonprofit professional organization whose sent does not extend to other kinds of copying, such as
Management Systems; Robert G. Staples, Monroe, VA; mission is to serve the professional interests of its col- for general distribution, resale, advertising and promo-
Charles J. Teplitz, University of San Diego; Veljko M. lective membership by: advancing the state of the art tion purposes, or for creating new collective works.
Torbica, PMP, Florida International University; Walter in the leadership and practice of managing projects Special written permission must be obtained from the
A. Wawruck, Vancouver, BC, Canada; Itzhak and programs; fostering professionalism in the man- publisher for such copying.
Wirth, Saint John’s University; Janet K. Yates, San Jose agement of projects; and advocating acceptance of Permissions. Requests to reprint articles pub-
State University project management as a profession and discipline. lished in PMJ must be made in writing to the publisher.
Membership in PMI is open to all at an annual dues of Reprints. Individual articles may be purchased
PUBLICATIONS ADVISORY BOARD $119 U.S. For information on PMI programs and mem- through the Knowledge & Wisdom Department
Linda C. Cherry, PMI Publisher; Greg Hutchins, Quality bership, or to report change of address or problems (www.pmi.org/k&wc) at a cost of $10.00 per article
Plus Engineering; William V. Leban Jr., PMP, Keller with your subscription, contact: for members and $15.00 per article for nonmembers.
Graduate School of Management; Charles J. Teplitz, Project Management Institute, Headquarters, Glossy Reprints. Requests for glossy reprints of
PMP, University of San Diego School of Business Four Campus Boulevard, Newtown Square, Pennsylvania individual articles in quantities of 100 or more can be
Administration; John Sullivan, Reynolds & Reynolds 19073-3299 USA; tel: +610-356-4600; fax: +610- sent to pmipub@pmi.org.
356-4647; e-mail: pmihq@pmi.org Bulk Copies of Current Issues. Copies of the cur-
rent PMJ can be obtained in quantities of 25 or more.
EDITORIAL & ADVERTISING SERVICES Orders must be placed 40 days prior to date of issue.
The cost is $5.50 per copy plus shipping.
Address manuscripts and other editorial submissions,
Back Issues. Back issues are available on request
advertising and mailing list rental inquiries, and
at $20 each. Call +610-356-4600 for availability.
requests for reprints/bulk copies/reprint permission to:
Janice Thomas, Athabasca University, Center for Innovative Management, #301 Grandin Park
Plaza, 22 Sir Winston Churchill Ave., St. Albert, Alberta T8N 1B4 Canada
W orldwide, firms are in a race to create value and survive in the increasingly
competitive marketplace. They combine various organizational assets in
unique ways to create value. In turn, value creation involves tangible assets, such as
financial, physical, and technological ones, and less tangible ones, such as human,
organizational, and social assets (Brush, Greene, Hart, & Haller, 2001).
This paper explores project management maturity models (MMs) as a project
management construct and assesses them as sources of competitive advantage.
With the widely publicized project failure rates and related cost overruns
(Standish Group, 2001), more companies are turning to approaches that may help
▼ them improve their project management practices. Firms are turning to project man-
Abstract agement as part of their competitive advantage strategies. This is evident in the expo-
Project management maturity models are nential increase of membership in project management associations such as the
important assessment tools for the profes- Project Management Institute (PMI). In addition, the billions of dollars invested in
sion. Maturity models identify organizational
strengths and weaknesses as well as pro- projects, along with the increasing attention on MMs, quantifying the value of pro-
vide benchmarking information. They cap- ject management, and developing classification systems for the discipline exemplify
ture explicit, codified practice (know-what), support for project management (Bounds, 1998; Ibbs & Kwak, 1997; PMI Standards
but do not include the intangible assets of
Committee, 2000; Schlichter, 2000).
project management (know-how). Some
have made the claim that project manage- Because investments are time, cost, and resource intensive, firms are willing to
ment maturity models (MMs) can lead to a take only a critical interest in those practices that will improve their competitive
competitive advantage for firms. This paper positions. Projects are an essential building block of business value. However, we
uses four resource-based frameworks to
have yet to understand the complex ties between project management and its
assess whether or not maturity models lead
to a sustained competitive advantage. In value as a strategic asset.
the context of the strategy domain, the Over the past decade, MMs emerged in the literature as concrete, tangible ways of
authors conclude that MMs can result in a assessing aspects of a firm’s project management maturity. They help firms compare
temporary competitive advantage but not
a sustained competitive advantage. Clearly,
explicit competences at the project and program level relative to a standard. The
a sustained competitive advantage is rooted models are gaining interest as companies and academics strive to make sense of why
in a combination of know-what and know-how. some projects succeed and others do not. Companies and academics are questioning
project success rates as well as the returns from an investment in project manage-
Keywords: resource-based view; project ment. Proponents claim that the MMs enable firms to achieve a competitive advan-
management maturity models; competitive tage. However, no thorough exploration of this claim has been made.
advantage; strategy
The purpose of this paper is to rigorously explore the foundations of these
claims. The authors posit that the essence of a firm’s competitive advantage rests
©2002 by the Project Management Institute
2002, Vol. 33, No. 4, 4–14 in its resources (strategic assets) that are valuable, rare, inimitable, and involve
8756–9728/02/$10.00 per article + $0.50 per page organizational focus (Barney, 2002). The authors begin with an overview of com-
petitive advantage and competitive convergence and then introduce MMs. This is
Strategic Assets
Core Competencies
Capabilities
Generic Resources
Source. Brush, C.G., Green, P.G., Hart, M.M., & Haller, H.S. (2001). From initial idea to unique advantage: The entrepreneurial challenge of
constructing a resource base. The Academy of Management Executive, 15 (1), 64–78.
social complexity, and ambiguity); immobile (firm specif- Isolating mechanisms are a blend of resource characteris-
ic); nonsubstitutable; durable (long lasting); low tradabili- tics and managerial practices (Barney, 1989; Grant, 1991).
ty; and organizational focus (corporate attention) (Amit & They make some resources more firm specific and less mobile.
Schoemaker, 1993; Barney, 2002; Collis & Montgomery, Some examples include copyrights, patents, trademark laws,
1995; Grant, 1991; Jugdev & Thomas, 2002a; Peteraf, 1993; invisible assets (the features or organizational practices that
Priem & Butler, 2001a). one takes for granted and are the unspoken or tacit attributes),
In economic terms, value is measured through decreasing and small decisions such as the micro steps involved in
product/service costs or differentiating it to charge a premium quality improvement practices (Collis, 1994; Grant, 1991;
price (Barney, 1998; Duncan, Ginter, & Swayne, 1998). Itami & Roehl, 1987). Strategic assets incorporate isolating
Valuable resources are more worthy. A resource has value when mechanisms as:
it exploits opportunities and neutralizes threats in the ■ Inimitable—Resources are difficult to copy, and firms may
environment (Barney, 1991). Common or generic resources undertake practices to keep competitors from mimicking
are not sources of competitive advantage; at best, they are a them. If resources can be copied, a firm stands to only achieve
source of competitive convergence. competitive parity through resource value and rarity (Collis &
However, rare resources can offer temporary competitive Montgomery, 1995);
advantages and are sources of strength (Mata, Fuerst, & Barney, ■ Path Dependency—History matters. A firm may develop
1995). Rare resources are heterogeneously distributed between or acquire resources in low-cost ways by being, for example, in
firms. If a resource is rare and a firm does not have it, the lack the right place at the right time. It becomes expensive for rival
is perceived to be a weakness (Duncan, Ginter, & Swayne, firms to recreate the conditions as they developed over time
1998). Imperfectly mobile resources are “sticky” to the firm, (Mata, Fuerst, & Barney, 1995). Because the paths companies
meaning that they are not tradable as commodities and do not take are unique, firms have idiosyncratic assets (Barney, 1991).
leave a company when people leave (Priem & Butler, 2001a). ■ Socially Complex—Resources are the result of the interre-
These resources are characterized as: lationships between skills and assets. Social complexity can
■ Nonsubstitutable—Those resources for which other refer to a firm’s culture, relationships, and reputation. These
resources cannot fulfill the same function (Priem & Butler, attributes are related to a firm’s history and cannot be acquired
2001b); quickly or changed rapidly. They can provide firms with tem-
■ Nontransferable—Firm-specific resources that fall in value porary advantages from low-cost imitators. “In complex, high-
when they are transferred (Grant, 1991). This is a synonym for ly interdependent human and technological situations, the
imperfectly mobile resources; causes of success and failure are often difficult to assign” (Reed
■ Embedded—Engrained resources within the company’s & DeFillippi, 1990, p. 12).
routines and processes in contrast to assets that are specific to ■ Causally Ambiguous—Resources that competitors may not
individuals (Grant, 1991). fully understand to be able to copy them. It is costly and time
Complex
Technological
Human
Simple
Financial
Physical
Utilitarian Instrumental
Source. Brush, C.G., Green, P.G., Hart, M.M., & Haller, H.S. (2001). From initial idea to unique advantage: The entrepreneurial challenge of construct-
ing a resource base. The Academy of Management Executive, 15 (1), 64–78.
consuming for rival firms to figure out the sources of the assets. The unique advantages for the firm occur when these
advantage or what makes some resources strategic. assets are valuable, rare, inimitable, and nonsubstitutable
Last, the organizational focus aspect of strategic assets is least (Brush, Greene, Hart, & Haller, 2001). Not all RBV sources dis-
developed in the RBV literature. It refers to the breadth of tinguish between core competences and strategic assets as
managerial commitment and support. Brush does. Many sources treat these terms as synonyms.
These terms reflect the complexity within the perspective. The pyramid indicates that companies have many resources
Several RBV frameworks draw on the concepts and reflect the but few strategic assets. It is a simple, two-dimensional por-
features of strategic assets. trayal of resources that helps us begin to think of MMs in the
RBV context as well as where they may fit in the hierarchy.
RBV Frameworks MMs focus on the know-what within the firm. They cap-
Because the RBV is an evolving perspective, a unified model on ture project management competences in the form of docu-
the concepts and processes of strategic assets is not in place. ments and procedures. MM competences consist of generic
However, a number of frameworks have been proposed in the resources, capabilities, and possibly some core competences.
literature, and they have merit for project management. Of the The generic resources include technology, supplies, and
four frameworks discussed, the first two are classified as frame- materials necessary to plan and execute on the methodology.
works on type, complexity, and use of organizational assets, and These resources could be the hardware, software, communica-
the latter two reflect competitive advantage pathways. tion technology, manuals, templates, and so forth. The
Brush developed two RBV frameworks: the Resource capabilities refer to the individual, team, and organizational
Pyramid of Value Creation and the Resource Development abilities supporting project management. Some examples
Pathway (Brush, Greene, Hart, & Haller, 2001). include individual knowledge and experience, effective
Resource Pyramid of Value Creation. In the pyramid teamwork, good communication routines, and organizational
(Figure 1), generic resources (supplies and materials) are at support for project management methods.
the base and often are easy to identify and access. Once com- Most MMs typically capture explicit knowledge as docu-
bined with other resources, generic ones become capabilities mented in the five levels. They do not capture the intangible
that enhance a firm’s ability to deploy resources. Capabilities assets. MMs do not appear to fit at the apex of the pyramid
are combinations of proprietary resources, knowledge, and where unique advantages or strategic assets are found.
skills that become institutionalized into operating routines Resource Development Pathway. Another framework by
and tacit knowledge. Brush addresses resource complexity and use (Figure 2). It
Beyond this, collections of specialized core competencies groups resources into six categories based on function (Brush,
that allow a firm to outperform rivals combine into strategic Greene, Hart, & Haller, 2001). The two axes map a company’s
Is It
No Heterogeneously Yes
Distributed Across
Competing Firms?
No Yes
Is It Imperfectly
Mobile?
Temporary Sustained
Competitive Competitive Competitive
Competitive
Disadvantage Parity Advantage
Advantage
resource development pathway. Simple resources are tangible, be used to map a firm’s resources over time and compare firms
discrete, and property based. Complex ones are intangible, sys- in terms of their pathways as well. With respect to MMs, a
temic, and knowledge based. firm’s project management resources may fit the simple-utili-
Within the framework, resources are characterized by their tarian or complex-utilitarian quadrants. MMs involve explicit
application to the productive process (utilitarian to instru- codified documents that describe competences. MMs are
mental). A firm applies utilitarian resources directly to the pro- as utilitarian because, unlike financial resources, they do
ductive process or combines them to develop other resources. not provide access to other resources. Instead, their use
For example, machinery is a physical, utilitarian resource is straightforward and geared toward improved project man-
because it produces services or products. However, financial agement processes and practices.
resources are instrumental as they provide access to other MM assets are tangible because they are in the form of doc-
resources. They also are flexible resources because they are uments, surveys, guidelines, templates, or manuals. Although
needed to purchase other resources. Proprietary technology is MMs may assess a department’s or firm’s skills (competence)
either instrumental or utilitarian depending on if it resides within certain project management knowledge areas, the
within a person (intangible), in which case it is instrumental, assessments are based on documented practices. Moving to the
or it is a patented process applied directly to a production next level within a MM often depends on having more
process, in which case it is utilitarian (Brush, Greene, Hart, & extensive documentation, project management procedures, or
Haller, 2001). codified processes in place. Moving to the next level does not
The matrix portrays complex-instrumental resources as take organizational know-how into account.
most intricate. These resources include the human, social, In contrast, strategic assets typically fit the complex-instru-
and organizational resources. Human resources are complex mental quadrant. Complex-instrumental resources are those
and intangible because they are harder to identify and mea- that often are more intricate and connote dynamism and
sure (Brush, Greene, Hart, & Haller, 2001). Complex regeneration. These involve human, social, and organizational
resources and knowledge assets are not like normal tradable resources. An excerpt from a MM instrument describes the
commodities that firms exchange on the market (Priem & competences at MM5:
Butler, 2001a; Wernerfelt, 1984). In general, the characteris-
tics are engrained within the firm’s learning routines making “Level 5: Optimizing Process. Processes are in
them harder for competitors to duplicate. Such resources are place and actively used to improve project man-
embedded within the firm. agement activities. Lessons learned are regularly
The resource development pathway is a more complex examined and used to improve project manage-
representation than the pyramid of value creation. It also can ment processes, standards, and documentation.
Competitive
1 No — — Below normal —
disadvantage
Competitive
2 Yes No — Normal Yes
parity
Temporary
3 Yes Yes No competitive Above normal Yes/No
advantage
Sustained
4 Yes Yes Yes competitive Above normal No
advantage
Source. Mata, F.J., Fuerst, W.L., & Barney, J.B. (1995). Information technology and sustained competitive advantage: A resource-based analysis.
MIS Quarterly, 19 (4), 487–507.
Management and the organization are not only financial metrics indicating that firms with higher MM scores
focused on effectively managing projects but also perform better and achieve more savings that those with lower
on continuous improvement. The metrics collect- MM scores (Ibbs & Kwak, 1998; Ibbs & Kwak, 2000).
ed during project execution are used to understand The second question of the framework asks if the resource
the performance of not only a project but also for is heterogeneously distributed among competing firms.
making organizational management decisions for Heterogeneous resources connote rarity as not all firms
the future” (Pennypacker, 2001, p. 3). have the same assets. Although MMs can be considered
heterogeneously distributed because not all firms use them,
The excerpt reflects consistent practices, incremental they are widely available to all firms, and it could be argued
improvements, and some managerial support but lacks a clear that they lead to competitive parity.
explanation of how practices are dynamic or regenerational. In The third and final question queries if the resource is
particular, it does not address project management know-how. imperfectly mobile. Recall that imperfectly mobile resources
The Brush matrix underscores a key limitation of MMs in that are characterized as being nonsubstitutable, nontransferable,
they cover the know-what of project management but not the and embedded. Going on the assumption that MMs are
know-how. If the MMs covered both kinds of knowledge, they heterogeneously distributed between firms, an assessment of
could fit the complex-instrumental quadrant. them at this step of the flow chart leads to the decision that
Unlike the two Brush frameworks, the next two frameworks they are mobile and lead to a temporary competitive advantage
move beyond classifying resources by type, complexity, and but not a sustained competitive advantage.
use. They assess the extent to which resources lead to a com- MMs are designed to be implemented in a variety of
petitive advantage. Both are variations of Barney’s framework organizations. They involve codified knowledge that makes
for a strategic asset that is called the VRIO model—valuable, them transferable between firms. The knowledge staff gains
rare, inimitable, and organizationally focused (Barney, 2002). from using the models is readily transferable to other firms.
Resource-Based Model of Competitive Advantage. Mata’s MMs are substitutable as there are a number of such products
framework evaluates assets using three key RBV features: value, on the market that all serve very similar functions. MMs also
heterogeneity, and immobility (Mata, Fuerst, & Barney, 1995). are not embedded within firms as they can leave a company
It is a decision flow chart (Figure 3). with staff turnover.
In assessing MMs according to Mata’s model, the first ques- Although MMs have merit, according to the Mata frame-
tion asks if the resources are valuable. MMs are valuable and work, they appear to result in a temporary competitive
have worth. In part, their worth is evident when companies advantage for some firms and competitive parity for most.
purchase and conduct maturity assessments, pay the consul- According to this flowchart, MMs do not lead to sustained
tant fees, software licensing fees, and provide staff training. The competitive advantages as purported in the literature
contribution of MMs is evident through some efficiency and (Association of Project Management, 2000; Australian Institute
This article is copyrighted material and has been reproduced with the permission of PMI. Unauthorized reproduction of this material is strictly prohibited.
T here is continuing interest in measuring and coping with project risk and uncer-
tainty by both academics (Fernandez, Armacost, & Pet-Edwards, 1998; Tavares,
Ferreira, & Coelho, 1998; Van Dorp & Duffey, 1999) and project management
(Hulett, 2000; Jones, 2000; Royer, 2000; Ruskin, 2000). Program evaluation and
review technique (PERT) and Monte Carlo simulation have been the traditional
approaches used to address project schedule uncertainty. Both methods are based on
critical path analysis and apply the probability theory to quantify the uncertainty
related to the time required to complete project activities. Concerning the use of
these methods, a recent survey of project management professionals (Pollack-
Johnson, & Liberatore, 1998) found that, of the respondents using analytical tech-
niques, nearly 90% used critical path analysis, and more than 20% used probabilis-
tic analysis and/or simulation. A 1999 software survey showed that about 20% of
project management software packages had Monte Carlo simulation capability
(Project Management Institute, 1999).
▼ However, there is an alternative approach for measuring schedule uncertainty
Abstract based on fuzzy logic. Fuzzy logic is an approach for measuring imprecision or vague-
ness in estimation and may be preferred to probability theory in capturing activity
There is continuing interest by academics and
practitioners alike in measuring and coping duration uncertainty in some situations. Several authors have proposed or investi-
with project schedule uncertainty. Fuzzy logic gated the fuzzy logic approach as applied to project schedule uncertainty (Buckley,
has been proposed as an alternate approach 1989; Chanas, 1982; Chanas, 1987; Chanas & Kamburowski, 1981; Chanas &
to probability theory for quantifying uncertainty
related to activity duration. However, the fuzzy
Kuchta,1998; Chanas & Zielinksi, 2001; Chang, Tsujimura, Gen, & Tozawa, 1995;
logic approach is not widely understood, and Dubois & Prade, 1979, 1980; Hapke, Jaszkiewicz, & Slowinski, 1994;
generally accepted computational approaches Kamburowski,1983; Lootsma, 1989; McCahon, 1993; McCahon & Lee, 1988; Prade,
are not available. This paper describes the dif- 1980; Rommelfanger, 1994; Shipley, de Korvin, & Omer, 1997). The dominant
ferences between the probabilistic and fuzzy
approaches and the advantages of the latter.
approach presented in the literature is the fuzzy extension of the standard critical
The paper also illustrates a straightforward path method algorithm, where forward and backward passes are taken in the net-
approach for applying fuzzy logic to assess pro- work to compute earliest and latest start times and slack. However, the backward pass
ject schedule uncertainty. has been reported to yield incorrect results when the minimum and subtraction
operations are replaced by their fuzzy counterparts (Chanas & Kuchta, 1998;
Keywords: activity duration estimating; Kamburowski, 1983; Rommelfanger, 1994).
critical path; schedule development; risk
management; fuzzy logic The direct use of the fuzzy backward pass was used commonly in some of the
early work on fuzzy critical path analysis and is still employed by some authors
©2002 by the Project Management Institute
2002, Vol. 33, No. 4, 15–22
(Hapke & Slowinski, 1996; Mares, 1994). Rommelfanger (1994) offers an alter-
8756–9728/02/$10.00 per article + $0.50 per page nate approach for computing the fuzzy backward pass that has some limitations.
A recent article by Chanas and Zielinski (2001) presents a method for analyzing
Degree of
Membership
0
175 180 185 190 195
Height in Centimeters
the degree of criticality of the fuzzy paths in the project net- we might assign a 0, implying certain exclusion from the set. Of
work but does not present a method for determining the course, because the meaning of the term “tall people” can vary
fuzzy criticality of project activities. by situation, a fuzzy set must be defined in context.
The fuzzy logic approach for project schedule uncertainty Formally, we could define the membership function
analysis is not widely understood, and as is evident, generally mT(194 cm) = 0.95 to represent the degree to which the indi-
accepted computational approaches are not available. This vidual with a height of 194 cm is a member of the set of tall
paper describes the differences between the probabilistic and (T) people. Using “x” to represent height, we could specify that
fuzzy approaches, the advantages of fuzzy approaches, and mT(x) = 1 if x ≥ 195 cm, and mT(x) = 0 if x ≤ 175 cm. For those
illustrates a straightforward approach for applying fuzzy logic values in-between, we could assess specific values for each
to assess project schedule uncertainty. height or use a formula such as mT(x) = [(x–175) / 20], if 175
< x < 195. For x = 193 cm, applying the formula yields 0.90.
Background on Fuzzy Logic This membership function is shown graphically as Figure 1.
Fuzzy logic is not fuzzy thinking. Zadeh (1965) introduced The theory and application of fuzzy sets has seen extensive
fuzzy sets to represent knowledge that is vague or imprecise, development (Dubois & Prade, 1980; Klir & Yuan, 1995; Yen &
that is, “fuzzy.” In a classical set theory, an element either is or Langari, 1999; Zadeh & Kacprzyk, 1992). There are many suc-
is not a member of a set. In contrast to the sharp or “crisp” cessful applications of fuzzy logic in control systems, including
boundaries of classical sets, fuzzy sets allow degrees of mem- water treatment, subway systems, washing machines, automo-
bership in a set, as expressed by a number between 0 and 1. tive transmission systems, and vacuum cleaners (Yen & Langari,
For example, suppose we wish to define membership in a 1999). Applications in management systems include project
set of tall people. In classical set theory, we might define the selection (Machacha & Bhattacharya, 2000), multi-
set of tall people as those individuals who are 195 cm (about criteria decision-making (Chiadamrong, 1999; Siskos, Lochard,
6'4 3/4'') or taller. But what if a person is 194 cm? This indi- & Lombard, 1984), cash flow analysis (Chiu & Park, 1994), and
vidual might naturally be described as somewhat tall, as job shop scheduling (Fortemps, 1997), to mention a few. The
would the person who is 193 cm. Yet, neither individual articles by Bagnoli and Smith (1997) on real estate decision-
would be a member of our set of tall people. This example making and Hutchinson (1998) on financial applications are
demonstrates that tallness might be better represented as a especially useful introductions to fuzzy business applications.
vague or imprecise concept, without sharp borders separating
the tall from those not tall. Fuzzy Logic vs. Probability
In a fuzzy set theory, we can accommodate this imprecision Probability theory often is used to model the uncertain dura-
by associating a degree of membership in the set of tall people tion of a project activity. We might say that the probability that
for individuals of varying heights. For example, we might asso- a given activity will take seven days to complete is 0.8 and the
ciate a value of 0.95 to represent the degree to which the 194 probability that it will take eight days is 0.2. Because these two
cm person is a member of the fuzzy set of tall people and a probabilities sum to 1.0, no other activity durations are possi-
lower value, 0.90, for the degree to which the 193 cm is a mem- ble in this example. The interpretation of these probability
ber. For the person who is 200 cm tall (nearly 6'7'') we would statements is that after repeating this activity many times, the
assign a membership value of 1.00, implying certain member- activity time was seven days in 80% of the cases, and the activ-
ship in the set. For the person who is 175 cm tall (about 5'9''), ity time was eight days in 20% of the cases. In essence, there is
G
A D
C E
for the critical path length is obtained by multiplying the activity on the critical path are identified. The belief that this
probabilities of all the individual activity times in that particular activity is critical is the maximum of the beliefs for
combination. The probability of a specific critical path all cases having this activity on its critical path.
length is the sum of the probabilities over all combina- 5. Compute the center of area (COA) for all paths, includ-
tions yielding that length. More importantly, the meaning ing the critical path to provide “crisp” values for their lengths.
and interpretation of the results of the probability and The COA is a special weighted average whose numerator is
fuzzy approaches are quite different. obtained by multiplying each path length by its belief and
summing over all possible path lengths, while its denominator
A Computational Approach is the sum of the beliefs of all possible path lengths (remem-
In this section, the author describes and illustrates a straight- ber that the beliefs will not sum to 1.0). The COA is our best
forward procedure for fuzzy critical path analysis using the estimate of each path length.
ideas from the previous section. A visual basic program has
been written to automate the enumeration procedure. This Example
enumeration algorithm is: A simple project is chosen to illustrate the application of
1. Enumerate all possible combinations of activity dura- the enumeration method and is explained with the help of
tions. Each combination is referred to as a case. The number of Figure 2 and Tables 1, 2, and 3. To avoid possible confusion
cases is found by multiplying the number of possible positive with activity duration, the activities were labeled using letters
belief values across all project activities. without subscripts. Using the precedence information given in
2. For each case, compute all path lengths and the case Table 1, the project network shown can be drawn as in Figure
belief value and identify the critical path. The belief value for 2. This network is displayed using activity on node (AON)
each case is the minimum of the activity beliefs comprising notation. Table 1 also provides data on the membership func-
this case. The longest path is the critical path for that case. The tions for each activity. The data used are discrete, as contrasted
path lengths and the critical path for this case all are assigned with the continuous approach described in our height example
the same case belief value. and shown in Figure 1.
3. Construct membership functions for the length of all Note that the durations for activities B, D, E, and F can be
paths including the critical path. For a given path, all cases are expressed as “about 2,” “about 3,” “about 7,” and “about 2,”
reviewed, and those having a particular path length are identi- respectively. In this example, the author is using three possible
fied. The belief for this particular length is the maximum of the values along with their belief values to define the uncertainty
beliefs of all cases having this length. This process is repeated associated with each of these four activities. Three-point
for all possible path lengths. estimates such as these are relatively easy to make. The infor-
4. Determine the fuzzy criticality for each activity. For a mation on the membership function for B can be interpreted
given activity, all cases are reviewed, and those having this as having possible values of 1, 2, and 3 with belief values of
0.8
0.6
Beliefs
0.4
0.3 0.3
0.2
0.1 0.1
0
14 15 16 17 18 19
Possible Critical Path Lengths
0.3, 1.0, and 0.5, respectively. In contrast, there is only one pos- certainly critical. On the other hand, B is certainly not critical
sible time with a belief of 1.0 for activities A, C, and G. That is, because it is only on path P2 and this path is certainly not crit-
it is assumed that these three time estimates are certain. The ical. The complete results are given in Table 3.
application of each of the five steps is: 5. The COAs for each path also are given in Table 3.
1. There are 1 x 3 x 1 x 3 x 3 x 3 x 1 = 81 possible cases that To further interpret the results, consider the case where all
must be evaluated; activities assume the length associated with a belief of 1.0
2. An examination of Figure 1 shows that there are three (A = 3, B = 2, C = 4, D = 3, E = 7, F = 2, and G = 2). The critical
possible paths in the network: path for this case is P1 with a length of 16 and a belief of 1.0 as
■ P1: A-C-E-G; shown in Table 2. The COA of 16.17 for the critical path is close
■ P2: A-B-D-F-G; to the value for this case because the network is small and the
■ P3: A-C-D-F-G. uncertainty is not that great. However, there is uncertainty in
For each of the 81 cases, the lengths of each path are completing this project, because P3 is moderately critical, and
computed, and the critical paths are identified. For example, as a result, D and F also are moderately critical. There also is
the belief for the case A = 3, B = 3, C = 4, D = 1, E = 7, F = 1, some uncertainty associated with the critical path length, which
and G = 2 is Min(1.0, 0.5,1.0, 0.1, 1.0, 0.5, 1.0) = 0.1. The com- has possible values ranging from 14 through 19 (Figure 3).
putations of the path lengths and beliefs, and the identification
of the critical paths are given in Table 2. Summary and Conclusions
3. Using the information given in Table 2, we can form the In this paper, the author has offered fuzzy logic as an alter-
necessary membership functions. For example, there are 18 cases native approach for modeling uncertainty in project sched-
where the critical path is 14. The beliefs over these cases are either ule analysis. The fuzzy logic approach has been compared
0.1 or 0.3, leading to a belief of 0.3 for a length of 14 because the and contrasted with probability theory. Fuzzy logic may be
maximum value is selected. The complete results are given in suitable in those situations where past data is either unavail-
Table 3 and shown graphically for the critical path in Figure 3. In able or not relevant, the definition of the activity itself is
a similar fashion, the path criticalities can be determined. somewhat unclear or the notion of the activity’s completion
4. Using the information given in Table 2, we can determine is vague. An enumeration procedure for analyzing a project
the activity criticalities. For example, D is on paths P2 and P3, network whose uncertainty is represented by fuzzy logic has
and an inspection of Table 2 shows that P2 is never critical and been presented and demonstrated. Future work is needed to
that P3 is critical in 23 cases. The maximum of the beliefs of better identify those circumstances where project managers
these 23 cases is 0.4, and so this value is the fuzzy criticality of prefer fuzzy logic to probability theory or where either
D. This same result can be obtained by noticing that the beliefs approach could be used. In addition, improved computa-
that P2 and P3 are critical are 0 and 0.4, respectively. Because tional procedures are needed so that fuzzy logic can be
D lies on both of these paths, we take the maximum (0.0.4) = applied to projects of arbitrary size. These efforts should lead
0.4 to obtain its criticality. The same result holds for F. The activ- to increased practitioner use of the fuzzy logic approach for
ities along P1 (A, C, E, G) are all certainly critical because P1 is measuring project schedule uncertainty.
This article is copyrighted material and has been reproduced with the permission of PMI. Unauthorized reproduction of this material is strictly prohibited.
Toward a Dynamic Simulation
Model for Strategic Decision-Making
in Life-Cycle Project Management
Hemanta Kumar Doloi, The University of Sydney, Department of Civil Engineering,
Building J05, Sydney NSW 2006 Australia
Ali Jaafari, The University of Sydney, Department of Civil Engineering, Building J05,
Sydney NSW 2006 Australia
Simulate
Processes
Optimize
Processes
Yes Yes
Proceed With Should We Invest Proceed With
the New Product in a New Product? the New Product
No
Define the New Product Define the New Product
Do Not Proceed With LCC Model as Baseline
LCC Model as Baseline
the New Product
No
Should We Implement the
Alternative Products?
Yes
Accept the Alternative
Repeat the Process and Make the Alternative
as Required the New Baseline
This paper introduces a holistic approach for effective Simulation Model and the Integrated
front-end planning and management of the project deliver- Facility Engineering System
ables by focusing on the business objectives during the Discrete event simulation is a powerful means for the design of
early phases of the project. The underlying methodology mathematical-logical models of the real-world system and
enables project managers in meeting some of these chal- experimentation with alternative scenarios on the computer
lenges and allows design, evaluation, and visualization of screen. Simulation makes the target facility more realistic in
new and existing processes throughout the project life terms of functionality, and the required total investment cost
cycle. Development of a dynamic simulation modeling can be forecasted more accurately. Figure 1 shows a typical
system (DSMS) allows simulation and optimization of process simulation in the life-cycle decision model. Investment
selected processes during planning, implementation, and decisions can be optimized on project facilities based on finan-
operation phases and evaluation of the impacts of changes cial LCOFs (Jaafari & Manivong, 1998). The technique facili-
on the life-cycle objective functions (LCOFs) (Manivong & tates improved understanding of the real-life situation during
Jaafari, 1999). the conceptualization and finalization phases of the project.
Simulation Level
Simulation Simulation Simulation Simulation
Whole Project
Process Simulation
Submit
Optimization Level
Scope Evaluation
Resubmit in SPMIS Framework Resubmit
No LCOFs No
Satisfied
Yes
Figure 2 shows the overall architecture of the Integrated The IFE system entails integrated modules including:
Facility Engineering (IFE) project, currently under develop- ■ A smart project management information system (SPMIS)
ment at the University of Sydney’s Department of Civil to facilitate the analysis of project management functions
Engineering. The IFE is a generic system that aids management (Jaafari & Manivong, 1998);
of capital projects in an integrated project management envi- ■ A visual design management (VDM) system to assist in visu-
ronment. Target values set for LCOFs are used as the criteria to alization/schedule simulation and management of the design
guide decision-making. The architecture has been conceptual- process (Chaaya & Jaafari, 1999);
ized and designed to facilitate the uptake of life-cycle project ■ A construction management information system (CMIS)
management (LCPM) methodology for the delivery of projects (Jaafari, Manivong, & Chaaya, 2000);
(Jaafari & Manivong, 1998). ■ A dynamic simulation modeling system (DSMS) to
The LCOFs include: enhance the strategic decision analysis for project viability
■ The project’s financial status and its profitability; (Doloi & Jaafari, 2000);
■ The operability, quality, or performance of the facility; ■ A soft issues management system (SIMS) to evaluate soft
■ The soft issues such as project short- and long-term functions such as community and stakeholder issues (Jaafari
liabilities, including occupational health and safety (OH&S) & Vlasic, 1999).
risks throughout project life, environmental impacts, and The IFE system has a unified project databank that estab-
third-party liabilities. lishes a multi-access intranet configuration. It allows entry
Table 1. Definition of Characteristic Functions for an Idealized Process Simulation Model Embodying a Project
Management Information System
* Percent comparability is found by expressing the average total score from Table 3 as percentage of 6 (ideal score). For example, an average total
score of 3 is equal to 50% comparability.
Table 2. Description of Simulation Systems and Their Comparison to the Idealized System
System Interfacing. Most of the systems do not provide full many CAD programs to display static background and to
integration of the characteristic functions shown in Table 1. define material handling paths. Taylor-II (Nordgren, 1995) can
However, the commercial systems in the sample generally pro- generate model documentation in text file format and process
vide better integration with existing information technology external files (ASCII-file) for the whole model.
(IT) tools compared to the prototype systems. For example, Point Comparison and Ranking. The authors have under-
AweSim (Prisker & O'Reilly, 1997) is built on a relational data- taken a point comparison of the systems in the sample to eval-
base and accessible through standard tools such as DBase, uate their technological and functional capabilities (Table 3).
Access, FoxPro, and Microsoft Excel. Input data can move easi- The scoring scale ranges from 1 to 6:
ly from an Excel worksheet to the AweSim input tables. FAC- ■ Score 1: Poor;
TOR/AIM (Ehrlich & Lilegdon, 1997) allows creating custom ■ Score 2: Fair;
reports and graphs using Access report wizards. Output data ■ Score 3: Moderate;
can be saved directly as an Excel spreadsheet for further analy- ■ Score 4: Good;
sis. Many provide explicit integration capabilities with CAD ■ Score 5: High;
programs and dynamic data exchange. SIMAN/Cinema ■ Score 6: Ideal.
(Glavach & Sturrock, 1993), ProModel (Harrel & Leavy, 1993), Table 3 shows the total average score derived for each
and WITNESS (Markt & Mayer, 1997) import drawings from system. Individual scores are shown in columns against each
Dynamic implementation
functions
Performance evaluation
Modeling environment
Product management
Life-cycle application
Facility management
Process optimization
Application methods
Proactive evaluation
Scope management
Resource utilization
System integration
Simulation
systems
Construction-oriented
1 MICROCYCLONE 2 4 4 2 1 3 1 1 3 1.24
2 COOPS 2 4 4 3 1 4 1 2 3 1.41
3 CIPROS 3 3 4 4 4 1 3 2 2 3 1.71
4 STROBOSCOPE 2 3 3 3 3 1 2 3 1.18
5 SimCon 1 3 4 4 3 1 4 1 2 3 1.53
6 SIREN 2 3 3 3 3 1 3 1.18
7 SEACONS 2 2 4 4 3 4 2 1 3 1.47
8 CRUISER 2 2 4 4 3 1 4 2 1 4 1.59
9 AP3 3 2 4 4 3 2 2 2 2 4 1.65
Manufacturing-oriented
10 MODSIM III 3 2 5 5 4 1 5 3 3 5 2.12
11 PROMODEL 2 4 5 5 5 3 5 3 5 5 2.47
12 SIMFACTORY II.5 2 4 4 3 1 4 3 3 4 1.65
13 FACTOR/AIM 2 3 5 5 5 2 5 3 4 5 2.29
14 WITNESS 2 3 5 5 5 1 5 3 4 5 2.24
15 Taylor II 2 3 5 5 5 1 5 2 4 5 2.18
16 EXTEND 2 2 5 5 4 1 5 3 3 5 2.06
17 PROSIM 3 4 4 4 1 3 3 4 4 1.76
18 SIMPROCESS III 2 3 5 5 4 1 4 3 1 4 1.88
General purpose
19 GPSS/WORLD 2 3 5 5 4 1 4 3 3 4 2.00
20 SLAM II 3 5 5 4 1 5 4 5 1.88
21 PETRI-NET 3 4 4 3 1 4 2 3 3 1.59
22 SIMOBJECT 3 4 4 5 1 3 3 4 1.59
23 SIMPLE ++ 4 4 5 5 5 2 4 3 5 4 2.41
24 C++/CSIM17 2 3 4 4 4 1 3 3 3 1.59
25 SIMSCRIPT II.5 3 4 4 4 1 3 4 1.35
26 SIMAN/Cinema V 2 3 5 5 4 1 5 5 5 2.06
27 AweSim 1 4 5 5 4 2 5 1 5 5 2.18
28 SIMULA 3 5 5 4 1 5 3 4 1.76
Table 3. Point Comparison Between Existing Systems and the Idealized System
Activities
Generate Delay Assemble Release Replenish Block Batch Unbatch Interruption Discard
function. A blank space in a column represents a nonexistent used to represent these elements. A total of ten activities with
capability. An examination of these columns shows that none unique block symbols (generate, delay, assemble, release,
of the systems sampled provides the full capabilities sought. replenish, block, batch, unbatch, interruption, and discard) are
The last column of Table 2 shows the ranking of used to represent the behavior of processes and subprocesses
samples against the idealized DSMS in descending order. in the model.
The system with the highest ranking possesses less than Figure 4 shows the generic relationships among model-
one-half of the total capabilities defined for the idealized ing elements in DSMS framework. Processes involve a log-
system. The construction-oriented simulation tool, ically and sequentially related set of activities on selected
Construction Integrated PROject and process planning entities. The entity changes and produces an output at the
Simulation system (CIPROS) (Tommelein, Odeh, & Carr, end. If a process involves more than one function within
1994) scored the highest with 29% comparability require- the system that becomes too complex, the process must
ments. In the manufacturing category, ProModel scored further break down into subprocesses. Activities take place
the highest with 41% comparability. Under the general- within processes and subprocesses. Queue establishes the
purpose simulation category, SIMPLE++ (Geuder, 1995) necessary logical links between processes and subprocess-
came in first with 40% comparability. es. Resources are allocated for each activity within
Most of the available systems facilitate one or a combi- processes. It is postulated that this methodology can be
nation of the characteristic functions in various contexts. utilized as a generic application in various service situa-
The authors have concluded that a generic simulation sys- tions. DSMS interfaces will allow users to represent the
tem for decision evaluation within an integrated environ- real-life situations as a simulated model without prior
ment with the required capabilities (Table 1) does not exist. simulation training.
Furthermore, none of the systems reviewed provide a facil- The DSMS model utilizes a hierarchical and modular struc-
ity for proactive project scope evaluation based on holistic ture (Figure 5). This structure enhances the program’s capabil-
considerations. Under the traditional project management ity in simulating varied design alternatives (Zeigler, 1987). The
approach, projects are not proactively scrutinized, and deci- WBS adopted for each project facilitates this requirement. The
sions are mainly based on short-term profitability criteria WBS used in DSMS is based on the following hierarchy:
(Riggs, 1989). This clearly indicates that there is potential 1. Project. The project is considered to be the highest compo-
to research and develop a system that can furnish a lot nent in the hierarchy model and focuses on the gross project
more capability and functionality than the best systems attributes. At this level, attributes are project identification and
currently available. overall management issues from startup to the end of the project.
2. Part. A part is considered to be a subproject and defined
DSMS Specifications either as a physical part of the project or an associated soft
The DSMS has been conceptualized to simulate typical service deliverable. Major parts are identified at this level.
and production processes throughout the project life cycle. Six 3. Process. At this level, processes for each constituent
broad elements are utilized to construct the process model part are identified. The dynamic composition for a part is
(Figure 3). These include entities, processes, subprocesses, described as a system in terms of activities, events, and
queues, activities, and resources. Different block symbols are processes. If the processes are not easily identifiable at the
Resources
Subprocesses
Subprocess Subprocess Resource
1.2.1 1.2.3 1
Subprocess Resource
Subprocess 1.2.4 2
1.2.2
Resource
3
Activities Activity
1.2.4.3
Activity
1.2.4.1 Activity
1.2.4.2
Figure 4. Process Modeling Elements and Their Relationship in the Dynamic Simulation Modeling System
part level, parts are broken down further into products, The process simulation engine produces a design solution
i.e., subsystems. Processes are used to represent whole or at both the part and project levels. A user can select the time
part of the life of temporary entities in the real system. units in seconds, minutes, hours, days, and weeks. The model
Usually, a number of processes exist in a system model at statistics can be collected at all levels of the hierarchy.
any point in time. The process level is very important with- Modeling variability associated with uncertainties or
in the project hierarchical structure and defines function- randomness, such as activity times, interarrival times, time-to-
ality of the project. failure, and time-to-repair values, are achieved using statistical
Operational and technological sequencing of the distribution functions. The model outcomes are produced in
processes must be identified at the process level (Abdalla, the form of reports, bar charts, pie charts, tables, histograms,
1991). These processes are broken down further into sub- time plot variables, and text files. The model outputs are uti-
processes, and logical links are established with user- lized in terms of performance analysis, capacity analysis, com-
defined interfaces. The logical links between processes and parison study, sensitivity analysis, optimization study, deci-
subprocesses in DSMS methodology are treated as a queue sion/response analysis, constraint analysis, and visualization.
(Figure 4). Every queue stands as an individual unit with its The accepted scenario from the model output is further
own specific properties. The queue links adopted in DSMS analyzed for life-cycle impact evaluation (Jaafari, 1988).
are one of the unique innovations in process simulation Functions can be linked across the life-cycle phases within the
theory. The activity is the smallest piece of operation with a IFE environment to get a comprehensive overview of the pro-
finite execution time. A sequence of activities, i.e., an oper- ject’s status in real time. This objective-based approach will
ation, is initiated when an event occurs. One or more activ- allow planners to verify the impact of any changes on the sub-
ities transform the state of a system. A basic set of activities ject project. If LCOFs are not satisfied at the project level, the
constitutes the subprocesses and allows modeling the corresponding process must be re-evaluated for alternative
sequence of events in chronological order. solutions, and the procedure is repeated (Figure 3).
Define Resource
Yes Define Library
Process Clear? Processes No
No
Parts No Yes
Subprocess Exist? Unified LCOFs
Level 2
Project Satisfied?
Yes
Database
Process Clear? Yes
Define
No Subprocesses
Products
Level 3
Define Activity
Define Activity
Properties
Figure 5. The Three-Level Hierarchy Structure in the Dynamic Simulation Modeling System
IT Outlook. The object-oriented environment using 1991) provides all the necessary flexibility, speed and industry
C++/Visual C++ programming languages have been used for support during the design of the model.
development of the DSMS as well as the IFE system. The The object-oriented simulation model consists of a set of
DSMS is designed to set up automatically project-based interrelated object classes. It interacts with other objects
process models against instructions from the users and uti- through exchanging messages. An object can have defined
lizing information extracted from the IFE system. Interactive properties through which it is possible to construct a com-
Windows facilitate users to determine the association plex assembly of objects in order to define the processes of a
between each object class in the DSMS. Data exchanging and facility (Abdalla, 1991).
dynamic linking capabilities allow the DSMS to integrate The DSMS approach utilizes the discrete event simulation
into the IFE system. technique (Pidd, 1984) with process interaction approach
Furthermore, DSMS’s interfacing capabilities provide an (Garrido, 1999). In discrete event process simulation, entities
added facility for entering and modifying system inputs, representing products or services are induced into the model;
reviewing facility needs, redefining project scope, and produc- they compete for resources and perform the activities and are
ing reports and maps. A database management system discarded at the end (Figure 5).
(ObjectStore®) is being utilized to capture the entire project In the process interaction simulation approach, a process is a
and process data within the IFE framework. sequence of logically related activities in time. A process can be
represented as a class. The attributes of a process are represented
Process Simulation Framework as attributes of the class. The behavior of a process is modeled by
Object Oriented Technologies. The Booch Object-Oriented the operations to be performed on the process. The modeler
Analysis and Design method (Booch, 1993) is a suitable declares process objects and specifies their behavior via available
solution for design of the simulation model. Due to its object- methods. Thus, the integration of all the defined process objects
oriented nature, the C++ programming language (Stroustrup, into a single process unit provides the simulation model.
User Interface
Knowledge-based
Process Generation
Module
Module
Link Identification
Subprocess and Generation
Generation Module Module
Process Interaction Approach. Processes are the set of implementation lead to viable business solutions. For exam-
abstract data structures and entities carrying out the operation ple, the pulp handling system in a paper mill is considered a
sequences within the system (Figure 4). Activities are assigned major system. Design and specification of this system must be
to each subprocess to represent the system behavior, and each optimized in respect of the overall project vs. LCOFs.
process maintains its own list of activities. Subprocesses act as In the DSMS methodology, the entire model library (con-
nodes linked with interactive queue connections. Each sub- taining invisible elements or submodels) resides in the main
process contains the respective event time and activities in the project’s repository. Dialog boxes allow users to define relevant
system. The processes can be in one of the several states: rules, constraints and logical connectivity of all processes and
■ Active—When its activities are being executed; subprocesses. The knowledge-based system facilitates the syn-
■ Ready—When the process is waiting to start; thesis of the required process model automatically.
■ Idle—When the process is not active; Of course, the needed level of abstraction may have to be
■ Terminated—When the process has exhausted its action and explicitly specified or deduced from clues provided by the user.
is not going to be active again in the run under consideration. As an example, if a production engineer in a factory wants to
Thus, the process interaction approach in simulation mod- increase production by 50% due to market demand, the sys-
eling represents the dynamic behavior of a group of processes tem can synthesize an appropriate model automatically using
carrying out the operations and interactions with one another the information supplied by the user.
(Garrido, 1999).
Project and Process Model. DSMS is needed to support Process Interaction Implementation
proactive management of the project, particularly with respect Process Activity Interaction. The process model set up by
to LCOFs. The LCOFs must be evaluated not only for a select- DSMS consists of four major types of objects: processes,
ed major process, but for the complete project to ensure that resources, transactions, and queues. Processes use available
decisions made on the formulation, design, specification, and resources to provide the definitive results. Activities interact
inclusively within the processes. The process object allows rep- Transaction Modeling. Transactions are the entities that
resenting several subprocesses in multiple levels. Every process flow through the processes. Transactions can be used to repre-
and subprocess has a name and identification defined in the sent physical objects such as an order, a customer, telephone
model interface. calls, or work units. Every entity has its own properties such as
The total eight activities (generate, delay, assemble, release, a name, identification, arrival time, and service time. Users can
replenish, block, batch, and unbatch) define the unique define the state of transactions such as periodic arrival, cyclical
dynamic behavior of the processes and subprocesses arrival, and random arrival along with the generation points, as
(Figure 3). In the DSMS approach, the activities cannot be required in the real-life situation.
broken down. Every activity has the user-defined interactive Process Connectivity. Queue links are utilized among
properties of its own. Resources can be allocated from the processes and subprocesses for logical connectivity in the
resource library for each process vis-à-vis the activity. The sim- DSMS framework. These are used for defining the workflow.
ulation engine manages the processes in the object database Queues are modeling objects with their own characteristics
by carrying out three main tasks: process placement, process such as a name, identification, maximum queue size, priority,
removal, and process rescheduling. The engine keeps track of etc. More than one queue is used to define the concurrent
the simulation clock and decides which process is for imme- processes. Users can define the logical links by selecting prede-
diate activation based on the process connectivity (Pidd, cessor and concurrent processes in the process interfaces. A
1984). For the sake of brevity, details of the activities have not process that has multiple incoming and outgoing queues can
been discussed in this paper. be modeled using multiple queue links. DSMS allows collect-
Resource Modeling. Resources are the agents that are ing the queue statistics during and after the simulation run.
required in performing a process or subprocess through an
activity. The performance of the dynamic processes usually is Proposed DSMS Architecture
constrained by the limited availability of the resources or Components of DSMS. Figure 6 illustrates a conceptual archi-
resource interdependencies, which form queues in the system. tecture of the simulation model. The DSMS comprises four
DSMS modeling capability allows defining type, identification, broad subsystems:
capacity, cost, schedule, and allocation for resources. The sim- ■ Process Generation Module. This model facilitates the
ulation engine keeps track of resource utilization and queu- hierarchical process generation following the top-down
ing/delays for each process during the simulation run. Users approach. The processes are broken down into subprocesses in
can select the resource schedule and resource interruption the next level of the hierarchy model. The progressive break
options in the resource interfaces. down continues until a sufficient level of details can describe
Figure 9. Process-Process Relationships and Queue Link Mechanism in the Dynamic Simulation Modeling System
the behavior of the subprocesses. Activities in the subprocesses and subprocesses in the overall model. The simulator manages
are defined in a different interface, and resources are allocated and runs the relevant processes with input and output require-
as required. The resource library maintains the overall resource ments. The simulator maintains the simulation clock. It pro-
data. The information on process data, site condition, vides a set of operations that can be called or invoked by the
historical data, production data, product properties, and simulation model. Functionality and operability offered by dif-
design constraints are identified during the model definition. ferent scenarios are observed and optimized based on the
Different scenarios are combined with various constraints dur- model behavior and simulation run statistics.
ing simulation model construction. ■ Knowledge-Based Intelligent System. The knowledge-based
■ Simulation and Optimization Module. The simulation intelligent system facilitates the establishment of logical links
engine facilitates systematic sequencing of different processes between processes and subprocesses based on the expected
This article is copyrighted material and has been reproduced with the permission of PMI. Unauthorized reproduction of this material is strictly prohibited.
Sender Receiver
Media
they are concerned primarily with describing the selection of In another well-known approach, McGrath and Hollingshead
an appropriate mechanism, i.e., medium or channel, to (1994) characterized communication by the task being
encode, transmit, and decode the message. accomplished in addition to classifying tools according to
Categorization of communication mechanisms is impor- dimensions of contiguousness and synchronization. They
tant as a means for describing the selection of the most described e-mail as a distal, asynchronous technology, which
appropriate channel given the requirements of a particular may cause the natural order of messages to be disrupted. They
message (Robbins, 1998). The selection of a communication considered text-based e-mail a poor fit for tasks of increasing
mechanism also can influence the quality of information information requirements based on its relatively low richness.
received. The relationship between communication mecha- Technologies judged more appropriate for such tasks were
nisms and effectiveness is recognized presently in the field of audio or video systems and face-to-face systems.
knowledge management, where viscosity, or channel rich- Limitations of Traditional Taxonomies. One limitation of
ness, is used to describe the ability of a communication classifying Internet tools by time and distance is the decreasing
mechanism to transfer information-rich content (Davenport relevance of those factors in the contemporary business work-
& Prusak, 1998). place. When groupware taxonomies were introduced in the
Some important research in the use of electronic tools for early 1990s, integration of local area network technology and
group communication preceded the widespread organization- Internet communication was not common. (Note that elec-
al use of Internet applications by several years. The tronic networks are identified as same place/same time in
taxonomies created by these researchers generally categorized Figure 2, while Internet tools are categorized as different
communication tools in terms of time/synchronization and place/different time.)
place/distance. However, in the late 1990s, Internet applications, particu-
First, Johansen, Sibbet, Benson, Martin, Mittman, and Saffo larly e-mail, became very popular business communication
(1991) defined electronic communication tools used by tools, even when traditional methods might be better suited
groups as same time/same place, same time/different place, due to factors of task, location, or relative effectiveness
different time/same place, or different time/different place (Jefferson, 2000). The present popularity of e-mail and
(Figure 2). Electronic tools for same-time/place communica- intranet-based Web technology has contributed to the
tion were considered synchronous because interactive commu- realization that electronic networks and Internet tools
nication occurs in real time. Theoretically, such communica- currently are considered anytime/anyplace technologies—a
tion therefore can substitute for face-to-face conversation. With concept considered futuristic by Johansen et al. (1991).
asynchronous, i.e., different-time communication, one party Internet tools that are useful in project management
cannot respond until the other has finished sending. E-mail is communication include e-mail, static Web sites, group-
an example of asynchronous communication. ware, conferencing, and discussion group applications
In different-place communication, geographic separation (Giffin & Stankosky, 2000). E-mail, Web sites, and
prohibits face-to-face communication even if the participants discussion applications use client/server technology, while
are communicating simultaneously. Conferencing and other conferencing applications may be implemented in peer-to-
electronic technologies for remote communication are exam- peer or client/server configurations. Groupware is
ples of different-place tools. deployed on the Internet through database-backed Web
With same-place technology, participants either use the servers that also may implement discussion and confer-
same electronic tools synchronously or asynchronously, such encing applications or employ other Internet technologies
as with an electronic meeting room (synchronous) or a work- such as file transfer. The characteristics of these underlying
station used in shifts (asynchronous). technologies have distinct implications for the types of
ou e
on im
s)
discussed in Table 1.
hr T
nc nt
sy re
(A iffe
Proposed Framework for
D
Categorization of Internet Applications Voice Mail
Requirements for Categorization Scheme. Like the E-mail
no e
ro im
)
us
approach of the traditional communications model, there
ch T
yn e
are advantages to making the message the central compo-
(S Sam
nent of a taxonomy for Internet communication in project Two-Way
management. Primarily, this strategy allows communication Video Team Rooms
to be defined independent of the project and its context, Shift Work
Screen
therefore creating the possibility that the new taxonomy can
Sharing
define many types of communication, all of which may be
e
Electronic
ac
used at various times in a project. In addition to the mes-
Pl
Networks
nt
sage-centricity of the traditional communications model, the
re
ffe
concept of synchronization from Johansen et al. (1991) also PC Projectors
Di
is useful. Notwithstanding the contemporary notion of
e
Internet tools anyplace/anytime, there are still important
ac
Pl
differences between the tools in terms of their ability to facil-
e
m
itate real-time communication.
Sa
However, the traditional taxonomies did not address other
relevant attributes of Internet applications and project
Source. Johansen, R., Sibbet, D., Benson, S., Martin, A., Mittman, R.,
management, e.g., the use of Internet applications in the and Saffo, P. (1991). Leading business teams: How teams can use
send-only mode and the suitability of certain applications for technology and group process tools to enhance performance. Boston,
this specific task, such as static Web pages. This consideration MA: Addison-Wesley Publishing Co.
can be addressed by including one-way communication as a
distinct class, the appropriateness of which is supported by Figure 2. Categorization of Electronic Tools
the project management function of disseminating informa-
tion or providing reference material.
Another consideration that should be addressed is the people communicating with many. The categories on this
relationship between the utility of Internet tools and the axis may be defined rigidly or loosely depending on the
number of recipients of the message. For example, a Web capabilities of the tools being classified. If justified, addi-
page may be a better method for delivering information to a tional divisions on this axis could include one person
large diverse group than e-mail, if the sender does not have communicating with a few, a few people communicating
an effective means of maintaining an e-mail distribution list. with a few others, a few people communicating with many,
This and similar types of situations may be considered by and so on.
including the number of senders and receivers as a dimen- The framework of Figure 3 also is independent of distance
sion of the taxonomy. and location. For example, a group of people in one room
Last, the proposed taxonomy must have a level of detail could carry on many one-to-one conversations about different
sufficient for describing a full range of communication needs subjects. Because the subjects are different, each conversation
in project management while having only the minimum represents an individual message, and the communication
number of categories necessary to differentiate between the type would be synchronous one-to-one. However, if all were
available Internet tools. conversing freely about the same subject, the communication
Proposed Categorization Framework. The identified type would be synchronous many-to-many. In fact, both types
requirements include the synchronization between the of communication could occur simultaneously in one place
sender and receiver, the number of senders and receivers, the because it is the message not the location that determines the
possibility of one-way communication, and the ability to communication type.
describe each of these about an individual message. Nine categories of communication are defined by the basic
Incorporating these into a two-dimensional taxonomy pro- framework of Figure 3:
duces the framework of Figure 3. This scheme accounts for ■ One-Way 1:1. One sender sends to one receiver;
the consideration of one-way communication as a special ■ One-Way 1:Many. One sender sends to more than one
category of synchronousness. Combined with asynchronous receiver;
and synchronous two-way communication, a full range of ■ One-Way Many:Many. More than one sender sends to more
communication possibilities is provided on one axis. than one receiver;
On the other axis, the number of senders and receivers ■ Asynchronous Two-Way 1:1. One sender communicates
varies from one person communicating with one to many non-interactively with one receiver;
■ Asynchronous Two-Way 1:Many. One sender communi- the communication to include contextual cues and personal
cates non-interactively with more than one receiver; reaction. Therefore, it is a measure of the suitability of a tool for
■ Asynchronous Two-Way Many:Many. More than one sender problem solving or knowledge transfer.
and more than one receiver communicate non-interactively; The other axis, which represents the number and relation-
■ Synchronous 1:1. One sender communicates interactively ship of senders and receivers, is important from both techno-
with one receiver; logical and management viewpoints. From the standpoint of
■ Synchronous 1:Many. One sender communicates interac- management, an increase in message recipients implies an
tively with more than one receiver; opportunity to reduce cost and risk by exploiting electronic
■ Synchronous Many:Many. More than one sender commu- communication. For example, the Internet is used in con-
nicates interactively with more than one receiver. struction projects to reduce cost associated with revising and
Relevance of Proposed Framework to Project Management publishing blueprints (Tedeschi, 2000). The cost savings from
Communication. The two axes of Figure 3 encompass several this activity will consequentially be more attractive as more
important criteria from the perspectives of Internet applications people need a set of plans. Similarly, the risk of someone not
and project management. First, the axis that describes the syn- receiving an important update will be reduced if all use a
chronization of senders and receivers also describes the ability of common system.
One-way
Two-way/
Asynchronous
Two-way/
Synchronous
From a technical standpoint, the cost of implementation real-time communication. Consequently, they are the most
and support also is related to the number of parties involved. appropriate tools for each type of synchronous communication
For example, if only a few need to communicate, e-mail may in the model. Conferencing applications may support text,
be the least expensive form of communication. However, if the audio, or video transmission, however audio and video
cost of implementing and maintaining a Web site can be transmissions are best suited for small groups with few con-
spread out over many, the Web becomes a more efficient mech- nections due to bandwidth requirements. Additionally, as the
anism for providing information. number of active participants in video and audioconferencing
grows, it becomes more difficult for individuals to follow the
The Completed Taxonomy conversation. Text conferencing is less vulnerable to this
Categorization of Internet Applications. To populate the limitation because all contributions to the conversation are
taxonomy, the characteristics of Internet tools in Table 1 displayed in the same window and participants can scroll
must be considered within the bounds of the framework of upward to view previous comments. Therefore, text con-
Figure 3. The optimal position of each tool in the matrix is a ferencing should be better suited than videoconferencing as
consequence of two factors. First, the technological charac- the number of participants increases;
teristics of some tools may prevent their inclusion in certain ■ E-mail and Web Communication. Compared to confer-
areas. For example, the only tools that facilitate real-time encing, e-mail and the remaining Internet applications are not
communication are conferencing technologies. The others real-time. However, by virtue of their lack of synchronousness,
are either one-way or asynchronous-only mechanisms. they allow the user to review and respond to messages at the
Second, the efficiency with which certain tools can be used most convenient time, thereby increasing efficiency. E-mail is
in different organizational environments should be considered. well suited for individual one- and two-way communication
For example, although it is possible to use e-mail to provide and may be used effectively with larger groups, particularly if
information to a large number of people, if the information is all participants are members of the same organization.
on a Web site, the sender is not burdened with keeping track of However, as the number of e-mail participants increases, the
e-mail addresses. Additionally, if the recipients are unknown to challenges of maintaining large distribution lists and support-
the sender, the Web is the only mechanism that is suitable. ing individual e-mail clients and external programs for viewing
Conversely, if two-way communication is required in large attachments becomes a burden. In particular, Web-based com-
groups, a Web-based groupware system that allows participants munication is more effective when a group is not associated
to subscribe to distribution lists may be the most effective. with a single organization. Unlike static Web sites, which are
The justification for categorizing the Internet applications primarily suited for one-way communication, database-backed
based on technical and management aspects (as summarized sites can implement two-way communication protocols and
in Table 2) is: provide customized content, security, and user authentication.
■ Conferencing Applications. Conferencing applications are Because they are more efficient than conferencing applications,
the only tools in Table 1 that are capable of supporting e-mail and Web communication are more appropriate for
Conferencing (Text) Useful for interactive communication between larger/more diverse groups where
audio/videoconferencing are not practical.
asynchronous and one-way communication. Web communi- communication, is useful for describing the case in which the
cation is preferable as the number of participants increases. In participants share some sort of relationship, e.g., they belong to
the case of one-way Web communication, static pages are most a common organization. For example, a common association
efficient and are preferable. Database-backed sites are required allows technologies such as e-mail to be used effectively,
for two-way communication; whereas unassociated groups would find e-mail less effective
■ Discussion Groups. Compared to Web and e-mail com- due to the absence of a common directory of addresses.
munication, discussion group applications allow many The completed taxonomy is presented in Figure 4. The areas
senders to collaborate on the content of the message. Like e- where each Internet application was deemed most appropriate
mail, discussion group applications are message-centric, i.e., are depicted with shaded ovals. Areas where the ovals overlap
participants in the conversation can view a list of replies to depict the expectation that more than one application may be
each individual message and respond to any comment as they used effectively.
choose. Unlike e-mail applications, messages are not delivered
automatically to a user’s in-box. However, as the number of Discussion
users increases, this attribute becomes less inconvenient Application of the Taxonomy. The completed taxonomy
because discussion group applications allow users to select should be viewed as a tool for structuring project communica-
messages to be read before their contents are retrieved. Last, tion systems based on the synchronization and relationship
discussion group applications allow senders to provide infor- characteristics of project communication requirements. For
mation or submit responses without knowing the identities of example, if the ability to collaborate by asking questions and
the other participants in the communication. For this reason, reviewing others’ results is important to the success of a pro-
discussion group applications are well suited for communica- ject, then technologies suited for asynchronous many:many
tion in which the participants belong to many organizations. If communication should be provided. According to Figure 4,
the participants wish to restrict access to particular messages or discussion group technologies suit these requirements.
restrict communication to certain participants, discussion- If the ability to provide information to many parties is
group applications also can be implemented through important to the success of the project, Web sites should be
database-backed Web sites that can provide both security and used for maximum efficiency. Conversely, if most of the project
authentication. Based on their ability to support collaboration, communication is expected to be 1:1, the use of e-mail should
discussion groups are most appropriate for many:many be encouraged and resources made available to facilitate
communication, except in real time. However, e-mail and Web audio/videoconferencing.
sites are more convenient and, therefore, more appropriate Table 3 demonstrates a hypothetical analysis that employs
when there is only one sender. the framework of the taxonomy to categorize project informa-
Population of the Taxonomy. To accommodate the attrib- tion elements. In Table 3, the dimensions of synchronization
utes summarized in Table 2, a subcategory of communication are defined in terms relevant to project communication. One-
must be added to the framework of Figure 3 between the 1:1 way communication includes providing information, as well
and 1:many classes. This additional class, represented as 1:few as the creation and maintenance of project documentation.
E-Mail Discussion
Groups
Two-way/ Web-based
Asynchronous GroupWare
Two-way/ Audio/Video
Synchronous Conferencing Text Conferencing
Two-way asynchronous communication is described as The taxonomy primarily identifies one application as opti-
offline communication to represent communication that might mal for each type of communication. However, this list implies
typically take place through channels such as regular mail, fac- that the application indicated is theoretically the best suited for
simile transmission, or interoffice memos. Synchronous com- that type of communication, not that other applications can-
munication primarily contemplates information that would oth- not be used successfully.
erwise be exchanged or created in meetings or on the telephone. Organizations that study the effectiveness of Internet com-
The steps involved in preparing an analysis (Table 3) munication may find that user satisfaction with Internet tools
include identifying project information elements, identifying increases when the tools are used within their realm of suit-
the personnel or teams responsible for creating or receiving ability as defined by the model of Figure 4. In the absence of
each element, and categorizing the relationship of those tools suited for specific tasks, individuals might substitute
involved as 1:1, 1:few, 1:many, or many:many. another mechanism of electronic communication that con-
The project information elements then should be plotted tributes less effectively to project outcomes. Therefore, firms
within the framework of the taxonomy, and Figure 4 should be also may benefit from providing a full complement of Internet
employed to determine which Internet technologies are most tools for project management communication.
applicable for the various project information and comm-
unication requirements. The completed table of information Conclusion
elements also could be used in conjunction with the project A taxonomy has been created to describe the ability of
organization chart and schedule to identify the number of Internet applications to support different types of communi-
users and the phases of the project for which various Internet cation in project management. Because the taxonomy
technologies might be employed most effectively. defines communication based on the requirements of an
Further Development. At present, the taxonomy meets the individual message, it can be used as a model for using
identified needs for project management communication multiple types of Internet applications with multiple types of
based on the available Internet applications. However, it also communication in any given project.
can be expanded if the availability of applications supports The axes of the taxonomy describe the numbers of
additional categories. For example, if communication applica- senders and receivers and the synchronization between
tions are dissected further according to their ability to carry them. Special cases of communication are used to describe
information-rich content, a third axis of viscosity could be one-way communication and to differentiate between com-
added to the scheme. In that case, e-mail with multimedia munication with groups under common association with
attachments could be considered a highly viscous mechanism the sender and communication in groups without common
for 1:1 asynchronous communication. Messaging that does not association. The taxonomy differs from previous categoriza-
support attachments would rate lower on the viscosity scale. tion schemes in that it supports the classification of Internet
tools as anytime/anyplace technologies and expands the Enghavanish, S. (1999). Analysis of the effectiveness of
concept of group communication to include one-way communication and information sharing in virtual project
dissemination of information. organizations. (Doctoral dissertation, The George
The taxonomy uses a graphical approach to depict the effec- Washington University, 1999).
tiveness of each application, with the region of the taxonomy Giffin, S.D., & Jefferson, T. (2000). Conducting project
encompassed by each Internet application defining the theo- management over the internet. Proceedings of the 21st Conference
retical realm of its maximum effectiveness. Applications do not of the American Society of Engineering Management, Washington,
overlap except for the case of 1:few communication. DC. Rolla, MO: ASEM.
Otherwise, only one application is considered to be a best fit. Giffin, S.D., & Stankosky, M. (2000). Internet applications
By analyzing the communication requirements of the for project management communication. Proceedings of the 21st
project management team and matching those requirements Conference of the American Society of Engineering Management,
with the appropriate communication applications, organiza- Washington, DC. Rolla, MO: ASEM.
tions may improve their ability to meet their goals for effec- Guss, C.L. (1998). Virtual project management: Tools
tive technology use. This taxonomy may serve as a theoreti- and the trade. Project Management Journal, 29 (1), 22–30.
cal model for project management organizations seeking to Jefferson, T. (2000). Virtual project teams in an academic
improve their effectiveness and efficiency. setting. Proceedings of the 21st Conference of the American
Society of Engineering Management, Washington, DC. Rolla,
References MO: ASEM.
Davenport, T.H., & Prusak, L. (1998). Working knowledge. Johansen, R., Sibbet, D., Benson, S., Martin, A., Mittman, R.,
Cambridge, MA: Harvard Business School Press. & Saffo, P. (1991). Leading business teams: How teams can use
Doherty, P. (1999). Site seeing. Civil Engineering, 69 technology and group process tools to enhance performance. Boston,
(5), 38–41. MA: Addison-Wesley Publishing Co.
This article is copyrighted material and has been reproduced with the permission of PMI. Unauthorized reproduction of this material is strictly prohibited.
To minimize work in progress, a precedence-feasible sched- path leading into it. Resource buffers (RB), usually in the form
ule is constructed by timing activities at their latest start of an advance warning, are placed whenever a resource has to
dates based on critical path calculations. If resource conflicts perform an activity on the critical chain, and the previous crit-
occur, they are resolved by moving activities earlier. The ical chain activity is done by a different resource.
critical chain then is defined as that chain of precedence and During project execution, both the critical chain and the
resource dependent activities that determines the overall baseline schedule should be fixed. Project activities are execut-
duration of a project. If there is more than one critical chain, ed according to the roadrunner mentality using an unbuffered
just select one. projected schedule. This schedule is early start-based, except for
The safety time that is eliminated from the critical chain activ- the gating tasks (activities without predecessors), which are to
ity durations by selecting aggressive duration estimates is shifted be started at their baseline scheduled start times. Early com-
to the end of the critical chain in the form of a project buffer pletion of activities must be reported, and activities should be
(PB). This PB should protect the project due date promised to started as soon as work becomes available. The execution of
the customer from variability in the critical chain activities. the project is managed by using the buffers as a proactive
Feeding buffers (FB) are inserted whenever a noncritical warning mechanism.
chain activity joins the critical chain. Their aim is to protect the As activities are completed, project management should
critical chain from disruptions on the activities feeding it and keep track of buffer consumption. As long as a predeter-
to allow critical chain activities to start early in case things go mined portion of the buffer remains [buffer consumption is
well. Although more detailed methods can be used for sizing only in the green (O.K.) zone], everything is assumed to go
the buffers (Newbold, 1998; Product Development Institute, well. If buffer consumption moves beyond a certain point, a
1999), the default procedure is to use the 50% buffer sizing warning is raised (the yellow watch-and-plan zone). If it
rule, i.e., to use a PB of half the project duration and to set the deteriorates past a critical point (the red act zone), corrective
size of a FB to half the duration of the longest noncritical chain action must be taken.
Units
3 6 3 5 6
2 7
2 2 8
1 5 4 1 4
7
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9
Time Time
Figure 2. Minimum Makespan Schedule for the Network Figure 3. Schedule Obtained Using Microsoft Project
of Figure 1
the initial baseline schedule. Goldratt (1997) suggests using from the ones present in the optimal duration schedule of
the median, while the Product Development Institute (1999) Figure 2: the chain 4–7 and the chain 5–6–7.
argues in favor of the mean. Herroelen and Leus (2001) con- If the project planner would rely on the ProChain software,
clude from their full factorial experiment that the use of the the schedule in Figure 4 would be obtained. The schedule now
mean activity duration provides the safest estimates of the has a length of eight time periods and has two critical chains: the
project duration. chain 3–5–8 and the chain 3–6–7. Both activity 2 and activity 4
To illustrate that the length of the baseline schedule and the have been right-shifted, i.e., scheduled as late as possible within
collection of possible critical chains is entirely dependent on the scheduling horizon determined by the critical chains. The
the scheduling procedure used, consider the example network software selects the chain 3–6–7 as the critical chain.
given in Figure 1 (Demeulemeester, Herroelen, Simpson, If the user relies on PS8 for generating the baseline sched-
Baroum, Patterson, & Yang, 1994). The numbers above each ule, the optimal seven-period schedule of Figure 2 results, with
node denote the activity duration of the corresponding activi- the critical chains 2–6–7, the chain 2–6–8–3 (8 and 3 have
ty (node 1 and node 9 are dummy nodes with zero duration). switched position), and the chain 5–4. The software selects the
It is assumed that these activity duration estimates do not critical chain 2–6–7.
include individual safety provisions. The numbers below each In short, four different baseline schedules have been
node of the network in Figure 1 represent the number of units obtained for the same project, ranging in duration from seven
of a renewable resource, e.g., workers, required during every to nine periods, and each providing the planner with a differ-
period of the corresponding activity’s duration interval. It is ent population of critical chains. Goldratt (1997) contends
assumed that the renewable resource has a constant availabili- that it does not really matter which critical chain is chosen.
ty of five units per period. This simple example proves the fact that generating a good
Figure 2 gives a minimum duration schedule. This can be baseline schedule does matter, and that statements such as “the
obtained, for example, by using the branch-and-bound pro- critical chain is never ambiguous” (Leach, 2001b) must be
cedure of Demeulemeester and Herroelen (1992, 1997). As interpreted with sufficient care. The length of the baseline
can be verified easily, the project duration is seven time peri- schedule, the set of candidate critical chains, and the activities
ods (by poor luck this corresponds to the length of the in the chosen critical chain all depend on the procedure used
longest critical path 1–2–6–7–9). The baseline schedule for generating the baseline schedule.
reveals three critical chains: the chain 2–6–7, the chain Adding a 50% integer length PB (according to CCS/BM
2–6–3–8, and the chain 5–4. logic) to the baseline schedules of Figures 2, 3 and 4 would
Figure 3 shows the baseline schedule generated by inspire management to make the customer promise to
Microsoft Project. This baseline schedule has a length of nine deliver the project in at least 11, 14 or 12 periods, respec-
time periods and reveals two possible critical chains, different tively. These should be interpreted as “lower bounds:” The
five periods and requires two units of the renewable resource activity 5, the buffer FB6 should be inserted after activity 6,
type during each of its execution periods. and the buffer FB7 should be placed after activity 7. If a
Figure 8(a) shows a minimum duration schedule for the person applies the 50% rule for buffer sizing and assumes
project. The critical chain is identified as chain 1–2–3–8 with integer length buffers, FB5 should have a length of four
a 17-period duration. As indicated in Figure 8(b), three FBs periods, FB6 should have a length of three periods, and FB7
must be inserted: The buffer FB5 should be inserted after should be two periods long.
1 2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Time
3
FB5
7 FB6
FB7
Units
2
6
1 5 2
4 3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time
3 FB5
7 4 FB6
FB7
Units
2 5
6
1 2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Time
6 4 2
1 4 8
A, E B, E E
3 2 1
5 6 7
F B, E C
same resource with unit availability; the same holds for D and schedule shown in Figure 11(b), which has a smaller
E. Resource usage of the other activities is not restrictive. One makespan. However, the new projected schedule actually is an
possible critical chain would be B–E–H–F–I. The correspond- implicit recognition of a new critical chain. It would be logical
ing buffered baseline schedule is shown in Figure 10(b). The to identify this chain and insert/resize the buffers accordingly.
unbuffered projected schedule is depicted in Figure 10(c). This also allows full exploitation of the updated PB size in
Suppose that at the foreseen end of activity D, this activity making reliable project completion time estimates. Similar
is perceived to take longer than expected. Keeping the critical observations were made as a result of the simulation experi-
chain unchanged would yield the new projected schedule ments in Herroelen and Leus (2001): If the critical chain must
shown in Figure 11(a). Clearly, the makespan of the projected be kept in series during project execution, the makespan can be
schedule has increased. improved by regularly reevaluating the “tightness” of the criti-
Assume that the critical chain no longer is required to cal chain, or in other words, one should avoid concentrating
remain in series. In this case, it no longer is necessary to sched- on an ex ante-derived critical chain that is no longer the real
ule activity H in front of activity F. The result is the projected constraint on project makespan.
This article is copyrighted material and has been reproduced with the permission of PMI. Unauthorized reproduction of this material is strictly prohibited.
Journaling: A Learning Tool for
Project Management Training
and Team-building
Robert Loo, The University of Lethbridge, Faculty of Management, Lethbridge,
Alberta T1K 3M4 Canada
T eamwork and managing teams can be very stressful because of interpersonal and
task conflicts, the pressures of deadlines, performance standards, and resource
scarcity, among other disagreements that typically threaten team effectiveness and
success. Project managers control such threats to team effectiveness partly through
measures such as timely training in leadership skills, conflict management skills,
communication skills, and the like. Another method to manage such threats is
through the implementation of various administrative procedures and practices such
as team meetings and performance reviews.
This paper focuses on the usefulness of reflective learning journals as a method of
promoting both individual and team performance in the context of project manage-
▼ ment training and team-building (Fleming & Koppelman, 1997; Goodge, 1998;
Abstract Zimmerer & Yasin, 1998). Critical reflection upon project team experiences through
journaling should facilitate learning specific skills including interpersonal commu-
This paper describes the application of
reflective learning journals to promote critical nications, conflict management, managing effective meetings, managing stress, and
self-awareness. This technique improved both leadership skills. The nature of reflective learning, the use of reflective learning jour-
individual and team performance among nals, and their application in teamwork are illustrated along with recommendations
members in nine management undergraduate
teams conducting project management
for the effective use of journaling in project management.
research projects. Qualitative analyses of
entries in the journals revealed seven major Theory and Practice of Reflective Learning
themes: interpersonal relationships, team Boyd and Fayles (1983) define reflection as “the process of internally examining and
communication, stress-time pressures, team
vs. individual work, personal criticisms and
exploring an issue of concern, triggered by an experience, which creates and clarifies
improvement, lessons learned and future meaning in terms of self and which results in a changed conceptual perspective”
actions to improve one’s performance, and (p. 100). Reflection is embedded within several theoretical frameworks (Hutchinson
evaluation and comments about journaling. & Allen, 1997; Kember, Jones, Loke, McKay, Sinclair, Tse, Webb, Wong, F., Wong,
Analyses of the journaling evaluation data
showed that participants found journaling a M., & Yeung, 1999; McCaugherty, 1991; Riley-Doucet & Wilson, 1997; Scanlon &
useful learning tool. Recommendations are Chernomas, 1997). This paper used Scanlon’s and Chernomas’s (1997) three-stage
presented for using reflective learning journals model because it is a relatively simple model that easily can be communicated and
as one tool in project management training
used. As seen in Figure 1, the first stage of reflection is awareness, where awareness
and team-building.
might be stimulated by some uncomfortable or positive thoughts or feelings about a
learning situation or event. Without such awareness, reflection cannot occur.
Keywords: learning journals; team-building;
training In the second stage, the individual critically analyzes the situation, bringing to
bear his or her relevant knowledge and experiences as well as the application
©2002 by the Project Management Institute
2002, Vol. 33, No. 4, 61–66
of new knowledge resulting from the analysis process. This stage should involve
8756–9728/02/$10.00 per article + $0.50 per page critical thinking and evaluation and self-examination with accompanying growing
self-awareness.
Personal Compatibility. Other students commented on the ■ “This lack of focus was mainly due to our lack of knowledge
good fit of personalities or effective aspects of group interactions. of each other personally.”
■ “As far as personalities go, our personalities coincide Friends vs. Colleagues. While students may believe initial-
very well;” ly that working with friends will make projects easier, they
■ “[We] possess similar personalities, so we enjoy learning soon discover that is not always the case.
from one another;” ■ “I also reaffirmed my belief that having team members with
■ “We seem to be getting along great so far this semester and whom you also share a social relationship with is not always
I hope that it continues throughout the project.” conducive to getting things accomplished.”
Safe Environment. Some comments reflected a “safe envi- Interpersonal Conflict. Some students noted conflict with
ronment” team in that the team climate was supportive rather other group members.
than threatening. ■ “For the first time in my university career, I feel like get-
■ “What really worked well during this meeting was the fact ting mad every time I have a meeting in this group. It’s not
that everyone contributed their ideas, and no one criticized that I have personal differences with anyone, in fact [we]
ideas offered ... I have found that this group has helped me to are good friends. It’s just that we seem to argue over the
come out of my shell;” stupidest things.”
■ “There was [sic] no criticism or accusations made ... I believe Self-Awareness and Interrelationships. Some students rec-
that this cohesion produced a fantastic paper and a great ognized aspects of their own personality and interpersonal
rapport between each of us;” style that hindered constructive interrelationships. Some made
■ “I found it great that we all accepted each other’s feedback efforts to change their behaviors.
and made the suggested changes to our individual work.” ■ “At these meetings, I can’t help but feel that I’m taking
Humor. There were a few comments about the positive charge too much. A nasty habit of mine, I usually try to dom-
effects of humor in group meetings. inate the conversation ... I am concerned that I may be
■ “I have a good sense of humor and like to display that, offending my other group members, and I have tried to talk
especially in a group setting. This sense of humor helped with about it with my group members. Nobody has made any
loosening our group up at the beginning of our meetings.” complaints so far;”
On the other hand, there were negative comments about ■ “I have traditionally hated working in groups ... Having
interactions within the team. been let down in the past, I seem to have developed a bit of a
Interactions Lacking Focus. Some students noted the frus- pessimistic outlook when it comes to relying on other people
trations when group meetings lacked focus or were discussions for work that will directly affect me;”
about unrelated matters. ■ “I can get aggravated quite easily if I am under pressure
■ “Made it difficult to get any work done because of talking and if things don’t go the expected way I planned, so I need
about unrelated subjects, disagreements, and differing out- to take a deep breath, step back, and remember that my team
looks, which had to be treaded on carefully in order to keep members are in the same boat as I am with school, work, and
from hurting any feelings;” social lives.“
Note. Adapted from the Atkins and Murphy (1993) three-stage model of reflective learning.
Group Communication Negative Reactions to E-mail. While some students had pos-
Not surprisingly, there were many comments related to itive evaluations of e-mail as a communications tool, others did
communication within teams, including comments about not like using e-mail or found the communications frustrating.
both effective and ineffective communications. ■ “We discussed after our first team exercise that e-mailing was
Usefulness of E-mail as a Communications Tool. Several a bad means of communicating ... Some team members were
journal entries emphasized the usefulness of e-mail as a com- not as keen as others on using e-mail for correspondence;”
munication tool for the group assignment. ■ “She said she would e-mail us a draft and we could look it
■ “I sent an e-mail to all team members, getting folks to start over and send it back to her on Thursday. That sounded like a
thinking about the kind of project they wanted to do. We great plan, but Netscape was down most of Wednesday and
e-mailed back and forth two or three times and finally all Thursday so I couldn’t get my mail.”
confirmed the topic;”
■ “We were all up to speed because we kept in touch Stress-Time Pressures
through e-mail;” Some students commented on stress mainly due to time pressures.
■ “I learned that I love e-mail!” ■ “We realized we started too late on this paper, and we’re
Effective Communication Leading to Effective Meetings. starting to get snappy and crabby at and with each other—
Some students noted examples of effective group communica- unjustified actions;”
tion leading to effective meetings. ■ “From the beginning, we failed to get organized. Now we
■ “During this meeting we had excellent communication and are all in a panic trying to find the easiest way to produce a
focus. The meeting was not long—just under an hour—but we good paper;”
got a lot accomplished and I think that we all came away from ■ “I have so much other work to do that I can’t organize or get
the meeting with a good feeling;” other people going and coordinate the proceedings.”
■ “What really worked well during this meeting was the fact On a positive note, one student stated:
that everyone contributed their ideas and no one criticized ■ “After a good meeting like this, my stress level just dropped
ideas offered.” two notches.“
There also were instances of ineffective communications The distressful effects of time pressures also were mentioned:
noted in some journal entries. ■ “I’m worried that we don’t have enough done yet, and I’m
Off-Task Communications. Some students noted the frus- worried that we are not going to finish this in time. We need to
trations of group meetings in which they felt that too much get going, but everyone’s schedule is so limited;”
time was spent on discussions not relevant to the task—the ■ “I think we may have been feeling the pressure because
group term paper. there is [sic] only two weeks before the paper is due.”
■ “We really haven’t done that much together, and when we
do get together, we tend to talk a lot about unrelated topics.” Team vs. Individual Work
Ineffective Communications. There were journal entries While the assignment was a group paper, some students
that noted ineffective communications in the sense that commented favorably on the ability to divide the work
communication was not focused, ambiguous, and the like. among members, or in other words, the ability to do indi-
■ “I think we must have been on different wavelengths the vidual as well as group work. Other students commented
evening we chose our topic. It was very different than what we that the group assignment allowed individual work too;
talked about;” therefore, students could break the project into individual
■ “Today we discussed our problem, and I realized that it was tasks and coordinate the work.
pretty bad. Ironically, by talking about it, we got even further ■ “The ability for the individual team members to work
off track without addressing the real problem. I guess it would individually, yet still [work] towards [sic] the common goal
help if we had an agenda and appointed someone to be the of the group saved everyone a lot of unnecessary meeting
watch dog for each meeting.” times and headaches in the end.”
Conclusions
Findings from this study show that reflective learning journals,
with the guidance and support of instructors, are useful tools
that can help staff learn as individuals as well as members of Robert Loo, PhD, is professor of management at the
teams. Significantly, this group of students provided evidence University of Lethbridge. Before joining the university
demonstrating they learned that identification of their short- in 1989, he worked extensively in both private and
comings is important and that it is essential to take steps to public sector organizations since the 1960s. Loo has
change behaviors to work compatibly within teams. The reflec- held both project and program management positions in organiza-
tive learning journals stimulated students to contemplate how tions such as the Canadian Department of National Defense, the
they would fit in as a contributing member of a team. Systems Division at Bell-Northern Research, and the federal Public
Service Commission of Canada. He has published more than 100
References journal articles and book chapters over the past 25 years.
Boyd, E.M.E., & Fayles, A.W. (1983). Reflective learning: Key
to learning from experience. Journal of Humanistic Psychology, 23
(2), 99–117.
Fleming, Q.W., & Koppelman, J.M. (1997). Integrated
project development teams: Another fad ... or a permanent
change. Project Management Journal, 28 (1), 4–11.
Goodge, P. (1998). How do we make management devel-
opment effective? Journal of Management Development, 17
(1), 83–87.
Hutchinson, D.J., & Allen, K.W. (1997). The reflection inte-
gration model: A process for facilitating reflective learning. The
Teacher Educator, 32 (4), 226–234.
Kember, D., Jones, A., Loke, A., McKay, J., Sinclair, K., Tse, H.,
Webb, C., Wong, F., Wong, M., & Yeung, E. (1999). Determining
the level of reflective thinking from students’ written journals
using a coding scheme based on the work of Mezirow.
International Journal of Lifelong Education, 18 (1), 18–30.
McCaugherty, D. (1991). The use of a teaching model to
promote reflection and the experimental integration of theory
This article is copyrighted material and has been reproduced with the permission of PMI. Unauthorized reproduction of this material is strictly prohibited.
Cover to Cover
Book Review Editor, Kenneth H. Rose, PMP
Managing Multiple Projects: Planning, scheduling, resources, and reporting, he provides a clear path
Scheduling, and Allocating Resources for to improved productivity.
Competitive Advantage In Part II, Mike McCauley, Ann Bundy, and William
edited by James S. Pennypacker and Lowell D. Dye Seidman debunk the myths of resource management.
Conventional wisdom has it that resource shortages may be
T he project management
office (PMO) is a matter
of more frequent consider-
but they provide a framework and foundation that does much to
remove the mystery associated with project organizations.
The discussion of maturity models shows that PMOs need
ation and cautious application. not leap to life fully formed. They may be tailored to fit the
Different views exist as to just needs and project maturity of the organization. The authors link
what a PMO is and the various maturity to metrics, showing how a PMO can play a lead role in
views have their own propo- organizational performance measurement by identifying key
nents and detractors. Parviz F. project and process issues, selecting and defining corresponding
Rad and Ginger Levin have metrics, and integrating the metrics into existing processes.
made some sense of this situa- Guidance for implementation begins with a summary of
tion with their highly rational PMO functions. An effective PMO performs functions that assist
and readable new book The projects and support organizationwide project management.
Advanced Project Management Rad and Levin describe 17 separate project-focused functions
Office: A Comprehensive Look at and 11 enterprisewide functions and areas of performance.
Function and Implementation. The authors’ approach to implementing a PMO addresses
In the introduction, the authors differentiate between a pro- motivation, cost and benefits, current practices, and organi-
ject office, which is limited in application to a specific project zational structure. They offer a six-step life cycle for imple-
and ceases to exist when the project ends, and a project man- mentation and discuss deployment of the PMO relative to its
agement office, which is enterprisewide in scope and exists sponsorship and intended influence.
independent of individual project duration. The book closes with a brief but contemporaneously rele-
The substance of the book stands in two parts: elements and vant discussion of the PMO and professional responsibility.
implementation. In four brief chapters, Rad and Levin address The PMO can provide mentoring, consulting, and guidelines
project performance facets, marginal project performance, pro- in areas of professionalism, integrity, and diversity.
ject manager competency, and maturity models. This all estab- Project management offices are not likely to replicate in
lishes a foundation. They follow with two chapters that address rapid cookie-cutter fashion. Divergent issues and interests
functions of the PMO and implementing a PMO. affect both startup and operation. The Advanced Project
The authors describe determining project performance as a Management Office does not prescribe a solution. Instead,
critical role of the PMO, but do not prescribe a specific, uni- it offers a solution method that may be applied flexibly accord-
versal set of attributes and ratings. Rather, they offer a method- ing to the need and desires of those who would improve
ology for formalizing and documenting attributes and ratings organization project management performance.
that responds to different needs among different projects.
They explain deliverables-focused client viewpoints and St. Lucie Press, 2002, ISBN: 1–57444–340–2, hardcover, 224
means-focused project team viewpoints in a clear, logical way. pp., $56.95 member, $59.95 regular.
Their summary of project evaluation criteria is clear and com-
plete and supports the implementation of an honest-broker Reviewed by Ken Rose, PMP, an instructor for ESI
element such as a PMO to apply, evaluate, and document the International residing in Hampton, VA, USA, and vice
criteria and results. president for programs, PMI Hampton Roads Chapter.
Correspondence
The Current State of Project Management vast array of government and educational institutions in
Research: Trends, Interpretations, and Predictions the United Kingdom, Australia, South Africa, India, etc., all
I have somewhat belatedly finished reading “The Current in English.
State of Project Management Research: Trends, My complaint is not so much that these other resources
Interpretations, and Predictions” [June 2002, Project have not been researched—the cost and workload would be
Management Journal, p. 5] and was amazed to find that the enormous—rather the exaggerated claim of having researched
authors (T.J. Kloppenborg and W.A. Opfer) seem to be all articles published in “English.” The more accurate descrip-
unaware of the fact that English is spoken in places outside tion of the research would be “a limited range of articles pub-
of the North American continent, in particular the United lished in the USA/Canada.”
Kingdom! Their claim to have reviewed all of the research —Patrick Weaver
published in English from 1960 to 1999 is patently false.
The major easily accessed information resource ignored
in the paper is the International Journal of Project Authors’ Response:
Management, published by Pergamon Press (ISSN There are a couple of minor points on which we feel that Mr.
0263–7863). This journal has, since the early 1980s, only Weaver has merit, but we disagree with most of what he asserts.
published refereed project management articles with a high We would first point out that, while the article certainly
proportion referring to research projects. does imply an extensive search of the English language sources,
Kloppenborg and Opfer also have apparently ignored “This research study project covers research published in
the vast amount of published literature available in other English from 1960 through 1999,” it does not make any claim
publications including the Association of Project “to have reviewed all of the research published in English from
Management (United Kingdom) and Australian Institute 1960 to 1999.” This is a somewhat semantic argument, but we
of Project Management, to mention only two, and from a think it is valid.
Index of Advertisers
Mindjet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C2
PMI Knowledge & Wisdom Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Primavera Systems Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C4