Journal of Business and Behavioral Sciences

Vol 33, No 1; Spring 2021

A TWO-STEP APPROACH IN PROJECT

PORTFOLIO SELECTION

Kamy Farahbod
Jay Varzandeh
California State University, San Bernardino

ABSTRACT

There is a fundamental difference between project management and project

portfolio management that has been recognized by all project managers and
the Project Management Institute (PMI) for many years. While project
management is about doing the selected projects right, project portfolio
management helps us to do the right projects. Given the mixture of many
qualitative and quantitative factors/concerns, such as the returns, risks, and
strategic issues, selecting the right portfolio could be a challenge for project
managers and the Project Management Office (PMO). The purpose of this
paper is to identify the various methods that are currently being used to
create the best project portfolio, and to further propose an approach that has
not yet been considered by neither academics nor professionals for this
purpose. It is suggested that a two-step approach consisting of Data
Envelopment Analysis (DEA) and 0-1 integer programming will enable the
project selection committees to objectively evaluate the potential project
portfolios and select the ones with the highest level of efficiency.

Key Words: Project Portfolio Management, Data Envelopment Analysis,

0-1 Integer Programming, Optimal Project Portfolio

INTRODUCTION
Historically, business organizations have used elaborate approaches to
evaluate, prioritize, and select their portfolio of projects, as shown in Table
1. While these approaches range from simple judgements to sophisticated
schemes, their importance to overall success of the organization cannot be
overstated. According to PMBoK, the strategy of an organization is an
action plan to achieve its business goals and objectives, and it determines
the portfolio of projects and programs that the organization will execute
(Sangher, 2019). Most organizations form committees within their project
management offices (PMOs) that are commissioned to meet their strategic

109
Farahbod and Varzandeh

objectives and further increase the likelihood of realizing the desired return
on investment. Of course, not all projects have a strategic disposition and
are evaluated carefully, but for the projects that are vital to the competitive
performance of the organization, project portfolio management is necessary
in determining the priority of the projects and making limited resources
available for their completion.
Project selection models should be realistic, capable, flexible, easy to use,
cost effective, and easily computerized (Pinto, 2010). By realistic, we mean
that the model should reflect objectives and strategic goals of organizations.
Capability refers to the ability of the model to simulate different scenarios
and optimize the decisions. Flexible models can be modified for application
in the presence of changes. Ease of use indicates the model is simple and
understandable in all areas of the organization. Cost effective refers to the
low cost of data gathering and modeling relative to the project cost. And
finally, easily computerized suggests the model can be easily applied to
many projects and solved effortlessly using computer technology for
comparison purposes.
Table 1 represents the most widely accepted methods for project selection.
Each method presents a particular combination of advantages and
disadvantages that must be considered by the project selection committee.
Archer, et al. (2004, p 244) suggest “scoring models are probably the easiest
to use, and weighted factor scoring method has been shown to be preferred
in different kind of projects and industry sectors.” However, the importance
of using more comprehensive methods that allow for a certain level of
sophistication to evaluate the duration and the mix of vital strategic projects
is self-evident. One such approach is data envelopment analysis.
Using a multi-step process in the portfolio selection process is not new.
Keren et al. (2014) used a combination of Analytical Hierarchy Process
(AHP) and DEA to select a project manager; and additionally, Dia (2009)
used a 4-step process in order to select a portfolio of stocks. It is important,
however, to ensure that the steps are easily understood, supported, and
applied in a project environment where the stakeholders usually come from
different functions with various personalities and educational/technical
abilities.

DATA ENVELOPMENT ANALYSIS

Frontier analysis, also called data envelopment analysis (DEA) has been
used extensively in many business settings. Malhotra, et al., (2010) use
DEA to rate corporate bonds. Soheilirad, et al. (2018) provide an example
of the application of DEA in supply chain research. This method was first
introduced in 1978 by Charnes, Cooper, and Rhodes to measure the relative

110
Journal of Business and Behavioral Sciences

efficiency of decision-making units (DMUs) in organizations. Using DEA

and mathematical programming, the performance of peer units can be
evaluated. DEA examines the resources used by the DMU’s (inputs) and
their outputs. A DMU is efficient if no other DMU can produce more
outputs for the same or less quantity of inputs, or can produce the same
output for less inputs. Efficient DMUs are located on the efficient frontier,
enabling the analyst to identify the reasons for the inefficiency of the other
DMUs. DEA helps us reveal the weaknesses and strengths of the DMUs,
prepare the DMUs to meet their customers’ needs, and pinpoint the
opportunities to improve the existing operations and to create new products,
services, and processes. It is a non-parametric approach, and as such, has
the advantage of not assuming any particular shape or form for the efficient
frontier function. This, plus the flexibility of DEA in the treatment of the
inputs and the

Table 1 – Project Selection Methods

111
Farahbod and Varzandeh

outputs, makes it one of the most widely used methodologies in business

analysis. Most recent applications of DEA deal with measuring the
efficiency and profitability of supply chains (Amirteimoori, et al., 2011).
The purpose of this research is to identify the most efficient projects, given
the time and budget constraints, so that they are selected for the company’s
project portfolio. One important advantage of DEA over the typical
statistical approaches that compare DMUs to an “average” DMU is that
DEA compares each DMU with the “best” DMU, given the amount of
inputs and outputs for each DMU. DEA can also pinpoint specific
inefficiencies of each DMU. However, this paper uses this methodology for
the sole purpose of identifying the most efficient projects and there is no
need for analyzing the efficiency of the inefficient projects. Once the most
efficient projects are identified, given the numerical and nonnumerical
constraints, a 0-1 integer programming model can be used to identify the
best portfolio.

0-1 INTEGER PROGRAMMING

It is well known that “0-1 integer variables can be used to model situations
in which k projects out of a set of n projects must be selected, as well as
situations in which the acceptance of one project is conditional on the
acceptance of another” (Anderson et a., 2019). The general formulation of
these models includes the optimization of an objective function subject to a
series of constraints that include binary variables that assume an optimal
solution of “1,” when the selection of a project is optimal, and assume an
optimal solution of “0,” when not selecting a project is optimal. This level
of flexibility of the 0-1 integer programming enables the project selection
committees to successfully create optimal portfolios (Ghasemzadeh et al.,
1999).
In this paper, we use 0-1 integer programming for the purpose of finding
the most efficient portfolio among all possible portfolios, given the
constraints that the projects are facing.

PORTFOLIO SELECTION METHODOLOGY

As mentioned earlier, the proposed methodology is divided into two
sequential steps. In the first step, the relevant inputs and outputs of the
DMUs (projects) are pinpointed and preferences of the project selection
committee are specified. Then, using DEA, efficiency ratios of the DMUs
are calculated. In the second step, a 0-1 integer programming model
designed to identify the most efficient portfolio of projects, given the
organization’s constraints, is formulated and solved.

112
Journal of Business and Behavioral Sciences

This methodology provides both the simplicity of the modeling and the
solving processes, while also providing ease of understanding and
application for the project stakeholders. The data collection process for
DEA and 0-1 integer programming is done simultaneously, allowing the
methodology to generate a portfolio that is objective and based directly on
organization’s value system.

AN ILLUSTRATIVE CASE
For illustration purposes, we have chosen an IT company that is currently
considering 39 projects that mostly require labor hours for consulting,
software development and more (costs), other costs (including materials,
permits, licenses, software, training, etc.), and time (in weeks). All of these
resource requirements and times (inputs in DEA) are estimated in the
preplanning phase of the projects’ life cycles. Hopefully, the estimated
budget at completion (BAC) is less than the contract value for each project,
providing a reasonable profit. Of course, as is aforementioned, some
projects are selected based on considerations other than profitability, and
therefore, may result in losses. Projects 10, 12, 14, 15, and 36 in our case
are expected to be completed at a loss. However, the management insists on
doing the first two for operating necessity, competitive necessity, or other
reasons.

Step 1: The Most Efficient Projects

In the first step, the most efficient projects are identified. Table 2 lists all 39
projects and the resulting efficiencies, using DEA. Projects 3, 9, 18, 21, and
22 are on the frontier of efficiency and are the most efficient. However, as
mentioned earlier, other considerations such as the total costs, time, and
other factors play decisive roles and impose constraints in project portfolio
selections. A 0-1 integer programming analysis will enable the selection
committee to find the most efficient portfolio, given all constraints and
considerations.

Step 2: The Most Efficient Portfolio

Once the efficiency of the projects is determined, a 0-1 (binary) integer
programming model can be used to find the best portfolio of projects. Each
decision variable Xi can represent the inclusion/absence of a particular
project in the portfolio, with Xi =1 implying the inclusion, and Xi =0
implying the absence of the project “i” in the optimal portfolio. Of course,
all constraints challenging the selection process must be included in the
model. As mentioned above, money losing projects 10 and 12 are selected

113
Farahbod and Varzandeh

due to other considerations. However, there is no need to include projects

14, 15, and 36 as they will be completed at a loss.

Table 2 – Data Inputs and Output and Efficiency of Projects

114
Journal of Business and Behavioral Sciences

If we assume a total budget of $3 million has been allocated and the

maximum allowable time to finish a project is one year (52 weeks), then a
0-1 integer programming model that can maximize total efficiency of the
portfolio given these constraints will identify the selected project for the
most efficient portfolio, as shown in Table 3. Realistically, it is quite
possible to receive partial payments for completed parts of the projects; if
this is the case, then the constraints of the model would have to be
modified to account for the funds as they become available in future
weeks.

CONCLUSION
While project portfolio selection committees have quite a few alternative
means for their selection processes, a more objective and comprehensive
approach based on optimization of important factors, such as efficiencies,
can improve the quality of the selected portfolio. This paper proposes a
simultaneously broader and easier two-step approach to the project portfolio
process. As shown in the illustrative case, data envelopment analysis can be
used in the first step to identify the most efficient projects. Given the list of
these projects, a 0-1 integer programming model that includes all the
constraints and challenges facing the selection committee can be used to
maximize the efficiency of the portfolio.

Given the power of the optimization methods that are used in this paper and
the ease of their use, there is no doubt that this two-step approach should be
added to the list of the available methods in Table 1 to enhance and improve
the selection process.

115
Farahbod and Varzandeh

Table 3 – The Portfolio of the Most Efficient Projects

Project Total Estimated Costs
4 $954,700.00
5 $56,900.00
6 $175,200.00
7 $155,600.00
8 $133,600.00
9 $94,800.00
10 $80,600.00
11 $66,100.00
12 $75,200.00
13 $15,100.00
15 $21,415.00
20 $665,400.00
21 $35,800.00
22 $7,600.00
23 $15,100.00
24 $11,900.00
25 $47,500.00
26 $59,900.00
27 $59,900.00
28 $50,100.00
29 $40,200.00
30 $14,900.00
31 $11,800.00
32 $21,100.00
33 $11,900.00
34 $17,300.00
35 $14,950.00
37 $8,900.00
38 $5,900.00
39 $10,000.00

Total Cost of the Portfolio $2,939,365.00

116
Journal of Business and Behavioral Sciences

REFERENCES

Amirteimoori, A. and L. Khoshandam. (2011). “A Data Envelopment Analysis

Approach to Supply Chain Efficiency.” Advances in Decision Sciences,
Volume 2011.
Anderson, D., Sweeney, D., Williams, T., Camm, J., Cochran, J., Fry, M.,
Ohlmann, J. (2013). Quantitative Methods for Business, 12th ed.,
Cengage Learning.
Archer, N., & Ghasemzadeh, F. (2004). “Project portfolio selection and
management.” In P. W. G. Morris & J. K. Pinto (Eds.), The Wiley
Guide to Managing Projects. (pp. 237–255). Hoboken, NJ: John Wiley
& Sons.
Charnes, A., W. Cooper, and E. Rhodes (1978). “Measuring the Efficiency of
Decision Making Units.” European Journal of Operational Research, 2,
pp. 429-444.
Dia, M. (2009). “A Portfolio Selection Methodology Based on Data
Envelopment Analysis.” INFOR. Vol. 47, No. 1, February, pp. 71-79.
Ghasemzadeh, F., Archer, N., & Iyogun, P. (1999). “A Zero-one model for
project portfolio selection and scheduling,” The Journal of the
Operation Research Society, 50 (7), pp. 745–755.
Keren, B., Hadad, Y., and Z. Laslo (2014). “Combining AHP and DEA Methods
for Selecting a Project Manager.” Management: Journal of Sustainable
Business and Management Solutions in Emerging Economies,
Belgrade Vol. 19, Iss. 71, pp. 17-28.
Malhorta, R., Malhorta, D, and R. Russel (2010). “Using Data Envelopment
Analysis to Rate Bonds.” Journal of Business & Economic Studies,
Volume 16, No. 1, 60-78.
Meredith, J. and S. Mantel (2017). Project Management: A Managerial
Approach, 9th ed. John Wiley & Sons.
Pinto, J. (2010). Project Management: Achieving Competitive
Advantage. Prentice Hall.
Sanghera, P. (2019). PMP® in Depth: Project Management Professional
Certification Study Guide for the PMP® Exam, Paul Sanghera.
Sarkis, J., Presley, A., and D. Liles (1997). “The strategic evaluation of candidate
business process reengineering projects.” International Journal of
Production Economics, 50(2–3), pp. 261–274.
Soheilirad, S., Govindan, K., Mardani, A., Zavaskas, E., Nilashi. M., and N.
Zakuan (2018). “Application of data envelopment analysis models in
supply chain management: a systematic review and meta-analysis.”
Annals of Operations Research, 271.
Strang, K. D. (2011). “Portfolio selection methodology for a nuclear
project.” Project Management Journal, 42(2), pp. 81–93.

117
Copyright of Journal of Business & Behavioral Sciences is the property of American Society
of Business & Behavioral Sciences and its content may not be copied or emailed to multiple
sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.