Page 1 of 22

Advances in Social Sciences Research Journal – Vol.8, No.1

Publication Date: January 25, 2021

DOI:10.14738/assrj.81.9568.

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94-

115.

94

Comprehensive Project Risk Management Methodology

Luis F. Copertari

Computer Engineering Program

Autonomous University of Zacatecas (UAZ). Zacatecas, México.

ABSTRACT

The objective of this paper is to introduce and discuss the basics of a

methodology called the Probabilistic Critical Path Method (PCPM) for

managing the previously identified risks (uncertainty) of the three

project management dimensions: time, cost and return (performance).

An interactive Graphic User Interface (GUI) has been designed for

visualizing the tradeoffs among these three dimensions as well as their

uncertainties on a flat computer screen. The user can choose to visualize

the probability of failure (exceeding some user given due date, budget

or not exceeding a given Minimally Attractive Rate of Return – MARR) or

the probability of success (not exceeding the due date and the budget

and exceeding the MARR). PCPM allows for comprehensive project risk

management and it constitutes a new integrative project risk

management framework. This paper shows that it is possible to

integrate all three project management dimensions (time, cost and

return) and show their known risks as well as determining the optimal

cost and the associated time and return for such optimal cost. Finally, it

is possible to interactively show all this multidimensional information

on a flat computer screen.

Keywords Risk management; project management; decision support

systems.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or

not-for-profit sectors.

INTRODUCTION

One of the most important tasks in project management, particularly in technology and information- based organizations, is to successfully manage risk. Thus, project risk management becomes critical

for successful project management (Buganová & Šimíčková, 2019; Khameneh, Taheri, & Ershadi,

2016; Odimabo & Oduoza, 2018; Serpella, Ferrada, Howard, & Rubio, 2014; Szymański, 2017).

The methodology introduced in this paper, called the Probabilistic Critical Path Method (PCPM)

allows to manage the three project management dimensions: time, cost and performance (the latter

being measured by the Internal Rate of Return, or simply return, of the project) and meaningfully

link them to the identified risks of each dimension in the form of the probability of failure or,

alternatively, the probability of success. It integrates these dimensions by allowing to visually see

their tradeoffs as well as the probability of failure, that is, of exceeding the due date and the budget

and the probability of not exceeding the Minimally Attractive Rate of Return (MARR), or

Page 2 of 22

95

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

alternatively, to visualize the probability of succeeding by not exceeding the due date or the budget

and exceeding the MARR. A specially designed and interactive Graphic User Interface (GUI) for

successful project risk management decision making was created and even the optimal cost can be

calculated given some specific indirect project cost slope (m). This methodology allows for

comprehensive project risk management. This paper explains the basics of PCPM. Calculations have

been automated by creating a Decision Support System (DSS) called Schedule1, which can be

downloaded for free from www.copertari.net/schedule.

Managing Risk And Uncertainty

A great deal of project management involves good risk management. Risk management in a project

can be defined as the consequence of the existence of significative uncertainty concerning the level of

project achievement (Chapman & Ward, 1997). Tight time, cost or performance goals increase time,

cost and performance risks, respectively. A risky situation is often considered as the existence of

potentially high and unacceptable costs due to events considered more or less likely to happen. This

negative approach to risk leads to the idea that risk management essentially deals with reducing or

removing the possibility of under-achievement. Risk analysis is not a ‘throwing the dice’ situation,

but rather an area of study where a proactive, creative and intelligent prior planning approach must

be used, instead of entrenching in a defensive position (Adams, 2001; Dey, 2001; McManus, 2001;

Schimmoller, 2001; Walker, 2001).

Within this context, it is important to distinguish between risk and uncertainty. Risk is the

possibility or probability of failure, whereas uncertainty is the variability of the relevant outcomes

for a given risk or eventuality. Brealey and Myers (1991) define risk saying that more things can

happen (at present time) that will happen (in the future). Uncertainty, on the other hand, is the

degree in which an identified threat or risk (at present time, after previous analysis) will

(presumably, based on experience, historical data or assumptions) vary. Uncertainty is an identified

(and quantified) risk. Even then, the degree in which such identified risk will vary is unknown.

Uncertainty constitutes the ‘known’ unknowns because although a specific risk has been identified,

its exact impact is still unknown. Unidentified risks are ‘unknown’ unknowns because, generally

speaking, a risk is unquantified uncertainty about something not yet being considered possible as a

future eventuality. It will be assumed throughout this paper that risk identification has been

successfully and completely carried out and will focus attention on the risks due to the uncertainty

for the most relevant variables previously identified by decision makers. Risk sources are any

factors that may affect the project dimensions (time, cost and performance). Setting a tight time

target such as an optimistic due date increases time risk. Likewise, an irrationally low budget

increases cost risk and setting a minimum Internal Rate of Return (IRR) increases performance risk.

On the other hand, setting slack times, emergency budget allocations or a lower IRR reduces time,

cost and performance risks, respectively (Dawson, 1998; Farrell, 1996; Lefley, 1997; Tavares,

1998).

1 The heuristic specially designed for probabilistic crashing is proprietary software technology developed for Schedule.

Page 3 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

96

Systemic Relationships Between The Project Dimensions: Time, Cost And Return

(Performance)

Although the relations among the project management dimensions vary from time to time and from

project to project, a systemic approach can be used to elucidate the nature of the underlying

balances (Icmeli, 1996; Johnson & Schou, 1990; Sunde & Lichtenberg, 1995).

Figure 1a illustrates the systemic relationships between time and cost using influence diagrams. If

the project is delayed (it takes longer) will cost more, so that there is a positive correlation between

time and cost. But if in order to deliver the project on time, additional resources are used for critical

activities, maintaining resources to a minimum for non-critical activities (which is called crashing)

there is a negative correlation between cost and time (Winston, 1994). The existence of both a

positive and a negative correlation between time and cost implies the existence of a balance point

in which an optimal project completion can be achieved at a minimum cost. Figure 1b illustrates

how the time/cost balancing is additionally influenced by performance. Improving the quality of the

product requires investing more resources, which will increase cost and increase time if those

resources are limited. But if more resources are invested and it takes longer to complete the project,

it costs more, so that the Internal Rate of Return (IRR) of the project measuring its profitability is

reduced. As a consequence, there must also be an optimal balance between time/cost achieving an

optimal performance as measured according to the project’s IRR.

Figure 1. Balances among time, cost and performance.

Project Scope And Time, Cost And Performance Risks

The first and most important task in the management of a project in particular is determining its

scope, that is, exactly what is intended to achieve with the project and exactly in what measure or

proportion. From the scope determination, the different aspects of the project are planned. A

procedure called Work Breakdown Structure (WBS) is used in order to identify the tasks and sub- tasks (or simply activities, being the total task the project itself) as well as activity precedence, that

is, which activity should go before which other or if it has no precedence and it can begin at the start

of the project. Two very important dimensions of the project are time and cost, which are closely

linked by a procedure called crashing. An additional dimension is performance, although some

speak about quality instead of performance. From the point of view of this paper, quality will be the

degree of success in the performance of the project, and the performance will be measured using

the Internal Rate of Return (IRR) of the project, for which the target is set based on the Minimally

Attractive Rate of Return or MARR (Copertari, 2014).

a. Time-cost balance

Time Cost

+

Delay

-

Crashing

b. Time/cost-performance

balance

Time/Cost Performance

-

Profit

+

Quality

Page 4 of 22

97

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

The Probabilistic Critical Path Method (PCPM) considers the project and its dimensions from a

probabilistic point of view. Thus, if the project is intended to be finished by a certain due date, the

probability of success in time and the risk of not fulfilling the due date are illustrated in Figure 2a.

Also, if the goal is to finish the project within a certain budget, project success in the cost dimension

and cost risk are shown in Figure 2b. Finally, if the project goal is to reach a minimum MARR,

performance success as measured by the project’s IRR as well as the risk of not reaching the MARR

are pictured in Figure 2c.

Figure 2. Project success and risk based on the time dimension (set by the due date), cost (set by the

budget) and performance (set by the MARR and measuring performance according to the project’s

IRR).

The Targets Of Project Management

When we talk about project management, the most important aspects to consider in this area of

study come up: project scope, time, cost and quality. Time and cost are clearly two of the most

important dimensions of all projects and they are intimately connected. The quality of the project is

something more difficult to measure. Generally speaking, it could be said that quality is satisfying

the customer’s expectations. However, who is the customer of a project? The project team members

could be considered the customers, or the people receiving the project deliverables, or maybe the

decision makers wishing the project to be successful so that they can innovate in their products or

services or even the company stockholders wishing a successful project in order to receive a

satisfactory return over their investment. It is precisely the latter point of view of quality

measurement in a project, the one of the stockholders, being considered in order to define the way

in which the project performance is measured. In this case, it is the Internal Rate of Return (IRR), or

simply return, which is the rate of return over the investment the project yields, that is, that rate at

which the sum of all cash flows is equal to zero, being those either positive (the final income to

Due date Time

(days)

Time risk

Probability of

delivering on

time

a. Time dimension

Budget Cost

($)

Cost risk

Probability of

delivering on

budget

b. Cost dimension

MARR Performance

(IRR)

Performance

risk

Probability of

exceeding the

MARR

c. Performance dimension

Page 5 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

98

obtain in a project) or negative (the different costs associated to each activity, being the activity

critical or not and whether it is decided to crash it or not).

Figure 3. The dimensions of project management as an ideal balancing among the time, cost

and performance dimensions.

In this way, for the purposes of this paper, there are three dimensions of the project to be measured:

time, cost and performance. In PCPM time is measured probabilistically and the due date is used as

a basis to indicate the project’s time risk (see Figure 2a). The cost is also measured probabilistically

and the budget is used as an evaluation point in order to know the probability of exceeding the

budget, which is the cost risk. Finally, there is performance, measured as a dimension to exceed

(while time and cost are dimensions not to exceed), having the IRR as indicator of performance and

the MARR as the indicator of the threshold for success or failure. Figure 3 illustrates the three

targets of project management and the way in which they can be visualized in a probabilistic three- dimensional chart.

THE SHAPE PARAMETERS EQUATION FOR THE BETA DISTRIBUTION

In general, we have that the mean () and the variance (2) can be calculated in practice according

to equations (1) and (2), respectively, where the minimum, most like and maximum duration times

are given as a, m, and b, respectively (Meredith & Mantel, 1995).

Performance

(IRR)

Cost

(Money)

Time

(Duration)

Due date

Budget

MARR

Performance

risk

Cost risk

Time risk

Target

Page 6 of 22

99

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

μ =

a+4m+b

6

(1)

σ

2 = (

b−a

6

)

2

(2)

Also, the general formula for the mean and the variance of a beta probability density function with

range (a and b) and shape ( and ) parameters (Hastings & Peacock, 1975) are given in equations

(3) and (4), respectively.

μ =

aβ+bα

α+β

(3)

σ

2 = (b − a)

2 αβ

(α+β)

2(α+β+1)

(4)

Since the values for  and 2 are known from equations (1) and (2) based on PERT’s three point

estimates (a, m and b), equations (3) and (4) constitute a non-linear system of two equations with

two unknowns ( and ), which can be solved. Equations (5) and (6) provide the solution to this

system of non-linear equations, which yield the values for  and , respectively.

α = (μ − a) (

(μ−a)(b−μ)−σ

2

(b−a)σ2

) , b > a (5)

β = (b − μ) (

(μ−a)(b−μ)−σ

2

(b−a)σ2

) , b > a (6)

Equations (5) and (6) are novel and specific findings of the Probabilistic Critical Path Method

(PCPM). What is the practical relevance of equations (5) and (6)? Now, it is possible to correctly

calculate the shape parameters ( and ) of any activity or any variable described using a beta

probability density function based on the minimum (a), maximum (b), mean () and variance (2)

values. In turn, the mean () and the variance (2) values can be calculated using equation (1) and

(2), respectively, which are based on three-point estimates for minimum (a), most likely (m) and

maximum (b) values related to the variable being modelled using the beta probability density

function.

THE STARTING OF ACTIVITIES IN A PROJECT MANAGEMENT NETWORK

Let k be any activity and N be the total number of activities in a project management network. In

order to calculate the projects’ IRR (or simply return) it is necessary to know when to apply activity

costs. The problem with CPM is that there is no specific starting or finishing activity completion

times if they have a slack greater than zero. Let dk, ESk, EFk, LSk, LFk and Sk be a specific duration,

early start, early finish, late start, late finish and slack for any activity k, where k = 1, 2., ..., N. Also,

let Ck = {1,0} indicate with a 1 if the activity is critical and with a 0 if it is not. A critical activity delays

the entire project, so they have a slack of zero. However, for non-critical activities, the slack is

greater than zero. The problem is how to specify when non-critical activities start. Let Startk (or

simply Stk) be the starting of activity k, where k = 1, 2, ..., N. The user decides how optimistic (0 ≤ O

≤ 1) or pessimistic (0 ≤ P ≤ 1) wants to be. A completely optimistic user (O = 1) would start the

activity at its early starting time (ESk), whereas a completely pessimistic user (P = 1) would start

the activity at its latest starting time (LSk). Clearly, O+P = 1 and combinations of O and P different

than one and zero are possible. Equation (7) indicate how to calculate the starting time (Startk or

simply Stk). Notice that equation (8) is also satisfied.

Startk = O × ESk + P × LSk (7)

Page 8 of 22

101

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

Figure 5. Minimum activity completion time calculations for the testing example.

Figure 6. Maximum activity completion time calculations for the testing example.

4 0 4.5

9 | I 1

4 0.5 4.5

5 | E 0

1 0.5 1.5

2 2 2.5

0.5 0.5 1.0

0 0.5 0.5

1 | A 0

0 0 1

2 | B 1

0 1 1

1 1 4

6 | F 1

1 3 4

3.5 1 4.5

8 | H 0

2.5 1 3.5

2.5 1 4.5

7 | G 0

2 3.5 s

1.5

Finish 10

0 4.5

Start 0

1.5 1 3.5

3 | C

0.5 2 2.5

0

2 1.5 2.5

4 | D

0.5 0.5 1

0

11 19

9 8 17

7 | G 0

2

4 5 9

5 | E 0

6 2 11

1 1 4

0 3 3

1 | A 0

0 4 4

2 | B 1

0 0 4

4 9 13

6 | F 1

4 0p

13

1 19 s

12

11 7 18

8 | H 0

13 6 19

9 | I 1

13 0 19

Finish 10

19

Start 0

0

3 8

4 1 12

11

3 | C 0

7 4

3 4 7

4 | D 0

11

Page 9 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

102

Figure 7. Mean activity completion time calculations for the testing example.

In Figures 5, 6 and 7, the critical path resulting in each case is highlighted using bold lines. The only

thing missing in order to be able to calculate beta distributed completion times is the variance (2)

in order to be able to successfully apply equations (5) and (6). However, we are not going to use the

PERT approach based on assuming the stochastic completion times of the activities to be

represented using the mean activity duration time () and CPM calculations. This is because we

want more than just a probabilistic project completion PERT time, which assumes a normally

distributed completion time that tends to underestimate the actual completion time, because we

also want the probabilistic beta distributed completion times of all the activities.

Criticality and Project Completion Variance

Now we need a new concept, called criticality. Criticality in a PCPM network is the probability for

each activity to be critical. Thus, Ck is no longer either zero or one, but a number between zero and

one, since activities now have a given probability of being critical. For that, we are going to use

simulation. Let s be the total number of simulation runs, Ck,i be the criticality obtained for activity k

(k = 1, 2, ..., N) in iteration i (i = 1, 2, ..., s), 2 the variance in the completion time for the whole

project, k

2 the variance of the completion time for activity k, path p be the set of all the activities

(non-repetitive) of all the alternative project network paths in which activity k exists. Then,

equations (9), (10) and (11) apply.

Ck =

1

s

∑ Ck,i

s

i=1

, k = 1, 2, ... ,N, where Ck,i = {

1, if the activity is critical

0, if the activity is not critical (9)

σ

2 = ∑ Ckσk

N 2

k=1

, 0 ≤ Ck ≤ 1 (10)

σk

2 = ∑ Chσh

2

h⊂p , 0 ≤ Ck ≤ 1 (11)

For the testing example shown in Table 1, a total of 1,000 runs and 10 re-runs where carried out.

Table 2 shows the results. Notice that a slack is zero if it is between a zero value of ±0.0001.

9 16

7 7 14

7 | G 0

2

3 4 7

5 | E 0

5 2 9

1 1 3

0 2 2

1 | A 0

0 3 3

2 | B 1

0 0 3

3 8 11

6 | F 1

3 0p

11

1 16 s

10

9 6 15

8 | H 0

11 5 16

9 | I 1

11 0 16

Finish 10

16

Start 0

0

2 7

3 1 10

9

3 | C 0

6 4

2 3 5

4 | D 0

9

Page 10 of 22

103

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

Table 2. Criticality example using simulation (s = 10,000, r = 10) where the activity is critical if

the slack is approximately zero.

k = 1 2 Block 1 3 4 5 6 Block 2 7 8 9 Block 3

r C1 C2 C1+C2 C3 C4 C5 C6

C3+C4+

C5+C6 C7 C8 C9

C7+C8+

C9

1 32.5% 67.5% 100.0% 29.1% 0.0% 6.5% 64.5% 100.0% 6.5% 29.1% 64.5% 100.0%

2 31.9% 68.1% 100.0% 28.6% 0.1% 6.5% 64.9% 100.0% 6.5% 28.6% 64.9% 100.0%

3 32.7% 67.3% 100.0% 28.7% 0.1% 5.8% 65.4% 100.0% 5.9% 28.7% 65.4% 100.0%

4 32.4% 67.6% 100.0% 28.9% 0.0% 6.0% 65.1% 100.0% 6.0% 28.9% 65.1% 100.0%

5 32.9% 67.1% 100.0% 29.4% 0.0% 6.3% 64.3% 100.0% 6.3% 29.4% 64.3% 100.0%

6 32.3% 67.7% 100.0% 28.7% 0.0% 5.8% 65.4% 100.0% 5.9% 28.7% 65.4% 100.0%

7 32.4% 67.6% 100.0% 28.8% 0.0% 6.1% 65.1% 100.0% 6.1% 28.8% 65.1% 100.0%

8 32.4% 67.6% 100.0% 28.9% 0.0% 6.2% 64.9% 100.0% 6.3% 28.9% 64.9% 100.0%

9 32.8% 67.2% 100.0% 29.4% 0.0% 6.9% 63.6% 100.0% 7.0% 29.4% 63.6% 100.0%

10 31.7% 68.3% 100.0% 28.2% 0.0% 6.5% 65.3% 100.0% 6.5% 28.2% 65.3% 100.0%

Ck 32.4% 67.6% 100.0% 28.9% 0.0% 6.3% 64.9% 100.0% 6.3% 28.9% 64.9% 100.0%

The advantage of using criticality is that now we have beta distributed probabilistic completion

times for selected activities marked as milestones in the project and not only a normally distributed

probabilistic completion time.

THE OPTIMAL COST WITH UNLIMITED RESOURCES

Let the project completion time be denoted as T. The result of applying the probabilistic crashing

heuristic is a series of points beginning with the maximum project completion (T = b) having the

minimum overall cost (CA or U) and proceeding in a series of iterations until the minimum project

completion (a) is reached having a maximum project cost (CB or V). For simplicity, crashing can be

assumed to be an inversely proportional relationship between time and cost (Babu & Surech, 1996;

Foldes & Soumis, 1993). Notice that crashing refers to a CPM technique, whereas probabilistic

crashing refers to a PCPM heuristic technique specially designed for the PCPM methodology.

Generally speaking, however, it can be said that there are two kinds of costs: direct and indirect

costs. Direct costs, which are inversely proportional to the project completion time (T) include the

cost of material, equipment, and direct labor required to perform an activity, although in this case

we will consider the entire project. Indirect costs include in addition to supervision and other

overhead costs, interest charges on the cumulative project investment and overdue penalty costs

(Auguston, 1993; Clark, 1992; Meredith & Mantel, 1995; Rivenbank, 2000; Tattersall, 1990).

Let D denote the direct costs of the whole project. Then, direct cost is inversely proportional to time

(T), as indicated in equation (12).

D ∝

1

T

(12)

In order to transform such proportionality relationship into an equality, proportionality constants

are added, as indicated in equation (13).

Page 11 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

104

D =

χ

T

+ φ (13)

In this case, there are for practical purposes, two direct costs: V, which is associated to time T = a,

and U, which is associated with time T = b. Thus, equations (14) and (15) are created.

V =

χ

a

+ φ (14)

U =

χ

b

+ φ (15)

These system of two equations have four known values (after probabilistic crashing and time

calculations): V, U, a and b and two unknowns:  and . Thus, it can easily be solved. The solution

for the unknown variables ( and ) are indicated in equations (16) and (17), respectively.

χ =

ab(V−U)

b−a

, b > a (16)

φ =

Ub−Va

b−a

, b > a (17)

Thus, substituting from equations (16) and (17) into equation (13) results in the direct cost

expression, shown in equation (18).

D =

ab

T

(V−U)

(b−a)

+

Ub−Va

b−a

(18)

The project indirect cost, I, is directly proportional to the project completion time (T), as indicated

in equation (19).

I ∝ T (19)

The proportionality expression in equation (19) can be transformed into an equality adding

proportionality constants. Let O denote the minimum overhead (indirect) cost at time T = 0, and m

the increments (slope) of the indirect cost as a function of time. Then, equation (20) indicates the

relationship between time and indirect cost.

I = O + m T (20)

The total cost (C) is the result of adding the indirect (I) and the direct (D) costs, as indicated in

equation (21).

C = I + D (21)

Substituting I from equation (20) and D from equation (18) into equation (21) yields equation (22),

which is the sum of the indirect and direct costs, or total cost.

C = (O + m T) + (

ab

T

(V−U)

(b−a)

+

Ub−Va

b−a

) (22)

Taking the derivative of the total cost (C) with respect to time (T) from equation (22) and equating

that to zero yields the optimal project completion time (T*). The derivative is shown in equation

(23).

δC

δT

= m −

ab

T2

(V−U)

(b−a)

= 0 (23)

Page 12 of 22

105

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

Solving for T from equation (23) yields the optimal completion time (T*) as indicated in equation

(24).

T

∗ = √

ab

m

(V−U)

(b−a)

(24)

Figure 8a illustrates the direct (D) and indirect (I) costs, whereas Figure 8b shows the total cost and

it highlights the fact that the derivative, when equating to zero, would result in the optimal project

completion time (T*) for the optimal minimal total cost (C*).

a. Direct and indirect project costs. b. Optimal project cost.

Figure 8. Time-cost tradeoffs.

TIME-COST TRADEOFF

When considering the total cost of the project, a concept that is central not only to CPM but also to

PCPM is the time-cost tradeoff. When considering there are unlimited resources, it is possible in

CPM as well as in PCPM to calculate a tradeoff between time and cost.

Generally speaking, the longer it takes to complete the project because fewer resources are used,

the lower the cost of such doing is. If we use additional resources, it should be possible to reduce

the time it takes to complete the project at the expense of incurring in additional costs. By following

an optimal algorithm, a curve similar to the one corresponding to the direct costs in Figure 8a is

generated. However, because there are too many activities in realistic projects, it is not possible to

consider all alternatives of activity crashing. A project with 10 activities would have 210 = 1,024

crashing combinations to consider. A project with 20 activities would have 220 = 1’048,576 crashing

combinations. Schedule allows to have a maximum of 254 activities. Considering 2254 crashing

combinations is out of any practical algorithm that could be devised. Even a relatively realistic

Time

(T)

Cost

(C)

a b

U

O

V

O+mb

m

D

I

Time

(T)

Cost

(C)

a b

C*

O+ma+V

T*

O+mb+U

Page 13 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

106

project with 50 activities (250 combinations) would mean disaster for a combinatorial algorithm to

find the optical time-cost tradeoff.

In any case, the theory holds that if such optimal time-cost crashing calculations could be performed,

the result should look like the direct cost curve in Figure 8a. Let iteration i = 0 correspond to the

case where there is no activity crashing at all having the maximum project completion time (b), and

the maximum number of iterations (I) be the case where all activities are crashed. Clearly, iteration

i = I is reached only after the minimum project completion time (a) has been obtained in the

previous iteration (that is, iteration i = I-1).

Let CAk be the normal activity duration cost (corresponding to the maximum activity duration time

bk) and CBk be the crashed activity duration cost (corresponding to the minimum activity duration

time ak) for activities k = 1, 2, ..., N. Also, let time ti be the completion time of the project for iteration

i and cost ci be the corresponding cost to such project completion time.

Figure 9 shows the ideal (optimal) time-cost tradeoff sequence of I iterations. For this case, the

minimum project cost (CA) is calculated according to equation (25) and the maximum project cost

(CB) is given according to equation (26).

Figure 9. Ideal (optimal) time-cost tradeoff crashing sequence of I iterations.

CA = ∑ CAk

N

k=1

, k = 1, 2, . . . ,N (25)

CB = ∑ CBk

N

k=1

, k = 1, 2, . . . ,N (26)

The variance in the cost of activity k (Ck

2) is given according to equation (27).

0

1

2

3

i-1

i

i+1

I

I-2

I-1

Project time

Project cost

a b

CA

CB

(t0,c0)

(t1

,c1

)

(t2

,c2

)

(t3

,c3

)

(ti-1

,ci-1

)

(ti

,ci

)

(ti+1,ci+1)

(tI-2

,cI-2

)

(tI-1

,cI-1

)

(tI

,cI

)

Page 14 of 22

107

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

Cσk

2 = (

CBk−CAK

6

)

2

, k = 1, 2, ... ,N (27)

The only information we need to be able to apply the beta distribution shape parameters equations

is the mean cost. For that, it is necessary to interpolate the mean cost (C) corresponding to the

mean project completion time ().

Now, we need to consider the case where there is a time t corresponding to a cost c between

iterations i and i-1. Equation (28) calculate the interpolated cost c corresponding to an interpolation

time t, whereas equation (29) yield the interpolated time t corresponding to an interpolation cost c.

Figure 10 illustrates the situation.

Figure 10. Ideal (optimal) time-cost tradeoff crashing sequence of I iterations for a given

interpolation time and cost (t,c).

c = ci −

(ti−1−t)(ci−ci−1

)

ti−ti−1

, ci ≥ c > ci−1 and ti ≤ t < ti−1 (28)

t = ti +

(ci−c)(ti−1−ti

)

ci−ci−1

,ti ≤ t < ti−1 and ci ≥ c > ci−1 (29)

Thus, having a mean interpolation time , it is possible to calculate the interpolated cost C

according to equation (30).

Cμ = ci −

(ti−1−μ)(ci−ci−1

)

ti−ti−1

, ci ≥ Cμ > ci−1 and ti ≤ μ < ti−1 (30)

Project cost

i

1 0

2

3

i-1

(t,c)

i+1

I-2

I-1

I

a b Project time

CA

CB

t

c

(ti,ci)

(ti-1

,ci-1

)

Page 15 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

108

Now, having the minimum project cost (CA), maximum project cost (CB), the mean project cost (C)

and the variance project cost (C2), we can calculate the shape parameters (C and C) for the

corresponding beta distributed cost probability density function as given by equations (31) and

(32).

Cα = (Cμ − CA) (

(Cμ−CA)(CA−Cμ)−Cσ

2

(CB−CA)Cσ2

) , CB > CA (31)

Cβ = (CB − Cμ) (

(Cμ−CA)(CB−Cμ)−Cσ

2

(CB−CA)Cσ2

) , CB > CA (32)

THE RETURN (PERFORMANCE) DIMENSION

The Net Present Value (NPV) and the Internal Rate of Return (IRR) or simply return

Let T be the project duration for certain iteration i (i = 0, 1, 2, ..., I) and the Startk (Stk as a shortcut)

the beginnings of each activity for k = 1, 2, ..., N. It is assumed a certain cost k (ck) occurs at the

beginning of each activity (at time Stk). Also, suppose there is an income, which is assumed to be

collected at the end of the project (at time T for each time-cost tradeoff iteration). The costs of each

activity are the ones collected for such iteration. Finally, such income (Income) is fixed and should

in principle (but not necessarily) be greater than the maximum project cost (CB). It is initially

assumed that the Internal Rate of Return (IRR) is, as a consequence, positive. In order to simplify

calculations, let X denote the value for 1+IRR, as indicated in equation (33).

X = 1 + IRR (33)

Then, the IRR is the rate at which the Net Present Value (NPV) equals zero. The equation that

explains the NPV calculation when it is zero is shown in equation (34).

NPV = 0 = Income − ∑ ckX

N T−Stk

k=1

(34)

Equation (34) obtains the value of X when the NPV = 0. From that X, based on equation (33), we can

calculate the value of the IRR. It is assumed there is one IRR for the project, regardless of the

duration units (days, weeks, months or years). This is because calculating the annualized rate of

return brings additional complications. Thus, we have a function of X, f(X), indicated in equation

(35).

f(X) = Income − ∑ ckX

N T−Stk

k=1

(35)

We need to find the value (and there should be only one because there is one sign change only in

the cash flows) for which f(X) = 0. For that, we rely on the Newton-Raphson method of assured

convergence (Burden & Faires, 1985). Beginning with any given value for X (such as 1.5, for

example), we proceed from such initial value (X0 = 1.5 for j = 0) and continue to the next iteration.

The process for each new iteration j is indicated in equation (36).

Xj+1 = Xj −

f(Xj)

f

′(Xj)

(36)

Page 16 of 22

109

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

We can see that equation (36) requires us to calculate the derivative of the function f(Xj), that is

f’(Xj). That can easily be done by derivation of equation (35), since X is the unknown variable, which

is shown in equation (37).

f

(X) = − ∑ ck

(T − Stk

)X

N T−Stk−1

k=1

(37)

The rate of return obtained (IRR) does not necessarily increases as the time-cost tradeoff iterations

proceed, because the result obtained according to Schedule’s heuristic is not necessarily convex.

Thus, we take the minimum such rate of return and call it RA as well as the maximum rate of return

and call it RB. Let RR be the rate of return corresponding to some given time completion value t.

Thus, we can relate the rate of return RR to a given time t and calculate a given time t for a given

rate of return RR, as indicated in equations (38) and (39), where a is the minimum (all crashed)

project completion time and b the maximum (normal) project completion time.

RR = RA +

(t−a)(RB−RA)

b−a

, b > a (38)

t = a +

(RR−RA)(b−a)

RB−RA

, RB > RA (39)

The mean rate of return (R) is assumed to be the rate of return at which the mean project

completion time () occurs. Thus, we have equation (40).

Rμ = RA +

(μ−a)(RB−RA)

b−a

, b > a (40)

The return variance (R2) is calculated according to the original PERT equation, which is shown in

equation (41).

2 = (

RB−RA

6

)

2

(41)

Since we have the minimum (RA), maximum (RB), mean (R) and variance (R2) of the IRR or

simply return, we can calculate the beta shape parameters (R and R) according to equations (42)

and (43).

Rα = (Rμ − RA) (

(Rμ−RA)(RA−Rμ)−Rσ

2

(RB−RA)Rσ2

) , RB > RA (42)

Rβ = (RB − Rμ) (

(Rμ−RA)(RB−Rμ)−Rσ

2

(RB−RA)Rσ2

) , RB > RA (43)

Testing The Theory With The Example From Table 1.

Table 3 shows the time (ti), cost (ci) and return (ri) for each iteration (where I = 15+1) calculated

for the example from Table 1 as given by Schedule.

Page 17 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

110

Table 3. Time, cost and return calculations from Schedule for the realistic enough example shown in

Table 1 where Income = $2,500.

I Time (ti) Cost (ci) Return (ri)

0 19 $ 1,245.0000 5.36516914%

1 18 $ 1,253.3333 5.52892210%

2 17 $ 1,278.3333 5.64622149%

3 15 $ 1,348.3333 6.03133936%

4 13 $ 1,418.3333 6.59978418%

5 12 $ 1,458.6364 6.99933285%

6 11 $ 1,498.9394 7.67486409%

7 10 $ 1,692.8788 6.50907407%

8 9 $ 1,732.8788 6.18065098%

9 8 $ 1,839.5455 5.23776127%

10 7.5 $ 1,892.8788 4.91881442%

11 7 $ 1,926.2121 4.87586633%

12 6.5 $ 1,959.5455 4.81307255%

13 6 $ 2,029.7835 4.47847154%

14 4.5 $ 2,260.9524 3.03750096%

15 4.5 $ 2,430.0000 0.94333776%

Figure 11 shows the time-cost tradeoffs and Figure 12 shows the return for the time/cost-return

calculations from Table 3.

Notice in Figure 11 that the time-cost tradeoff obtained according to Schedule’s proprietary

probabilistic crashing heuristic is not a convex function as shown in Figure 8a, Figure 9 and Figure

10. However, as the completion time decreases, the cost always increases.

Also, notice that in Figure 12 the returns obtained are not always increasing as the completion time

increases. This is because of the complexity of the functions in equations (35) and (37).

Page 18 of 22

111

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

Figure 11. Time-cost tradeoff for the realistic enough example from Table 1.

Figure 12. Time-return calculations for the realistic enough example from Table 1.

THE INTERACTIVE GRAPHIC USER INTERFACE (GUI) FOR SCHEDULE

As we have previously seen, it is possible to calculate beta distributed time, cost and return

probability density functions, which give us the “risk” of failure (and alternatively success) of each

dimension. Furthermore, these dimensions are related through the time-cost tradeoff curve. Figure

13 shows a scheme of this interactive user interface.

$1200.0

$1300.0

$1400.0

$1500.0

$1600.0

$1700.0

$1800.0

$1900.0

$2000.0

$2100.0

$2200.0

$2300.0

$2400.0

$2500.0

4.5

5

5.5

6

6.5

7

7.5

8

8.5

9

9.5

10

10.5

11

11.5

12

12.5

13

13.5

14

14.5

15

15.5

16

16.5

17

17.5

18

18.5

19

Total cost

Time

1 0

2

3

4

5

6

7

8

9

10 11 12

13

14

15

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

4.5

5

5.5

6

6.5

7

7.5

8

8.5

9

9.5

10

10.5

11

11.5

12

12.5

13

13.5

14

14.5

15

15.5

16

16.5

17

17.5

18

18.5

19

Rate of return (IRR)

Project completion (Time)

0

2 1

3

4

5

6

7

8

9

11 10

12

13

14

15

Page 19 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

112

Figure 13. Interactive GUI for Schedule.

These dimensions are related according to equations (28), (29) and (38). If we change the due date

(t), it immediately changes the budget (c) and the return (RR) and if we change the budget (c) it

changes the due date (t) and the return (RR). To avoid a cyclical user interaction, we can change the

MARR at will and see the effect on the bubble.

DISCUSSION AND CONCLUSION

CPM presents a considerable constraint: activity durations must be deterministic. Besides, there is

the inherent problem with the CPM that for the non-critical activities, they can begin at any time

between the Early Start (ES) of the activity and such ES plus the slack (S). Also, activities can finish

at any time between the Early Finish (EF) of the activity and such EF plus the slack (S). In all these

cases, when do the actual beginning and end of the non-critical activities actually occurs?

We require, then, a modification to CPM. Using the optimistic percentage (O) and its corresponding

pessimistic one (P = 1- O), it is possible to specify values given the starting and finishing times of

the activities to know when it is they occur in each iteration of the simulation required to calculate

the criticality of each activity.

c

t

RR = MARR

RB

Return

a b

CA

CB

RA

0% 100%

100% 100%

Cost

Time

Page 20 of 22

113

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

The CPM has, nevertheless, something very important, which is the crashing method, which can be

extended to PCPM (calling it probabilistic crashing using the proprietary heuristic for the

computations) to obtain a time and cost relationship and also a time-cost and return relationship,

which allows to link all variables among them, even probabilistically, so that the risk element for

each dimension is added.

PERT, on the other hand, offers the idea of minimum, most likely and maximum times for the

duration of each activity, which is used in PCPM to calculate the mean and the variance of each

activity according to the PERT textbook formula. However, PERT tends to lead to optimistic

estimates of the duration of the activities by assuming a normal distribution (and not a beta

distribution, as with PCPM) representing the project duration. In principle, this is theoretically

wrong, because there is an absolute minimum project duration (called a in PCPM) and an absolute

maximum project duration (called b in PCPM). The normal distribution (unlike the beta

distribution) does not have minimum nor maximum duration, but rather a duration time that goes

from - to +, despite the fact that the probabilities of having these extreme durations are

infinitesimal.

There is, nonetheless, an additional limitation to PERT: what is the project duration variance when

there is more than one critical path? We could assume that in this case the sum of the variances of

the activities in each and every critical path are used and choose whichever is the highest of all

possible critical paths. However, PERT does not provide such consideration for assuming it unlikely

to happen, which constitutes a serious limitation of PERT. On the other hand, PCPM considers the

variance of the project based on the criticalities and it does not assume a normal distribution for

such duration, but rather a beta distribution, so that it does not lead to optimistic (nor pessimistic)

assumptions concerning the project duration. Instead, it leads to a precise probabilistic assessment.

Furthermore, it allows to calculate the beta distributed completion times of all the activities

considered to be milestones.

Notice that in this case the return calculation (IRR) was consistent with the theory, that is, as time

increases, the related cost is reduced and the return increases. It is possible that the return

calculations are not always following this precise trend. It is possible to have fluctuations such that

with time, return not always increases. In any case, the lowest project duration (a) is associated

with the minimum return (RA) and the largest project duration (b) is associated with the highest

return obtained (RB). However, the actual results can be consulted to find out the correct return for

any given time and cost, even for the optimal time and the optimal cost.

Finally, PCPM is the only method leading to a project duration, cost and return described by a beta

distribution, so that it is possible to estimate the time, cost and return risks for given due dates,

budgets or MARR, respectively, and consequently, define probabilistic tradeoffs among these

dimensions. In conclusion, PCPM not only emphasizes the risk element to the project, but it also

offers a comprehensive approach to Project Management (Project Risk Management to be more

specific) by considering the three dimensions and their inherent tradeoffs: time, cost and return

(the latter one measuring project performance).

Page 21 of 22

Advances in Social Sciences Research Journal (ASSRJ) Vol.8, Issue 1, January-2021

114

References

Adams, C. (2001). Risk aversion. Project Finance, 2-5.

Auguston, K. A. (1993). Activity-based costing adds up to big benefits. Modern Materials Handling, 48(7), 57-59.

Babu, A. J. G. & Surech, N. (1996). Project management with time, cost, and quality considerations. European Journal of

Operational Research, 88(2), 320-327.

Brealey, R. A. & Myers, S. C. (1991). Principles of Corporate Finance. New York: McGraw Hill.

Buganová, K. & Šimíčková, J. (2019). Risk management in traditional and agile project management. Transportation

Research Procedia, 40, 986-993.

Burden, R. L. & Faires, J. D. (1985). Análisis Numérico (Numerical Analysis). Belmont, California: Grupo Editorial

Iberoamérica.

Chapman, C., & Ward, S. (1997). Project Risk Management: Processes, Techniques and Insights. John Wiley & Sons.

Clark, R. H. (1992). Benefit/Cost Analysis for Trenchless Alternatives. Transactions of the American Association of Cost

Engineers, 2, Q.1.1-Q.1.8.

Copertari, L. (2014). Evaluación Financiera de Proyectos (Financial Project Evaluation). Zapopan, Jalisco, México:

Editorial Umbral.

Dawson, R. J. (1998). Practical proposals for managing uncertainty and risk in project planning. International Journal

of Project Management, 16(5), 299-310.

Dey, P. K. (2001). Project time risk analysis through simulation. Cost Engineering, 43(7), 24-32.

Farrell, L. M. (1996). The impact of illiquidity and uncertainty of the multiperiod project investment decision process.

Project Management Journal, 27(3), 35-45.

Foldes, S. & Soumis, F. (1993). PERT and crashing revisited: mathematical generalizations. European Journal of

Operational Research, 64(2), 286-294.

Hastings, N. A. J. & Peacock, J. B. (1975). Statistical Distributions. New York: John Wiley & Sons.

Icmeli, O. (1996). The resource constrained time/cost tradeoff project scheduling problem with discounted cash flows.

Journal of Operations Management, 14(3), 255-275.

Johnson, G. A., & Schou, C. D. (1990). Expediting projects in PERT with stochastic time estimates. Project Management

Journal, 21(2), 29-34.

Khameneh, A., Taheri, A., & Ershadi, M. (2016). Offering a framework for evaluating the performance of project risk

management system. Procedia – Social and Behavioral Sciences, 226, 82-90.

Lefley, F. (1997). Approaches to risk and uncertainty in the appraisal of new technology capital projects. International

Journal of Production Economics, 53(1), 21-43.

McManus, J. (2001). Risk in software projects. Management Services, 45(10), 6-10.

Meredith, J. R., & Mantel, Jr., S. J. (1995). Project Management: A Managerial Approach. New York: John Wiley & Sons.

Odimabo, O. & Oduoza, C. (2018). Guidelines to Aid Project Managers in Conceptualising and Implementing Risk

Management in Building Projects. Procedia Manufacturing, 17, 515-522.

Rivenbank, W. C. (2000). Benchmarking and cost accounting: the North Carolina approach. Journal of Public Budgeting,

Accounting and Financial Management, 12(1), 125-147.

Schimmoller, B. K. (2001). The shifting sands of risk management. Power Engineering, 105(9), 3.

Serpella, A. F., Ferrada, X., Howard, R., & Rubio, L. (2014). Risk management in construction projects: a knowledge- based approach. Procedia – Social and Behavioral Sciences, 119, 653-662.

Sunde, L., & Lichtenberg, S. (1995). Net-present value cost/time tradeoff. International Journal of Project Management,

13(1), 45-49.

Page 22 of 22

115

Copertari, L. F. (2021). Comprehensive Project Risk Management Methodology. Advances in Social Sciences Research Journal, 8(1) 94- 115.

URL: http://dx.doi.org/10.14738/assrj.81.9568

Szymański, P. (2017). Risk management in construction projects. Procedia Engineering, 208, 174-182.

Tattersall, J. (1990). Understanding overheads – their management and control. Management Decision, 28(4), 4-8.

Tavares, L. V. (1998). On the optimal management of project risk. European Journal of Operational Research, 107(2),

451-469.

Walker, L. E. (2001). Advantages of risk-based project management. Occupational Health & Safety, 70(9), 161-164.

Winston, W. L. (1994). Operations Research: Applications and Algorithms. Belmont, California: Duxbury Press.