Automation in Construction 59 (2015) 81–95

Contents lists available at ScienceDirect

Automation in Construction

journal homepage: www.elsevier.com/locate/autcon

Optimization of overlapping activities in the design phase of

construction projects
Reza Dehghan ⁎, Kamran Hazini, Janaka Ruwanpura
Center for Project Management Excellence, Schulich School of Engineering, University of Calgary, Calgary T2N 1N4, Canada

a r t i c l e i n f o a b s t r a c t

Article history: A well-known practice to accelerate construction projects is to overlap the design phase activities. For a typical
Received 22 November 2014 construction project, a number of overlapping strategies exist during the design phase which all can result in
Received in revised form 13 August 2015 timesaving. However, the cost of these strategies varies significantly depending on the total rework and complex-
Accepted 14 August 2015
ity they generate. A favorable overlapping strategy is one that generates the required timesaving at the minimum
Available online 5 September 2015
cost. To find such a strategy, the question “Which activities have to be overlapped and to what extent to reduce
Keywords:
the project duration at the minimum cost?” should be answered. This research aimed at answering the question
Overlapping through generating an overlapping optimization algorithm. The algorithm works based on the principles of
Fast-tracking genetic algorithms (GAs). The algorithm explained in the paper is unique compared to previous algorithms
Optimization algorithms and frameworks available in the literature, as it can optimize multi-path networks and can handle all types of
Time–cost trade-off activity dependencies (i.e. finish-to-start, start-to-start, and finish-to-finish). It also takes both critical and
Rework non-critical activities into account and follows the critical path if the critical path changes or new critical paths
Construction management emerge. A computer tool was also developed to run, examine and validate the overlapping optimization algorithm.
This paper introduces the algorithm and the computer tool in detail and explains the results of their validation
through optimizing a real-world project schedule.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction The remainder of this paper is structured as follows: A detailed liter-

An effective technique for earlier completion of projects is to overlap
the project activities that normally would be performed in sequence.
This process can result in rework and increased risk, which raise costs.
In the design phase of a typical project, a huge number of overlapping
strategies exist, which all can result in the same timesaving. However,
the cost of these strategies varies significantly depending on the total
amount of rework and complexity they generate. The best strategy is
one generating the required timesaving at the minimum cost. To find
this strategy, the question “which activities have to be overlapped and
to what extent to reduce the project duration at the minimum cost?”
should be answered. However, since the number of possible overlap-
ping arrangements can be extremely high, manual methods are not
helpful. In addition, available planning and scheduling tools (e.g. Prima-
vera, MS Project) lack the capability to evaluate the cost of overlapping
and identify the most favorable overlapping strategy. Therefore, devel-
oping a decision support tool capable of finding optimized overlapping
strategies is a step toward a new generation of planning and scheduling
tools. This paper introduces an overlapping optimization algorithm and
its associated computer tool, which fulfill the above need.
⁎ Corresponding author at: Center for Project Management Excellence, Department of
Civil Engineering, Schulich School of Engineering, University of Calgary, Calgary, T2N
1N4, Canada.
ature review is provided to position this research in the literature. Then
the overlapping mechanism is briefly reviewed and the overlapping
time–cost trade-off function central to solve the overlapping optimization
problem is introduced. Next, the overlapping algorithm and its
computerization—the major contribution of this paper—are explained in
detail, to allow the work to be reproduced by other researchers. Finally,
the results of experiments with the computer tool and the validation
process are described.
2. Background
The available literature about overlapping can be categorized into two
main areas: product development and project execution. Since the
manufacturing industry began to utilize concurrent engineering long
ago, the research for overlapping in product development is older and
more extensive than that for project execution, particularly construction
projects [1]. Several researchers such as Krishnan [2], Loch and Terwiesch
[3], Nicoletti and Nicolo [4], Prasad [5] and Terwiesch et al. [6] have inves-
tigated the inherent nature of activity overlapping in product develop-
ment. Roemer et al. [7,8] have tried to determine optimal overlapping
policies. Some researchers such as Eppinger [9], Pena-Mora and Li [10],
Bogus et al. [11–13] and Blacud et al. [14] have used the models and
frameworks developed in product development research studies to de-
velop similar models and frameworks for the construction industry.

http://dx.doi.org/10.1016/j.autcon.2015.08.004
0926-5805/© 2015 Elsevier B.V. All rights reserved.
82 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95

Table 1, which is adapted from Dehghan and Ruwanpura [1] and further
Construction–construction
completed, shows a brief comparative review of the existing literature.
The next sections provide further details.

2.1. Overlapping in product development

Research by Nicoletti and Nicolo [4] contributes to planning concur-

rent execution by enhancing information flow between interdependent

X
activities. Loch and Terwiesch [3] and Terwiesch et al. [6] studied the
Design–construction

importance of communication and information exchange in concurrent

engineering and presented an analytical model to address two ques-
tions: 1) how much is the optimal overlap? and 2) how should the over-
lapped activities be coordinated? The model of concurrent engineering
offered by Loch and Terwiesch [3] is more analytical than numerical. In
Type of overlapping

addition, they do not include the cost of rework in their model, which is
Design–design

very important as often the cost of rework is higher than the cost of
extra communication.
Information exchange and coordination are the focus of all research
performed by Prasad [5], Ha and Porteus [15], Nicoletti and Nicolo [4],
X
X
X
X
X
X
X

X
X
X
X

Loch and Terwiesch [3] and Terwiesch et al. [6]. Activity characteristics
Cascade of

are less investigated in their research. However, other researchers,

overlaps

such as Krishnan et al. [2,16,17] have focused on activity characteristics

to find out which activities are the best fit for overlapping. Krishnan
X
X

et al. [17] introduce and formulate two characteristics for upstream

Multi-predecessor

and downstream activities, and use them to determine how activities

must be overlapped. The first characteristic is upstream evolution,
which refers to how fast the upstream information is refined and final-
overlaps

ized. The second characteristic is downstream sensitivity, which refers

X
X
X
X

to how much downstream activity is sensitive to possible changes in

Group of overlaps

upstream activity and how quickly downstream activity can accommo-

A single chain
of activities

date those changes. The significance of Krishnan's model is that in spite

of its simplicity, it is a good representation of real-world practice. The
concepts of evolution and sensitivity, and their combination to define
X
X
X
X
X
X

different situations for overlapping, make a large contribution to under-

Type of study

overlapping

standing the mechanism of overlapping.

Individual
isolated

2.2. Overlapping in construction projects

X
X
X
X
X
X
X
X
dependent

Pena-Mora and Li [10] have conducted a highly contributing study

activities
Between

about overlapping in construction projects. They use the concepts of

upstream task evolution and downstream task sensitivity formerly
X
X
X

X
X
X
X
X
X
X
X
X

developed by Eppinger [9] and Krishnan et al. [17], to generate a frame-

Type of activities

interdependent

work suitable for construction activities. Instead of the concepts of

upstream task evolution and downstream task sensitivity in product
activities
Between

development, Pena-Mora and Li considered upstream/downstream

production rate, upstream production reliability, and downstream task
X

sensitivity. Their framework provides guidance for any type of overlap-

Construction

ping including design–design, design–construction, and construction–

construction.
The research by Bogus et al. [11–13] adds more details and more
X
X
X

X
X
X
X
X

insight to the research by Krishnan et al. [17], but with an orientation

Type of industry

toward construction projects. The research is limited to the design

Manufacturing
Comparative review of the existing major literature.

phase and considers information dependency between activities. Its

results are more useful for design managers to decide subjectively if a
pair of activities are suitable for overlapping. Blacud et al. [14] added
X
X
X
X
X

more to the research of Bogus et al. [12] by expanding the concept of

Dehghan and Ruwanpura (2014)

evolution and sensitivity to the construction phase. Blacud et al. [14]

Berthaut et al. (2011a, 2011b)

studied the overlapping of design activities as predecessor activities

Nicoletti and Nicolo (1998)

Loch and Terwiesch (1998)

Roemer et al. (2000, 2004)

with construction activities as successor activities. They focused their

Pena-Mora and Li (2001)

Gerk and Qassim (2008)

Terwiesch et al. (2002)

Cho and Hastak (2013)

research on determining the factors contributing to the sensitivity of

Krishan et al. (1997)

Blacud et al. (2009)

Bogus et al. (2005)

construction activities.
Eppinger (1997)

Common to all research studies performed by Krishnan et al. [17],

Researchers

Pena-Mora and Li [10], Bogus et al. [12], and Blacud et al. [14] is that
This Paper

they focus on one individual overlap and do not consider the overlap-
Table 1

ping in the context of a project schedule and with regard to other over-
laps. They do not provide any clue to which activities are better to be
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95
overlapped and which activities are not; nor among those which are
overlapped, which activities can be more overlapped than others since
they generate less risk and less cost. Roemer et al. [7] and Roemer and
Ahmadi [8] have tried to provide answers to these questions.
2.3. Overlapping time–cost trade-off
Roemer et al. [7] and Roemer and Ahmadi [8] took a different
approach from Krishnan et al. [17] and tried to address the design
costs of overlapping by evaluating the trade-off between overlapping
lead times and overlapping costs (Roemer and Ahmadi's research has
a focus on product development, not construction projects). Roemer
and Ahmadi [8] could clearly explain the concept of generating rework
as a result of overlapping. Furthermore, their effort to optimize overlaps
in a full chain of activities distinguishes them from the majority of
researchers who only investigated overlapping between two activities
isolated from other activities. However, they limited their research
to only one chain of activities (i.e. one path). Therefore, they did not
study a network of activities with multiple paths and with multiple
predecessors to some activities. Another limitation is that their optimi-
zation does not cover a condition in which two overlaps overlap
(concurrent overlapping or cascade of overlaps).
Similar research has been conducted by Gerk and Qassim [18] who
introduced a mixed-integer nonlinear programming model for the
acceleration of projects. They suggest that project acceleration is possi-
ble through three different techniques: activity crashing, activity
overlapping, and activity substitution. Gerk and Qassim [18] modeled
this problem and developed an objective function aimed at minimizing
the total cost of project acceleration, and a large number of constraint
functions. The details of network calculations and optimization compu-
tations were not explained in Gerk and Qassim's paper, but the exam-
ples they have offered indicate that their model has similar limitations
to Roemer and Ahamadi's model [8].
The most recent research, with similarities to the research by
Roemer and Ahamadi [8] and Gerk and Qassim [18], is conducted by
Berthaut et al. [19,20]. This research investigated the different feasible
modes of overlapping, taking resource constraints into account. They
showed the close interaction of overlapping modes and resource con-
straints. In their model, they have not considered the effect of parallel
overlaps, but they have addressed the effect of multi-predecessor over-
laps. Their approach to a multi-predecessor effect in general differs from
what Dehghan [21] suggested, and Dehghan and Ruwanpura [1] later
criticized it for lack of accuracy in evaluating the rework duration
resulting from multiple predecessors.
Cho and Hastak [22] developed their time- and cost-optimized
decision support model (TACTICS) based on fast-tracking methodology
and genetic algorithms. In their model, they tried to optimize overlap-
ping at the level of design and construction work packages, not activi-
ties. In addition, they further divided the construction work packages
into more specific work packages by considering the concept of space
zoning — availability of working space in the construction site — which
is a new approach. The major flaw in their study relates to the effect of
rework: central to the concept of activity overlapping is that overlapping
increases the risk of changes and rework, and rework negatively affects
both project cost and project duration. This is the fact that makes over-
lapping a difficult area for research. As Cho and Hastak [22] did not
consider rework resulted from overlapping (which can be drastically
high for construction activities), their model makes only a limited contri-
bution to both theory and practice.
Khoueiry et al. [23] and Srour et al. [24] presented overlapping
optimization models based on activity dependency information; that
is, the evaluation rate of upstream and sensitivity of downstream
activities. These research studies also assume that a project has only
one chain of activities.
Hazini et al. [25] explained a heuristic method to determine optimum
degree of activity accelerating and overlapping in schedule compression
83
by considering all network paths, and Hussain and Chua [26] presented
an optimization model to overlap design and construction activities
with minimal rework. The advantage of their effort over previous efforts
is that they consider multiple dependencies (multi-predecessors) when
optimizing the project schedule. However, they still do not address the
cascade of overlaps, in which overlaps themselves overlap.
Dehghan and Ruwanpura [1,27] have introduced a model of trade-
off between overlapping and rework of design activities addressing
how to handle both multi-predecessors and cascades of overlaps.
Studies conducted by Dehghan [21], Dehghan et al. [28,29] and
Dehghan and Ruwanpura [1,27] focus on the costs and benefits of over-
lapping and how to formulate them. Dehghan and Ruwanpura [1] have
the most recent study in the field. Their study is, to some extent, inspired
by the models developed by Roemer et al. [7], Roemer and Ahamdi [8]
and Gerk and Qassim [18], in that all have proposed models optimizing
overlaps together, not separately. What differentiates Dehghan and
Ruwanpura [1] is that they also demonstrated: 1) how to handle multi-
path networks through introducing a “multi-predecessor effect” and
2) how to handle concurrent overlapping through introducing the “cas-
cade effect”. The results of their work are the starting point of this paper.
Specifically, we have developed and computerized an overlapping
optimization algorithm based on the model developed by Dehghan and
Ruwanpura [1,27]. However, their model like any model is only a repre-
sentation of reality, clarifying the mechanism and characteristics of
design activity overlapping, but is unable to identify practically which
activities have to be overlapped and to what extent to reduce the project
duration at the minimum cost. For this purpose, an algorithm and a
practical tool are required to bridge the model to the real world applica-
tions. The objective of this paper is to introduce such an algorithm and
tool which are able to:
• Determine which activities have to be overlapped
• Determine the degree of overlapping
• Handle multi-path networks
• Handle cascade of overlaps
• Take into account all activities, critical and non-critical, and follow the
critical path if the critical path changes or new critical paths emerge
• Handle all types of activity dependencies including finish-to-start
(FS), start-to-start (SS), finish-to-finish (FF), and start-to-finish (SF).
3. Method: overlapping time–cost tradeoff
Fig. 1 [1] shows the mechanism of overlapping two dependent activ-
ities, in which the start of an activity depends on the finish of another
Fig. 1. The mechanism of activity overlapping [1].
84 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95

activity and the second activity can only be started if the first activity is Lij Duration of the overlapped interval between predecessor
finished completely. This is because the successor needs the information activity i and successor activity j
generated by the predecessor (e.g.: the predecessor activity can be “pre- Pij The probability that a change happens for predecessor activity i
paring pump layout drawings” by the mechanical engineer and the suc- during its overlapping with successor activity j and the change
cessor activity can be “preparing pump foundation drawings” by the causes some rework for successor activity j
structural engineer). However, to compress the schedule, the successor Tij The extended duration added to successor activity j, as a
activity may be intentionally started before the completion of its prede- result of rework originating from the changes made by prede-
cessor. This becomes possible if the predecessor activity releases some cessor activity i, during its overlap with successor activity j
preliminary information before its completion to the successor activity. Rij The equivalent rework duration for successor activity j, as a
Therefore, the successor can start sooner, using the preliminary infor- result of its overlapping with predecessor activity i.
mation and making necessary assumptions and predictions. The two Based on the explanation of the mechanism of overlapping, both the
activities can proceed in parallel for a while and, during this period, probability, Pij, and amount of rework, Tij, are functions of the overlap-
some intermediate information may also be transferred until the prede- ping duration, Lij. Therefore, the equivalent rework, Rij, is also a function
cessor is completed; then, the predecessor will release its final informa- of Lij.
tion to the successor. At this point, it is likely that the final information is


different from the preliminary or intermediate information and therefore, Ri j ¼ f L i j ð3Þ
changes and adjustments must be made to the successor to make it com-
patible with the final information. The changes and adjustments will take
some additional coordination and work (rework) in form of extra person-
hours (i.e. extra cost and time), which means an increase in the duration 3.1. Overlapping time–cost trade-off objective function
of the successor activity compared to its normal duration [1].
Two points are noteworthy. First, overlapping has a maximum dura- Dehghan and Ruwanpura [1] also formulated the overlapping time–
tion beyond which further overlapping is actually impossible because cost trade-off (TCT) objective function which is central to solve the
no preliminary information of any kind can be produced by the prede- overlapping optimization problem as follows:
cessor activity. In the current research, this amount is called maximum
n X
X 
overlapping or maximum allowable overlapping [1]. m
 C l f ðT−T t Þ if TNT t otherwise 0:0
C¼ P i j T i j W j þ Ei j þ : ð4Þ
Second, no rework occurs if the final, preliminary, and intermediate Be f ðT−T t Þ if TbT t otherwise 0:0
i¼1 j¼1
information are compatible. Rework is probable to take place. The prob-
ability of rework depends on several factors. The literature review
shows—and the interviews and focus groups endorse—that these factors In which:
are the type and complexity of overlapped activities, their relation with
other activities in the project schedule, and the amount of overlapping. C Cost (Benefit) of overlap
On the other hand, the amount (the duration) of additional work is a Pij The probability that a change happens for predecessor activity i
function of the overlapping duration, the strength of the successor activ- during its overlapping with successor activity j and the change
ity dependency upon the predecessor activity, and the intensity of causes some rework for successor activity j
nonconformity between final and preliminary information. The maxi- Tij The extended duration added to successor activity j, as a
mum rework may happen when the final information is significantly result of rework originating from the changes made by prede-
contradictory with the preliminary information, and the successor is cessor activity i, during its overlap with successor activity j
required to apply a major change. In such a situation, the worst scenario Wj Total daily wage for successor activity j, including daily
is that the successor must disregard all its progress during overlapping salaries and daily overheads
and start over. Therefore, the rework duration cannot be logically Eij Extra costs, other than daily wages and overheads, imposed
more than the overlapping duration [1,27]. on successor activity j or on other project areas (design,
Since the impact of the change on the predecessor activity i in Fig. 2 procurement, construction, etc.) because of the changes
is the extended (rework) period for successor activity j, the equivalent made by predecessor activity i during its overlapping with
rework is defined by multiplying the probability of rework and the successor activity j
duration of rework as below: Bef Daily benefits of project early finish
Clf Daily costs of project late finish
Ri j ¼ P i j  T i j : ð1Þ T Project duration
Tt Project target duration.
Subject to:
Noting that:
0 b Pij ≤ 1 sets the limit for the rework probability
0≤P i j ≤1⇒0≤Ri j ≤T i j : ð2Þ Clf ≥ 0 ensures no reward for late finish
Bef ≤ 0 ensures no penalty for early finish
In which: Tt N 0 maintains the target duration not be negative or zero
Pij = 0 if i is not the direct or indirect predecessor to j
Lij b Lij,max ensures no overlapping beyond the maximum allowable.
In which:

Lij Duration of the overlapped interval between predecessor

Fig. 2. Overlapping duration vs. successor added duration [1].
activity i and successor activity j
Lij,max Maximum allowable overlapping between predecessor activity
i and successor activity j
Eq. (4) simply states that the total cost of an overlapping strategy,
is the sum of the individual overlap costs, plus the cost of project late
completion (or minus the benefit of project early completion).
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95
The main objective of this paper is to indicate how overlapping
optimization can actually be performed using the overlapping
time–cost trade-off objective function (Eq. (4)). This is inherently a
multi-objective (bi-objective) optimization problem, as both time and
cost have to be minimized at the same time; while not independent,
they are intricately related. Due to the large number of possible overlaps
and extremely large number of possible strategies, the optimization
process will be complex and extensive. As a result, a robust optimization
technique or algorithm (for example, an evolutionary optimization al-
gorithm) is required to deal with such a complex, multi-objective and
multi-parametric problem. The objective of this paper is to introduce
and explain such an optimization algorithm and its computerization.
The computer tool works based on the overlapping optimization
algorithm, assessing various overlapping strategies and identifying the
least expensive strategies. The tool is actually a cost evaluation module
linked to a commercial project scheduling software (MS Project). This
computer tool is so user-friendly that any scheduler or cost controller
can easily run it and modify the schedule accordingly. The computer
tool is unique and new, as so far no similar tools exist in industry or
academia. It can optimize overlaps in large and complex project sched-
ules in fairly short processing times. It is able to handle multi-path
networks and all types of activity dependencies. The tool takes all activ-
ities, critical and non-critical, into account and follows the critical path if
the critical path changes or new critical paths emerge. The tool can also
take into account resource limitations and schedule constraints.
Our suggested algorithm uses the principles of genetic algorithms
and is partially inspired by the algorithm developed by Hegazy [30]
for activity crashing: both use the principles of genetic algorithms as
the optimization module; however, our algorithm is more complex
than Hegazy's, as the nature of activity overlapping is more complicated
than activity crashing. These complexities are related to the existence of
overlapping rework, heterogeneous information inside genes, multi-
predecessor and cascade effects [1], and modifications and relaxations
required on network calculations. Therefore, the majority of this paper
explains and clarifies the different elements of the overlapping optimi-
zation algorithm. We use an illustrative sample network to explain the
suggested algorithm step by step and with full details. This network
was used by Dehghan et al. [21,28,29] to introduce the preliminary
results of their research and was designed in a way that includes a vari-
ety of network phenomena: various types of activity dependencies
(FS, SS, FF), multi-predecessor overlaps, cascade overlaps, and change
of critical paths. The intention of using the same network in this paper
is that the interested reader can better understand the algorithm and
can implement it on similar problems. Later in the paper, however, we
will optimize a real-world network to validate the suggested algorithm.
3.2. Illustrative sample network
Fig. 3 shows a network including seven activities and nine finish-to-
start relationships. The original duration of each activity is also included.
Fig. 3. The sample network.
85
This network encompasses activities with one, two and three prede-
cessors (e.g. activity 2, activity 6 and activity 7 respectively). In addition,
activity durations have been defined in such a way that all activities may
become critical as a result of increasing overlapping between them. The
normal duration of this network when there is no overlapping between
activities is 110 days and its critical path is: 1 → 2 → 6 → 7.
The objective is to compress the network by applying overlaps
between activities. This can be achieved in two different ways: The
first is when a finish time is given and the least expensive overlapping
strategy to meet the finish time is requested. The second is when an
additional cost is given and the overlapping strategy with the shortest
duration is desired. For both cases, some additional information is
required. For instance, the maximum allowable overlapping between
each pair of activities, the amount of rework generated by each overlap,
and the estimated cost of each overlap must all be known. This informa-
tion, as well as other necessary information about the sample case study,
were suggested by Dehghan et al. [21,28,29] and presented in Table 2.
The table shows activities, their durations, the predecessor(s) for
each activity, and the type of relation activities have with each other,
before and after overlapping. In real-world practice, overlapping is
applied mainly through finish-to-finish (FF) or start-to-start (SS) rela-
tionships with a positive lag, and sometimes through finish-to-start
(FS) relationship with a negative lag. In the sample case study, a similar
strategy is utilized. As shown in Table 2, some activities are overlapped
by means of FS relationships with negative lags (e.g. activities 1 and 2),
some of them are overlapped using SS relationships with positive lags
(e.g. activities 1 and 4), and the rest are overlapped through FF relation-
ships with positive lags (e.g. activities 1 and 4). For lag times, Table 2
also highlights the minimum allowable: for example, the minimum al-
lowable lag time for the FS relation between activity 1 and activity 3 is
−10. The amount for the SS relation between activity 2 and activity 5 is 8.
With this information, the maximum allowable overlapping can
easily be calculated. Any overlapping beyond the maximum allowable
is considered to be technically impossible. Alternatively, any overlap-
ping beyond the minimum lag time is technically impossible, because
for each of FS, FF, and SS relations the minimum lag is equal to the max-
imum overlap as per Eqs. (5), (6) and (7). The relation between the lag
time and the overlapping duration are further detailed in Section 3.2
and particularly Eqs. (6), (7) and (8).
Overlapping rework durations as a function of overlapped amount
and overlapping cost values as a function of rework durations have
been also provided in Table 2. These functions are required to perform
time–cost trade-off computations. The cost of overlapping and the
rework duration are functions of overlapping degree. In this sample
network, various arbitrary types of functions have been used to present
the overlapping rework and overlapping cost. These functions are in
compliance with the data extracted from the design phase of industrial
construction projects as stated by Dehghan and Ruwanpura [1]. This
means: 1) The rework ranges are between 5 and 30% of overlapping
durations; 2) The cost values have a linear relation with rework dura-
tions (therefore, cost functions are ultimately a function of overlapping
durations) [1].
In addition, the following values are given [21,28,29]:
– Project targeted completion date (Tt) = 105 days
– Project daily benefits for early completion (Bef) = $1000 per day
– Project daily losses for late completion (Clf) = $1000 per day
– Change transfer ratio (rABC) = 0 (means that changes causing rework
only affect the immediate successor activity within a cascade of
overlaps).
4. Overlapping algorithm
In this section and its subsections the overlapping algorithm is
explained in detail. However, it is emphasized that the algorithm has
86 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95

Table 2
General attributes of the sample case study.

Activity Duration Predecessor Original relation original lag Relation after overlap Minimum allowable lag Overlapping rework function Overlapping cost function
Rij = f(Lij) Cij = g(Rij)

1 36 – – – – – – –
2 30 1 FS 0 FS −17 R12† = 0.2L12* C12‡ = 900R12
3 28 1 FS 0 FS −10 R13 = 0.1L13 C13 = 850R13
4 48 1 FS 0 SS 19 R14 = 0.15L14 C14 = 1050R14
5 22 2 FS 0 SS 8 R25 = 0.3L25 C25 = 800R25
6 24 2 FS 0 FS −20 R26 = 0.25L26 C26 = 1000R26
6 24 3 FS 0 FF 6 R36 = 0.2L36 C36 = 1000R36
7 20 4 FS 0 FF 8 R47 = 0.15L47 C47 = 950R47
7 20 5 FS 0 FS −15 R57 = 0.05L57 C57 = 950R57
7 20 6 FS 0 SS 10 R67 = 0.25L67 C67 = 950R67

* Lij: The duration of overlapping between i and j.

† Rij: The rework duration of the successor activity j as a function of Lij, Rij = f(Lij).
‡ Cij: The cost of overlapping between i and j as a function of Rij, Cij = g(Rij).

two main variations, as the overlapping time–cost trade-off problem
has two main variations. According to Dehghan and Ruwanpura [1],
the first variation is when the project target duration (Tt) is given, and
an overlapping strategy with the minimum cost (or maximum profit)
is required. Therefore, the time–cost trade-off problem is to find over-
lapping strategies that generate the minimum project costs.
The second variation is when a project target cost Ct (maximum
acceptable cost or minimum desirable profit) is given, and an overlap-
ping strategy resulting in minimum project duration is asked for.
In this case, the time–cost trade-off problem is finding overlapping
strategies that minimize project duration but do not exceed the
predetermined project cost limit (Ct). The first variation is a more
common problem in the design and execution of construction projects.
In this paper, the algorithm for the first variation is explained in
detail; the algorithm for the second variation is only slightly different.
Fig. 4 shows the flowchart of the suggested overlapping algorithm
for minimizing the project cost within the project target duration. A
part of the algorithm uses the principles of genetic algorithms (GAs)
optimization in order to solve overlapping time–cost trade-offs. The
algorithm encompasses extensive recursive calculations and numerous
trial and errors that require computer implementation. Two major
calculation modules exist in the algorithm. One module performs
cost calculations and the other performs schedule computations and
evaluates project time. The algorithm itself has two main steps. The
first step is to generate a collection of random potential solutions
(shown as “Generating the initial population” in Fig. 4) and the sec-
ond step is to evolve the random solutions to more favorable solu-
tions (shown as “Generating the offspring population” in Fig. 4).
The following subsections further clarify the flowchart (algorithm)
of Fig. 4.
4.1. Input data
From the required input data to start the algorithm, activity dura-
tions and activity dependencies are primary network variables required
to build the project network. Minimum allowable lag times, rework
duration functions (or overlapping rework durations), and change
transfer ratios are complementary network variables. Overlapping cost
functions (or cost values for different degrees of overlapping), project
daily benefits for early completion (Bef) and project daily losses for
late completion (Clf) are cost variables that, along with the targeted
project duration (Tt), are used to calculate the total costs/benefits
of overlapping. GA parameters are genetic algorithm variables and
their variations can change the efficiency of the calculations. Main
GA parameters are the desired initial population size, the desired
number of offspring reproductions, and the mutation-to-crossover
ratio. These parameters will be further clarified through the next
sections.
4.2. Generating initial population
An overlapping strategy is a collection of individual overlaps. Each
overlapping strategy is presented by a chromosome and is a potential
solution to the problem of reducing the project duration to the targeted
time. For the first step, a number of random chromosomes (overlapping
strategies) that meet the “time” criterion should be generated.
4.2.1. Generating random chromosomes
Each chromosome represents an overlapping strategy and encom-
passes several genes. Each gene represents a relation between two
activities and each relation is a potential overlap. Like a real natural
gene which holds the information required to build and maintain an
organism's cells, the overlap genes hold the required information to
build the project network and evaluate its duration and cost. Each
gene consists of eight pieces of information as follows:
1. Predecessor activity
2. Successor activity
3. Successor activity duration
4. Type of relation between the predecessor and the successor
5. Lag time
6. Overlapping duration between activities
7. Rework duration for the successor activity
8. Cost of overlapping as a result of rework.
Fig. 5 shows a random chromosome from the sample network with
the above eight items. The chromosome has 9 genes and each gene has
8 pieces of information. Item 1 indicates which activity is the predecessor
and item 2 indicates which activity is the successor. Item 3 shows the
original duration of the successor activity and item 4 determines what
the relation between the two activities will be if they overlap. These
four pieces of information are extracted from the original project
network—i.e. the network at its normal condition. The fifth piece of infor-
mation is lag time. The lag time is generated quite randomly because the
chromosomes of the initial population must be random. Then the over-
lapping duration (item 6) is calculated using the lag time, the predecessor
activity duration and the successor activity duration. Calculating over-
lapping duration depends on the type of dependency between the
two activities. For example, the overlapping duration of Gene no. 4 is
30 − 11 = 19 because for an SS relation:
Overlapping duration ¼ predecessor duration–SS lag: ð5Þ
For an FS relation the overlapping duration is calculated as fol-
lows:
Overlapping duration ¼ –FS lag: ð6Þ
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95 87

Fig. 4. The overlapping optimization algorithm for minimizing project cost.

And for an FF relation:
Overlapping duration ¼ successor duration− FF lag:
The overlapping duration can be any amount between zero
(no overlapping) and the maximum allowable overlapping. While
integer numbers are the preferred way of showing activity durations
in this research, the developed algorithm can deal with any type of
ð7Þ numbers, whether integer or real. As it is normal practice to express
duration in terms of “day” in construction projects, this time unit is
used all throughout this paper; since one day is small enough for a
construction project, half-days or hours are not used. Therefore, the
88 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95

Fig. 6. The random chromosome with corrected rework durations.

Fig. 5. A random chromosome.
durations can be expressed as integer numbers presenting the number
of days. For example, the overlapping duration between activity 1 and
activity 2 can be any integer number between 0 and 8.
When the amount of overlap is available, its associated rework and
cost for the successor activity can be obtained through rework and
cost functions (Table 2). For example, the overlapping value of gene
#4 is 19. According to Table 2, the rework function for this gene is
R25 = 0.3L25, which means that the rework for activity 5 should be
6 days:
The above rework corrections are applied on the chromosome and
result in the chromosome shown in Fig. 6. With these corrections, the
associated costs will also be changed. Details on how to calculate the
total cost of overlaps will be explained in the next sections.
4.2.3. Cascade effect
The rework durations may still need corrections if any cascades of
overlaps exist. The sample network has the potential to generate some;
in fact, any paths with three or more activities may have a cascade of over-
laps. The overlapping algorithm identifies such paths and checks whether
any cascades of overlaps exist. Potential paths that may generate cascades
of overlaps can be identified from the chromosome:

R25 ¼ 0:3  19 ¼ 5:7—roundup→R25 ¼ 6: Path 1: 1 → 2 → 5 → 7

According to Table 2, the cost function for this gene is C25 = 800R25.
Therefore, the cost of rework is $4800:
C25 ¼ 800R25 ¼ 800  6 ¼ 4800:
Similar to lag and overlap durations, the rework durations should be
integer numbers. Therefore, the resulted values from rework functions
are rounded to the closest integer number.
4.2.2. Multi-predecessors effect
With regard to rework duration, genes #5 and #6 have a different
story. They reflect a successor activity (activity 6) with two predecessor
activities (activity 2 and activity 3). In Fig. 5, the rework duration of
activity 6 in gene 5 is 5 days, while the same rework in gene #6 is
3 days. But activity 6 should only have a unique rework duration. There-
fore, its rework should be corrected by calculating the equivalent
rework duration. According to [1], the equivalent rework for activity 6
can be calculated using Eq. (5):
 
R11 þ R21 R11 R21
R1 ¼ −  L11 :
L11 L11 2
Path 2: 1 → 2 → 6 → 7
Path 3: 1 → 3 → 6 → 7
Path 4: 1 → 4 → 7.
A cascade of overlaps happen if the duration of an activity is less than
its total overlap duration with its predecessor and successor activities.
To identify all cascades, every 3 consecutive activities on each path
should be checked. For cascades of more than 2 overlaps, a similar
approach can be taken. Mathematically, if the total duration of all over-
laps is more than the total duration of all intermediate activities bounded
by the first and last activities, a cascade of overlaps occurs [1].
4.2.4. Network (schedule) calculations
The next step in the overlapping algorithm (Fig. 4) is forming
the network and performing scheduling calculations. For this purpose,
activities original durations, dependencies (FS, SS, or FF), lag times,
and rework durations are used to form the network.
For example, the sample chromosome from Fig. 6 contains the
required data to form the network of Fig. 7.
To form the network, activity durations and activity relations are
adjusted according to the data within the genes. The rework duration
ð8Þ

20 − 202 Þ  20 ¼ 7:25≈7:
Rework for activity 6 ¼ ð5þ3 53

Likewise, genes #7, #8 and #9 show that activities 4, 5 and 6 are

three predecessors to activity 7. The rework duration of activity 7 should
be corrected as well. According to [1], the equivalent rework for activity
7 can be calculated using Eq. (6):
 
R11 þ R21 þ R31 R11 R21 þ R11 R31 þ R21 R31 R11 R21 R31
R1 ¼ − þ  L11 : ð9Þ
L11 L11 2 L11 3

Rework for activity 7:

 
4þ1þ1 41þ41þ11 411
¼ − 2
þ 3
 14 ¼ 5:38≈5:
14 14 14 Fig. 7. The network of the sample chromosome of Fig. 6.
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95
of each activity must be added to its original duration. As a result, activity
durations might be different from their original amounts. For example,
the duration of activity 4 is now 50 days instead of 48 days, because
2 days rework is considered a result of 14 days overlap between activity
4 and activity 1. The durations of activities 5, 6 and 7 are likewise
increased.
Rework durations affect finish-to-finish dependencies as well.
Finish-to-finish relations should be adjusted because when the successor
activity duration is increased, its FF lag time with the predecessor activity
is also increased. As shown in Fig. 7, the lag time of the FF relation
between activity 3 and activity 6 is now 15 days instead of 8 days, as
the duration of activity 6 is now 7 days longer due to rework. Conse-
quently, activity 6's FF relation with activity 3 should be 7 days longer
as well. Likewise, the lag time of the FF relation between activity 4 and
activity 7 is now 12 + 5 = 17.
The rework durations do not affect FS and SS lag times and these lag
times remain unchanged.
Network calculations encompass forward and backward calcula-
tions in order to determine the earliest and latest start and finish time
of each activity, free floats and total floats of each activity, total network
duration, and the critical path(s). The sample chromosome leads to a
network with 89 days total duration. The critical path is changed from
1 → 2 → 6 → 7 (Fig. 3) to 1 → 4 → 7 (Fig. 8). The earliest and latest
start and finish time of each activity along with activities free floats
and total floats are shown in Fig. 8.
According to the network logic, some of the dependencies are
driving dependencies and others are not. For example, activity 6 has
two dependencies with its predecessors. The FF15 dependency between
activity 3 and activity 6 is a driving dependency, because the start of
activity 6 depends on it. By contrast, the FS-20 dependency between
activity 2 and activity 6 is not a driving dependency, because it does
not trigger the start of activity 6. If the lag time of this dependency is in-
creased to −16, which is equal to reducing the overlap between activity
2 and activity 6 by 4 days, this dependency becomes a driving depen-
dency. Therefore, there are 4 days of unnecessary overlapping between
activity 2 and activity 6 that may generate extra cost, but do not affect
the start date of activity 6 and consequently the project finish time. To
avoid these types of overlaps, the lag times of all non-driving dependen-
cies should be increased until those dependencies become driving. By
such a relaxation on lag times, unnecessary overlaps which do not
reduce the project finish time but impose extra risks and costs are
avoided. Therefore, the lag times modified based on the network logic
become different from the original random lag times. Fig. 9 shows the
modified network using modified lag times. Although some of the over-
laps have been reduced, the total project duration is still 89 days.
Changes in lag times may cause the associated rework durations to
change as well. This means that activity durations may also change.
For example, when the dependency between activity 2 and 6 is changed
from FS-20 to FS-16, the resulting rework for activity 6 will be reduced
from 5 to 4 in keeping with the relevant rework function from Table 2,
Fig. 8. The sample network with the earliest start and finish times.
89
Fig. 9. The network of Fig. 8 whose lag times are modified to avoid unnecessary overlaps.
R26 = 0.25L26. Consequently, the duration of activity 6 is decreased as
well.
Therefore, rework durations and activity durations have to be
recalculated and the chromosome should be constructed again. The
chromosome associated with the network in Fig. 9 is reconstructed in
Fig. 10. Multi-predecessor and cascade effects have been incorporated
in this chromosome.
Comparing Fig. 10 with Fig. 6 shows that the rework durations, and
consequently the cost values are decreased as a result of modifying lag
times. This modified chromosome has to go through the same process
previously explained: forming the network, recalculating activity dura-
tions, recalculating FF dependencies durations, determining the earliest
and latest start and finish times of each activity, recalculating activities
free floats and total floats, recalculating the total network duration
and determining the critical path(s). If necessary, lag times should be re-
laxed again to prevent unnecessary overlaps. Therefore, the above pro-
cess might be repeated several times, until all lag times are completely
modified and remain unchanged. Then the final network/chromosome
is ready for a time check followed by a cost evaluation.
4.2.5. Time check
Each chromosome results in one project duration, which is defined by
network calculations. The algorithm can perform a time check if required.
For example, the requirement might be only accepting project durations
equal to or less than a targeted project duration. If the duration is equal
or less than the project target duration (T ≤ Tt), then the chromosome is
acceptable and passes to the next step, otherwise the chromosome is ig-
nored and the previous step to generate a new chromosome is repeated.
In the sample case study, the target duration is 105 days which means
that project durations more than 105 days are not acceptable. To investi-
gate the total cost for all project durations, the time check can be waived.
4.2.6. Cost calculation
As explained above, each chromosome represents a solution, which
is a unique overlapping strategy. Each chromosome generates costs
because of overlaps and at the same time may generate benefit because
Fig. 10. The chromosome associated with the network in Fig. 9.
90 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95
of timesaving. The total amount of cost (or benefit) of chromosomes can
be calculated using the main objective function (Eq. (4)). Individual
costs of overlapping are obtained from the genes (Fig. 10) and totalled.
Also, the cost/benefit of a project late or early finish time is calculated
and added to overlapping costs.
For the chromosome of Fig. 10, the total cost is calculated by sum-
ming up the cost of rework for all 7 activities. Cost of rework for activity
1 is zero as it does not have any rework (this cost is not reflected in the
chromosome). Cost of rework for activities 2, 3, 4, 5, 6, and 7 is 0, 0, 2100,
4800, 6000 and 2850 respectively. Therefore:
Overlapping cost ¼ 0 þ 0 þ 0 þ 2100 þ 4800 þ 6000 þ 2850 ¼ 15750:
Of which, $6000 is the cost value of genes 5 and 6. Since both genes
include the rework cost for activity 6, only one of them is taken into
account to calculate the total cost. The same is true for $2850 which is
the cost value of genes 7, 8 and 9. Since all of them are the rework cost
for activity 7, only one of them should be taken into account to calculate
the total cost.
The timesaving benefit for early completion is calculated as follows:
Timesaving benefit ¼ 1000 dollars=day  ð105–89Þ days ¼ 16000:
Finally, the total net cost is:
Total cost ¼ 15750–16000 ¼ −250:
Therefore, the chromosome represents an overlapping strategy
which results in $250 net benefit.
4.2.7. Initial population
The generated chromosome, along with the project total duration
and project total cost it generates, is recorded. The process of generating
random chromosomes is repeated until a suitable number of chromo-
somes, perhaps 100, is generated. These chromosomes form the initial
population of solutions. The population encompasses completely
random overlapping strategies with various time and cost attributes.
In the next steps, the initial population is used to reproduce offspring
population with targeted cost and time attributes.
4.3. Generating offspring population
The initial population of solutions must evolve to generate better
solutions. This is possible by exchanging the stronger solutions' genes
with each other (marriage or crossover) to generate new solutions
(offspring genes). New solutions are evaluated and, if they are better
than the weakest solutions in the population, will replace them. This
process of generating better and better solutions is repeated until a
satisfactory population is generated. The fittest members of the popula-
tion are the best solutions.
4.3.1. Generating offspring chromosomes
Generating offspring chromosomes in genetic algorithms is per-
formed through crossover and mutation, resembling natural evolution.
To generate an offspring chromosome via crossover, the algorithm
randomly selects two members of the initial population as parent chro-
mosomes, and interchanges their genes. The crossover takes place by
randomly selecting some of the genes of one chromosome and exchang-
ing with the same genes in the other chromosome [31–34].
Unlike crossover, which resembles the common natural way of
reproduction [31], mutation is an odd phenomenon that resembles
the process of a sudden generation of an exceptional offspring that
may turn out to be particularly fit. Mutation is performed by randomly
selecting one chromosome from the initial population and then arbi-
trarily changing some of its genes. Essentially, what mutation does is
release the optimization process from being trapped in local optima
[30]. Chromosomes generated through crossover or mutation have to
undergo the same process as initial random chromosomes.
4.3.2. Network (schedule) calculations, time check, and cost calculation
The associated calculations with an offspring chromosome including
network calculations, time check, and cost calculations are essentially
the same as such calculations for a parent chromosome. Therefore,
what was described in previous sections are also applicable for offspring
chromosomes.
4.3.3. Cost check
When an offspring is generated by crossover or mutation and its cost
is calculated, a fitness evaluation should take place to ensure the
offspring chromosome is acceptable. This is performed by comparing
the cost (benefit) of the offspring chromosome with other chromo-
somes; if it is better (lower cost or higher benefit) than the worst chro-
mosome in the population, the offspring replaces the old chromosome.
Otherwise, the offspring is ignored.
4.3.4. Offspring population
The process of generating offspring chromosomes should be con-
tinued until a number of optimum chromosomes that best fit the objec-
tive function emerge. Miscellaneous termination criteria exist to stop
offspring generation. In the current algorithm, the number of genera-
tions determines when the process should be stopped. The final off-
spring population includes a range of best solutions. At this point, the
first variation of the overlapping optimization algorithm ends.
4.4. Overlapping algorithm: second variation
The second variation, the overlapping algorithm for minimizing the
project duration within the project targeted cost, is only slightly different
from the first variation. As shown in Fig. 4, the “time check” in the “gener-
ating the initial population” step of the first variation is changed to a “cost
check” in the second variation. The checking ensures that the generated
chromosome meets the cost constraint imposed. The rest of the flow-
charts are similar to each other. However, the time check and cost
check in the “generating the offspring population” step are performed in
different ways for different objectives. In the first variation, the time
check is to ensure that the imposed time constraint is met and the cost
check is to verify if the generated chromosome is better than the worst
chromosome in the population. In the second variation, the cost check is
to ensure that the imposed cost constraint is met and the time check is
to make sure the generated chromosome is shorter in duration than the
longest duration chromosome in the population.
Both variations of the algorithm are able to handle both continuous
and discrete values for activity durations, lag times, rework durations
and costs. Therefore, both can perform continuous and discrete
optimization.
5. Computerization
Both variations of overlapping optimization algorithm encompass
extensive recursive calculations and numerous trial and errors that
require computer implementation. To maximize the benefit, Microsoft
Excel (Excel) and Microsoft Project (MSP) were coupled to implement
the algorithm. The GA module was generated in Excel as a template,
and MSP was used to perform network calculations. This approach
greatly facilitated the implementation process, as network CPM calcula-
tions and other computations such as resource leveling are all included
as built-in features in MSP and are not required to be programmed inde-
pendently. Therefore, practitioners are provided with an automated tool
that integrates the current overlapping optimization algorithm into the
powerful features of their familiar software. Fig. 11 presents the flow of
information between Excel, MSP and the user.
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95 91

Fig. 11. Flow of information between MS Excel and MSP.

For several reasons Excel was selected as the GA algorithm platform: Microsoft Project applications as well as its flexibility to incorporate
first, it is a suitable platform for applying a wide variety of variables changes and perform sensitivity analyses.
and conducting complex and iterative computations; second, it gives In this section, the computer implementation of the first variation of
flexibility to the researcher to easily change the variables and review overlapping optimization algorithm is described. The second variation is
and analyze the results; third, it is quite compatible with MSP and a quite similar to the first. The required input data according to Fig. 11
flow of information can easily take place between them. is entered into the Excel template. From this, activity durations and
Visual Basic for Applications (VBA) has been used for automating the activity dependencies are primary network variables and are exported
calculations and coding and programming the simulation model. The to MSP to build the network. Minimum allowable lag times and over-
advantage of VBA is its compatibility with both Microsoft Excel and lapping rework functions are complementary network variables also
exported to MSP from Excel. Overlapping cost functions, project daily
benefits for early completion and project daily losses for late completion
Table 3
Results of the experiment. are cost variables and will be used by Excel to calculate the total costs/
benefits of overlapping. Project target duration is a constraint defined
No. Population Max. benefit Duration Calculation Min. benefit
by the user. The GA parameters are genetic algorithm variables and
($) (days) time (sec.) ($)
their variations can change the efficiency of the calculations.
1 Initial 4700 93 22 −5600 During optimization process, activity durations, activity dependen-
2 1st CO–MU 5400 93 31 −200
3 2nd CO–MU 5400 99 28 850
cies and suggested lag times are sent from Excel to MSP, which performs
4 3rd CO–MU 5400 99 29 1850 network calculations. Project duration and new lag times are then
5 4th CO–MU 5650 94 28 2650 returned to Excel. If project duration (T) is more than project target
6 5th CO–MU 6500 95 25 3050 duration (Tt), then the chromosome does not meet the time constraint
7 6th CO–MU 6500 95 27 3400
and the above steps should be repeated to generate a new chromosome.
8 7th CO–MU 6500 95 25 3750
9 8th CO–MU 6500 95 23 3850 However, if project duration (T) is less than project target duration (Tt),
10 9th CO–MU 6500 95 24 4350 the Excel template uses new lags to form a modified chromosome and
11 10th CO–MU 7700 93 24 4400 recalculate the total cost. This process is repeated until a desired num-
12 11th CO–MU 7700 93 25 4600 ber of first generation solutions (perhaps 100) is produced.
13 12th CO–MU 7700 93 22 4650
14 13th CO–MU 7700 93 23 4750
In the next step, the initial population should evolve to generate
15 14th CO–MU 7700 93 24 5350 better solutions. For this purpose, Excel selects two chromosomes ran-
16 15th CO–MU 8700 93 23 5400 domly as parent chromosomes and exchanges their genetic information
17 16th CO–MU 8700 93 22 5900 to reproduce an offspring chromosome. Once in a while Excel performs
18 17th CO–MU 8700 93 20 6350
mutation as well, according to a mutation to crossover ratio. The off-
19 18th CO–MU 8700 93 19 6350
20 19th CO–MU 8700 93 19 6500 spring chromosome is exported to MSP to perform network calculations
21 20th CO–MU 8700 93 19 6500 and determine project duration, modify lag times, and return the infor-
22 30th CO–MU 8700 93 206 6600 mation to Excel. Again, if project duration (T) is less than project target
23 40th CO–MU 8700 93 198 6700 duration (Tt), the Excel template uses the new lags to form a modified
24 50th CO–MU 8700 93 200 6700
offspring chromosome and recalculate the total cost. If the total cost is
92 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95

less than the total cost of a chromosome in the population with the
highest total cost, the old chromosome is replaced by the new chromo-
some. The above process—crossover-mutation and cost recalculation in
Excel and network calculations in MSP—is repeated until a satisfactory
number of offspring chromosomes are reproduced (perhaps 1000).
The final population is a collection of low-cost chromosomes. The
chromosome with the lowest total cost is the best solution. However,
other chromosomes in the vicinity are alternative solutions depending
on project conditions.

5.1. Experiment with the computer tool

In this section, the results of optimizing the sample case study are
presented. For the experiment, a personal computer (PC) with the
following attributes was used:
– Manufacturer: DELL
– Model: Studio XPS 9100
– Processor: Intel(R) Core(TM) i7 CPU 960 @ 3.2 GHz
– Installed memory (RAM): 12.0 GB
– System type: 64-bit Operating System
– Windows: Windows 7 Professional.
The mutation rate, crossover rate and the population size are the key
parameters of a GA. For obtaining good performance, these parameters
have to be “tuned” based on results obtained. Typically, a very small
mutation rate may lead to stagnation in local optima. A too-high mutation
rate may lead to loss of good solutions and an essentially random search.
No general theory exists to determine good rates. In this case study,
several trials were performed and the following GA parameters were
found to be suitable:
GA parameters:
Fig. 12. The best solution.
chromosome in the populations, on the other hand, was improved from
$5400 in the 15th generation to $6500 benefit in the 50th generation.
Therefore, even though the best solution remains unchanged, later gener-
ations encompass better solutions than earlier generations.
Fig. 13 shows some of the results achieved during the experiment. To
keep the figure readable, only the results of the initial population and
the 1st, 10th and 20th crossover–mutation populations are shown.
As shown in the figure, costly solutions happen more with longer
project durations such as 101 or 102 days. However, beneficial solutions
happen more with shorter project durations such as 93 or 94 days.
Therefore, for the current case maximum benefits happen at lowest pro-
ject durations and the optimization process improves the results from
costly long durations to beneficial short durations.
The total optimization calculation time to reach the best solution,
which was first obtained in the 15th population, is the sum of previous
calculation times of the previous populations. Therefore, the total opti-
mization calculation time is 403 s, less than 7 min.

– Initial population size: 100 6. Validation

– Crossover–mutation population size: 100
– Crossover rate: 98% To investigate the applicability of the algorithm and computer tool in
– Mutation rate: 2% the real world, a real project schedule was examined. A reputable engi-
– Termination criteria: Termination after 100, 200, 300, …, 2000, 3000, neering and construction company in Calgary, Canada, supported the
4000 and 5000 runs. research by assigning seven experienced individuals from the piping,
structure, planning and scheduling departments to the research.
Unfortunately, a major research limitation meant that the computer
The results of the optimization are presented in Table 3. tool could not be examined on a full-sized project with hundreds of
The maximum benefit, the project duration associated with the activities. Generating “overlapping functions” is difficult and time-
maximum benefit, and the optimization calculation time are the most consuming, particularly if there is no historical data available for the
important results. The minimum benefit is also presented for further overlaps, so that expert judgment is the only available source. Due to
analysis (project duration associated with the minimum benefit is not this limitation, only a section of a real schedule for an oil-sand project
shown). In any population (or generation), the best chromosome repre- could be used for validation.
sents the maximum benefit, while the worst chromosome represents
the minimum benefit.
According to Table 3, an initial population with 100 chromosomes
was generated. The processing time was 22 s. The best chromosome
resulted in a project maximum benefit of $4700 and a project duration
of 93 days. The worst chromosome in the population generated $5600
cost (i.e. −$5600 benefit). The initial population was used to generate
100 new chromosomes (1st CO–MU population) by means of crossover
and mutation (CO–MU). The calculation time was 31 s. In the new gen-
eration, the best chromosome had a maximum benefit of $5400 and the
worst chromosome generated $200 cost (i.e. −$200 benefit). Therefore,
both the best and worst solution improved significantly compared to
the initial population solutions.
By generating more chromosomes, better solutions were obtained. In
the 15th generation, after generating 1500 chromosomes through cross-
over and mutation, a chromosome emerged that had a maximum benefit
of $8700 and 93 days project duration. This chromosome was the best so-
lution achievable: even in the 50th generation no better chromosome
could be generated. The best solution is shown in Fig. 12. The worst Fig. 13. The initial population and the 1st, 10th and 20th CO–MU population.
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95 93

Table 4
List of activities to be overlapped.

Predecessor activity Original duration Successor activity Original duration Rework functions

Plot plan—Re IFDa 38 Module key plan—Re IFD 40 R = 0.05L

Plot plan IFCb 20 Equipment location plan—IFC 36 R = 0 for 0 b L b 6
R = 0.25L for 6 b L
60% model review tag resolution 20 Pipe rack module ISOs 148 R = 0.2L
60% model review tag resolution 20 Process module ISOs 48 R = 0.2L
a
IFD = issued for design.
b
IFC = issued for construction.

The section used had 40 activities including the piping and structural
design. Two full sessions were spent educating the experts what was
meant by “an overlapping rework function”, and four further sessions
were needed to review all 40 activities, identify those pairs that could
be overlapped, and develop their overlapping rework functions. With
the design of the project just finished and construction under progress,
the company was not able to provide more extensive support to the
research. Many other researchers have reported similar difficulties in
their research, and developing large numbers of overlapping rework
and cost functions remains a hotly-discussed topic in the field of activity
overlapping and concurrent engineering [7,13,14].
For better understanding, a part of the project schedule which
includes the above activities and their dependencies with each other
is schematically shown in Fig. 14.
6.2. Cost functions
Cost functions were generated by obtaining the direct and indirect
costs of performing the successor activities from the company. The
cost values (and functions) are reflected in Table 5.
The company also provided the project parameters as follows:

6.1. Rework functions
The project schedule was reviewed and four pairs of activities which
could be overlapped were identified. Other activities were either not
able to be overlapped or were so short in duration (less than 3 days)
that their overlapping could not generate a tangible timesaving. The
equivalent rework durations as a function of overlapping duration
were also determined. The experts were comfortable and confident
with generating rough estimates about the most likely “equivalent
rework”. These four pairs of activities and their associated rework func-
tions are listed in Table 4.
– Daily benefit for each day of early completion (Bef) = $1000 per day
– Daily loss for each day of late completion (Clf) = $1000 per day
– Project targeted completion date (Tt) = 730 days.
With the rework functions, cost functions, and other project para-
meters available, the optimization was performed. The results are
shown in Table 6.
According to Table 6, an initial population with 100 chromosomes
was generated. The processing time was 57 s. The best chromosome
resulted in a project maximum benefit of $9600 and a project duration
of 713 days. The worst chromosome in the population generated

Fig. 14. Activities to be overlapped and their dependencies with each other.
94 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95

Table 5 determining the optimal overlapping in the design phase of construc-

Cost functions. tion projects. For this purpose, an overlapping optimization algorithm
Successor activity Cost ($/day) Cost function which works based on the principles of genetic algorithms was devel-
Module key plan—Re IFD 1000 C = 1000R
oped. In this paper, different elements of the algorithm were explained
Equipment location plan—IFC 1800 C = 1800R in detail and the implementation was shown on an illustrative sample
Pipe rack module ISOs 2500 C = 2500R network. Both main variations of the algorithm—minimizing project du-
Process module ISOs 2500 C = 2500R ration within a predefined cost and minimizing project cost within a
predefined time frame—were addressed. In addition, computerization
of the algorithm was described and the results of an experiment with
$18,000 cost (−$18,000 benefit). In the table, the 1st, 2nd, 3rd, 4th, the computer tool were shown. Finally, the algorithm was validated
10th, 20th, 30th, 40th and 50th CO–MU populations are shown as well. using a real world project schedule.
In the 1st generation, after generating 100 chromosomes through The algorithm (and the computer tool) has abilities that differentiate
crossover and mutation, a chromosome emerged that had a maximum it from the other algorithms and frameworks available in the literature:
benefit of $17,000 and 711 days project duration. This chromosome It can optimize multi-path networks. This ability goes back to the fact
was the best solution achievable as even in the 50th generation no that two special situations, effect of multi-predecessor overlaps on re-
better chromosome could be generated. Therefore, the program found work and effect of cascades of overlaps on rework, have been addressed
the best solution very quickly and within 106 s (57 + 39), after gener- in the algorithm. During overlapping optimization process, critical paths
ating the initial population and the first crossover–mutation generation. may change and new critical paths may emerge. The overlapping opti-
The final results were shown to the experts in the focus group who mization algorithm has the ability to follow the critical path(s). In addi-
had helped generate the rework and cost functions. They admitted tion, a feature exists in the overlapping optimization algorithm that let it
that the results were practical and that they could work with the apply overlaps only on critical activities and relax noncritical overlaps
resulting fast-tracked schedule. through a recursive process. The overlapping optimization algorithm
To further examine the robustness of the final results, an experiment can also handle different types of activity dependency, i.e. finish-to-
was arranged with two senior schedulers (with over 15 years of work start, start-to-start and finish-to-finish. The necessary activity duration
experience) from another company. The project schedule and the adjustments and rework duration calculations related to each type of
required information such as maximum possible overlaps and overlap- activity dependency have been predicted in the algorithm. Despite its
ping costs and reworks were given to the schedulers, and they were novelty, practitioners, (project planners, schedulers, engineers, and
asked to generate the maximum timesaving with the minimum costs managers) can use the overlapping optimization computer tool and
through overlapping. They had one hour for this exercise and they apply overlapping on activities in project schedules. Specifically, MS
used Microsoft Project as the scheduling software. The best results Excel and MS Project, which were coupled for the tool, are familiar to
each of them could obtain were as follows: practitioners.
A comprehensive and robust schedule reduction strategy is one that
One scheduler could reduce the duration to 714 days with $12,200 includes both overlapping and crashing. Crashing, sometimes called
benefit. acceleration, refers to reducing activity duration by adding more
The other could reduce the duration to 713 days with $10,700 resources to the activity; therefore, the cost increases. A very important
benefit. step forward to this research is to develop an algorithm and a computer
The best results the experienced schedulers could reach after an tool that can identify the best combined crashing-overlapping strategy
hour were worse, from both time and cost standpoints, than the for project schedule reduction.
best result obtained by the overlapping computer tool (711 days,
References
$17,000 benefit) in less than two minutes.
[1] R. Dehghan, J. Ruwanpura, Model of trade-off between overlapping and rework of
In addition, both schedulers made a few mistakes in calculating the design activities, ASCE J. Constr. Eng. Manag. 140 (2) (2014)http://dx.doi.org/10.
project duration; manually calculating rework durations was particularly 1061/(ASCE)CO.1943-7862.0000786.
difficult. This shows another superiority of the overlapping computer [2] V. Krishnan, Managing the simultaneous execution of coupled phases in concurrent
product development, IEEE Trans. Eng. Manag. 43 (2) (1996) 210–217, http://dx.
tool over classic scheduling.
doi.org/10.1109/17.509986.
[3] C.H. Loch, C. Terwiesch, Communication and uncertainty in concurrent engineering,
7. Conclusion Manag. Sci. 44 (8) (1998) 1032–1048, http://dx.doi.org/10.1287/mnsc.44.8.1032.
[4] S. Nicoletti, F. Nicolo, A concurrent engineering decision model: management of the
project activities information flows, Int. J. Prod. Econ. 54 (1998) 115–127, http://dx.
This paper described the results of a research into activity overlap- doi.org/10.1016/S0925-5273(97)00108-4.
ping and project acceleration. The main objective of the research [5] B. Prasad, Concurrent engineering fundamentals: integrated product and process
was to develop a systematic and practical approach to assessing and organization, Prentice Hall, Upper Saddle River, N.J., 1996
[6] C. Terwiesch, C.H. Loch, A. De Meyer, Exchanging preliminary information in concurrent
engineering: alternative coordination strategies, Organ. Sci. 13 (4) (2002) 402–419,
http://dx.doi.org/10.1287/orsc. 13.4.402.2948.
Table 6 [7] T.A. Roemer, R. Ahmadi, R.H. Wang, Time–cost trade-offs in overlapped product
Optimization results of the real project. development, Oper. Res. 48 (6) (2000) 858–865, http://dx.doi.org/10.1287/opre.
48.6.858.12396.
N. Population Max. benefit Duration Calculation time Min. benefit [8] T.A. Roemer, R. Ahmadi, Concurrent crashing and overlapping in product develop-
($) (days) (sec.) ($) ment, Oper. Res. 52 (4) (2004) 606–622, http://dx.doi.org/10.1287/opre.1040.0125.
[9] S.D. Eppinger, Three concurrent engineering problems in product development
1 Initial 9600 713 57 −18,000 seminar, Massachusetts Institute of Technology, Sloan School of Management, C,
2 1st CO–MU 17,000 711 39 −5500 Cambridge, Mass, 1997.
3 2nd CO–MU 17,000 711 40 −500 [10] F. Pena-Mora, M. Li, Dynamic planning and control methodology for design/build
4 3rd CO–MU 17,000 711 39 2300 fast track construction projects, ASCE J. Constr. Eng. Manag. 127 (2001) N1, http://
5 4th CO–MU 17,000 711 40 4800 dx.doi.org/10.1061/(ASCE)0733-9364(2001)127:1(1).
6 10th CO–MU 17,000 711 368 8900 [11] S. Bogus, J. Diekmann, K. Molenaar, C. Harper, S. Patil, J. Lee, Simulation of over-
7 20th CO–MU 17,000 711 355 10,200 lapping design activities in concurrent engineering, ASCE J. Constr. Eng. Manag.
8 30th CO–MU 17,000 711 358 10,200 137 (11) (2011) 950–957.
9 40th CO–MU 17,000 711 383 10,500 [12] S.M. Bogus, K.R. Molenaar, J.E. Diekmann, Concurrent engineering approach to
reducing design delivery time, ASCE J. Constr. Eng. Manag. 131 (11) (2005)
10 50th CO–MU 17,000 711 379 10,700
1179–1185, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000363.
 R. Dehghan et al. / Automation in Construction 59 (2015) 81–95
[13] S.M. Bogus, K.R. Molenaar, J.E. Diekmann, Strategies for overlapping dependent
design activities, Construction Management and Economics, August 2006 (24)
(2006) 829–837, http://dx.doi.org/10.1080/01446190600658529.
[14] N.A. Blacud, S.M. Bogus, J.E. Diekmann, K.R. Molenaar, Sensitivity of construction ac-
tivities under design uncertainty, ASCE J. Constr. Eng. Manag. 135 (3) (2009)
199–206, http://dx.doi.org/10.1061/(ASCE)0733-9364(2009)135:3(199).
[15] A.Y. Ha, E.L. Porteus, Optimal timing of reviews in concurrent design for manufactur-
ability, Manag. Sci. 41 (9) (1995) 1431–1447, http://dx.doi.org/10.1287/mnsc.41.9.
1431.
[16] V. Krishnan, S.D. Eppinger, D.E. Whitney, Accelerating product development by the
exchange of preliminary product design information, ASME J. Mech. Des. 117 (1995)
491–498, http://dx.doi.org/10.1115/1.2826709.
[17] V. Krishnan, S.D. Eppinger, D.E. Whitney, A model based framework to overlap product
development activities, Manag. Sci. 43 (4) (1997) 437–451, http://dx.doi.org/10.1287/
mnsc.43.4.437.
[18] J.E.V. Gerk, R.Y. Qassim, Project acceleration via activity crashing, overlapping, and
substitution, IEEE Trans. Eng. Manag. 55 (4) (2008) 590–601, http://dx.doi.org/10.
1109/TEM.2008.927786.
[19] F. Berthaut, L. Greze, R. Pellerin, N. Perrier, A. Hajji, Optimal resource-constraint
project scheduling with overlapping modes, CIRRELT-2011-092011.
[20] F. Berthaut, R. Pellerin, N. Perrier, A. Hajji, Time–cost trade-offs in resource-
constraint project scheduling problems with overlapping modes, CIRRELT-2011-
102011.
[21] R. Dehghan, A new model, algorithm and computer tool to optimize overlapping of
design activities in construction projectsPh.D. Dissertation Department of Civil Engi-
neering, Schulich School of Engineering, University of Calgary, AB, Canada, 2011.
[22] A.K. Cho, M. Hastak, Time and cost-optimized decision support model for fast-track
projects, ASCE J. Construction Engineering and Management. 139 (1) (2013)
90–101, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000570.
[23] Y. Khoueiry, I. Srour, A. Yassine, An optimization-based model for maximizing
the benefits of fast-track construction activities, J. Oper. Res. Soc. 64 (8) (2013)
1137–1146, http://dx.doi.org/10.1057/jors.2013.30.
[24] I.M. Srour, M.A.U. Abdul-Malak, A.A. Yassine, M. Ramadan, A methodology for
scheduling overlapped design activities based on dependency information, Autom.
Constr. 29 (2013) 1–11, http://dx.doi.org/10.1016/j.autcon.2012.08.001.
95
[25] K. Hazini, R. Dehghan, J.Y. Ruwanpura, A heuristic method to determine optimum
degree of activity accelerating and overlapping in schedule compression, Can.
J. Civ. Eng. 40 (4) (2013) 382–391, http://dx.doi.org/10.1139/cjce-2012-0380.
[26] M.A. Hossain, D.K.H. Chua, Overlapping design and construction activities and an
optimization approach to minimize rework, Int. J. Proj. Manag. 32 (6)
(2013)http://dx.doi.org/10.1016/j.ijproman.2013.10.019.
[27] R. Dehghan, J.Y. Ruwanpura, The mechanism of design activity overlapping in
construction projects and the time–cost trade-off function, The 12th East Asia-Pacific
Conference on Structuralq Engineering and Construction (EASEC-12), January 26–28,
2011, Hong Kong, 2011http://dx.doi.org/10.1016/j.proeng.2011.07.246.
[28] R. Dehghan, K. Hazini, J.Y. Ruwanpura, Optimum activity overlapping using genetic
algorithms, The Sixth International Structural Engineering and Construction Confer-
ence (ISEC-6), June 21–26, 2011, Zürich, Switzerland, 2011http://dx.doi.org/10.
3850/978-981-08-7920-4_S1-CP19-cd.
[29] R. Dehghan, J.Y. Ruwanpura, K. Hazini, A new model to minimize project costs
through optimum activity overlapping, Proceedings of the Pacific Association of
Quantity Surveyors (PAQS) 15th Annual Congress, July 23–26, 2011, Colombo, Sri
Lanka, 2011.
[30] T. Hegazy, Optimization of construction time–cost trade-off analysis using genetic
algorithms, Can. J. Civ. Eng. 26 (1999) 685–697, http://dx.doi.org/10.1139/cjce-26-
6-685.
[31] D.E. Goldberg, Genetic algorithms in search, optimization and machine learning,
Addison-Wesley Publishing Co., Reading, Mass, 1989.
[32] J. Zhou, P.E.D. Love, X. Wang, K.L. Teo, Z. Irani, A review of methods and algorithms
for optimizing construction scheduling, J. Oper. Res. Soc. 2013 (64) (2013)
1091–1105, http://dx.doi.org/10.1057/jors.2012.174.
[33] D.X.M. Zheng, S.T. Ng, M.M. Kumaraswamy, Applying a genetic algorithm-based
multiobjective approach for time–cost optimization, ASCE Journal of Construction
Engineering and Management 130 (2) (2004) 168–176, http://dx.doi.org/10.1061/
(ASCE)0733-9364(2004)130:2(168).
[34] C.W. Feng, L. Liu, S.A. Burns, Using genetic algorithms to solve construction time–
cost trade-off problems, ASCE Journal of Construction Engineering and Management
123 (3) (1997) 184–189, http://dx.doi.org/10.1061/(ASCE)0887-3801(1997)11:
3(184).