The Journal of Supercomputing

https://doi.org/10.1007/s11227-019-03055-6

Resource levelling problem in construction projects

under neutrosophic environment

Mohamed Abdel‑Basset1 · Mumtaz Ali2 · Asma Atef1

© Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract
Planning and managing resources is one of the most important topics in project
management science. Resource leveling is used for improving work efficiency and
minimizing cost throughout the life of the project. Fuzzy resource leveling mod-
els assume only truth-membership functions to deal uncertainties conditions sur-
rounded by the projects and their activities duration. In this paper, we consider the
objective function of scheduling problem is to minimize the costs of daily resource
fluctuations using the precedence relationships during the project completion time.
We design a resource leveling model based on neutrosophic set to overcome the
ambiguity caused by the real-world problems. In this model, trapezoidal neutro-
sophic numbers are used to estimate the activities durations. The crisp model for
activities time is obtained by applying score and accuracy functions. A numerical
example is developed to illustrate the validation of the proposed model in this study.

Keywords  Project scheduling · Resource aggregation · Resource leveling ·

Neutrosophic theory · Trapezoidal neutrosophic number

* Mohamed Abdel‑Basset
analyst_mohamed@zu.edu.eg
Mumtaz Ali
mumtaz.ali@deakin.edu.au
Asma Atef
asmaatef@fci.zu.edu.eg
1
Faculty of Computers and Informatics, Zagazig University, Sharqiyah, 44519, Egypt
2
Deakin‑SWU Joint Research Centre on Big Data, School of Information Technology, Deakin
University, Burwood, VIC 3125, Australia

13
Vol.:(0123456789)
 M. Abdel‑Basset et al.

1 Introduction and related works

A project, as abroad definition, is an undertaking or a venture to accomplish

some goal or objective. A project is considered as a set of interrelated jobs whose
accomplishments lead to the completion of the project. Jobs or activities consume
time and resources and are governed by precedence relations. Activities are actu-
ally responsible for project completion. Project scheduling is the process of allo-
cate the scare resources to activities to make the project be completed to attain
specific objectives such as minimize the completion time of the project, mini-
mize the cost of the project execution and/or maximize the quality of the project.
Resources in general are classified into two categories of resources; these cat-
egories are the renewable resources and non-renewable resources. The distinction
between the two main categories of resources is shown in Fig. 1. Non-renewable
resources are those resources which are often referred to as consumable resources
as well. Non- renewable resources such as money, energy, fuel, raw materials are
best handled by time–cost trade-offs. That means the objective is to minimize the
total consumption of non-renewable resources over the feasible range of project
durations. That means whenever we are dealing with consumable resources or
non-renewable resources those which get consumed like cost we resort to proce-
dures like time–cost trade-offs which are used for trying to find out the minimum
cost schedule or trying to find out the best trade-off between the project dura-
tion and the project time. However, when we talk about renewable resources in
projects, which is our major concern in this paper, renewable resources like man
power, machines and various other capital equipment which are necessary in the
implementation of a project and some other kinds of special considerations are
required to plan for various types of resources in a project. The first of these con-
siderations is the construction of the resource profiles. Resource profile construc-
tion can start with a project network and then impose on the project network the
activity times, and it is very easy to determine a project schedule by doing a sim-
ple forward pass and a backward pass which can identify the early start and the

RENEWABLE NON- RENEWABLE

RESOURCES RESOURCES

Examples:Manpower, Examples:Money, Fuel

Power, Machines, and Fuel ,Energy, and Raw Materials
Flow

Handeled by
Resource Aggregation.
Resource Levelling.
Resource Allocation.
Objective: Minimize the max
peak or the project durations
Handeled by
Time-Cost Trade offs.
Objetive: Minimize the
Total Consumption over
Project durations

Fig. 1  Categories of resources types

13
Resource levelling problem in construction projects under…

late start schedules. That means for each activity in the project we can identify
the time that particular activity is going to be taking place. These activities might
require some types of resources such as man power, special type of machines or
equipment. If we want to find out the manner in which we are going to be using
that equipment, we will have to refer to these basic schedules.
From the basic schedule of activities, we can focus on this the activity resource
data. Depending upon which activities are active at what particular point of time, we
can determine the resource usage profile as shown in Fig. 2. This particular graph is
called a resource usage profile. It shows the variation in the resources which is going
to be needed to accomplish the schedule of the project.
This information is very important to the planner, the project manager, because
it would help him to identify exactly what is going to be, the resource require-
ment for the particular project and exactly how much resource is required during
what particular periods of time. The project manager can make his provisions
for hiring and firing appropriately. Renewable resources are basically handled
through three different mechanisms. The simplest way of handling these resources
is through what we call resource aggregation. Aggregation, simply means as the
name suggests, is the process of working out the resource usage profile from the

Project Network Activity Times

Resource
Usage
Schedule

Activity Resource Data

8
Resource Usage Profile

T1 T2 T3 TN
Time

Fig. 2  Resource profile

13
 M. Abdel‑Basset et al.

schedule. Resource aggregation is actually simply trying to find out the resource
requirements for a particular schedule. It does not in any way try to influence
those resource requirements. It is simply like taking a snapshot of a particular
schedule as shown in Fig. 2. On the other hand, resource leveling is more detailed
in the sense that it looks at the resource profile and we might not find the resource
profile very satisfactory. It might want to change the schedule to get a better more
acceptable resource profile as shown in Fig. 3, by minimizing the max peak and
also minimize the fluctuates as much as possible while make the completion time
of the project constant. Essentially, that is the process of resource leveling and
limited resource allocation is the process of trying to conform to resource avail-
abilities and trying to minimize the project duration as shown in Fig. 4. There are
procedures available for resource leveling. There are both analytical procedures
as well as heuristic procedures but the difficulty with the analytical procedures is
that they are capable of handling only a small number of jobs or very small pro-
jects and for real-life projects it is not possible to do resource leveling for a large
project which involves hundreds of activities. For this reason, generally, resort
is made to heuristic procedures. Although heuristic procedures do not guarantee
optimal solutions nevertheless, they are practical procedures which can be used
for this particular situation.
There are many studies [1–11] discussed the traditional procedures for leveling
resources. Many heuristics and meta-heuristics studies [12–23] also discussed the
resource leveling problem. Many fuzzy models [24–27] are also study the prob-
lem of resource leveling under uncertain conditions. In this paper, we will use
neutrosophic heuristic procedure for resource leveling.

R (T) Original
Resource Profile

Obtained
Resource Profile

T Time

Fig. 3  Resource leveling

13
Resource levelling problem in construction projects under…

Original
R (T) Resource Profile

Best Leveled
Resource Profile

Resource
Allocation Profile

T TN TI Time

Fig. 4  Resource allocation

2 Preliminaries

Neutrosophic theory was developed by Florentin Smarandache in 1998. We pre-

sent definitions concerning neutrosophic sets, trapezoidal neutrosophic numbers
and single valued neutrosophic sets.

Definition 1 [31] Let be a space of points and  x ∈ X . A neutrosophic set A in X
is defined by a truth-membership function TA(x), an indeterminacy-membership
function IA(x) and a falsity-membership function FA(x), TA(x), IA(x) and FA(x) are
real standard or real nonstandard subsets of [− 0, 1+]. That is TA(x):→ [− 0, 1 +] IA
(x):→[− 0 1 +] and FA(x):→[− 0, 1 +]. There is no restriction on the sum of TA(x), IA
(x) and FA(x), so 0 −≤ sup (x) + sup  ≤ 3 +.

Definition 2 [30, 31, 32] Let X be a universe of discourse. A single valued neutro-
sophic set A over X is an object taking the form A = {〈x, TA(x), IA(x), FA(x),〉:x ∈ X},
where TA(x):X→ [0,  1], IA(x):X→ [0,  1] and FA(x):X→[0,  1] with 0 ≤ TA(x) + IA
(x) + FA(x) ≤ 3 for all x  X. The intervals TA(x), IA(x) and FA(x) represent the truth-
membership degree, the indeterminacy-membership degree and the falsity-mem-
bership degree of x to A, respectively. For convenience, a single valued neu-
trosophic (SVN) number is represented by A = (a, b, c), where a, b, c  [0, 1] and
a+b+c ≤ 3.

13
 M. Abdel‑Basset et al.

Definition 3 [34] Suppose 𝛼ã  , 𝜃ã  , 𝛽ã ∈ [0,1]  and a1 , a2 , a3 , a4 𝜖 R, where
a1 ≤ a2 ≤ a3 ≤ a4 . Then, a single valued trapezoidal neutrosophic number ã  = 〈(a1 ,
a2 , a3 , a4 ); 𝛼ã  , 𝜃ã  , 𝛽ã  〉 is a special neutrosophic set on the real line set R, whose
truth-membership, indeterminacy membership and falsity-membership functions are
defined as:
� �
⎧ 𝛼 x−a1 (a1 ≤ x ≤ a2 )
⎪ ã a2 −a1
⎪𝛼 (a2 ≤ x ≤ a3 )
Tã (x) = ⎨ ã � a −x � (1)
⎪ 𝛼ã a4 −a3
4
(a3 ≤ x ≤ a4 )
⎪0 otherwise
⎩

⎧ (a2 −x+𝜃ã (x−a1 )) (a1 ≤ x ≤ a2 )

⎪ (a2 −a1 )
⎪ 𝛼ã (a2 ≤ x ≤ a3 )
Iã (x) = ⎨ (x−a +𝜃 (a4−x)) , (2)
⎪
3 ã
(a3 ≤ x ≤ a4 )
⎪ (a4 −a3 )
⎩1 otherwise

⎧ (a2 −x+𝛽ã (x−a1 )) (a1 ≤ x ≤ a2 )

⎪ (a2 −a1 )
⎪ 𝛼ã (a2 ≤ x ≤ a3 )
Fã (x) = ⎨ (x−a +𝛽 (a4−x)) , (3)
⎪
3 ã
(a3 ≤ x ≤ a4 )
⎪ (a4 −a3 )
⎩1 otherwise

where 𝛼ã  , 𝜃ã and 𝛽ã typify the maximum truth-membership degree, the minimum
indeterminacy-membership degree and the minimum falsity-membership degree,
respectively. A single valued trapezoidal neutrosophic number ã  = 〈(a1 , a2 , a3 , a4 );
𝛼ã  , 𝜃ã  , 𝛽ã 〉 may express an ill-defined quantity of the range, which is approximately
equal to the interval [ a2 , a3].

3 Design of proposed plan

Resource aggregation, resource leveling and resource allocations are three ways for
handling the renewable resources. Resource aggregation is just construct resource
usage profile from the project schedule and the activities resource data. Resource
allocation aimed to minimize the project duration with fixed resource availabilities.
Resource leveling, the core of this research, aimed to smooth resource usage profile
by shifting the tasks relative to their available slack or floats without worsening the

13
Resource levelling problem in construction projects under…

duration of the project. There are different criteria affected leveling resources, such
as the peak level of the resources usage, as used in workload smoothing heuristic
developed by Levy, Thampson and Wiest [28]; another criteria that may affect the
resource leveling is the sum of squares of resources usage as used by Burgess and
Killebrew heuristic [29]. For more reality, we present the proposed Neutrosophic—
Burgess heuristic algorithm for the purpose of best resource leveling to minimize
the fluctuations in day to day of resources used during the project execution time in
uncertain situations as follows:

3.1 Neutrosophic—Burgess heuristic algorithm

Step 1 Estimate the activities duration times based on trapezoidal neutrosophic

number.
Step 2 Obtain crisp model of Neutrosophic duration times and resources required
of activities. The method of comparing any two single valued trapezoidal neutro-
sophic numbers is basically based on the score function and the
( accuracy ) function.
Let ã   (a single valued trapezoidal neutrosophic number) = a1 , b1 , c1 , 𝛼ã , 𝜃ã , 𝛽ã  ,
then
1 ( )
Score function, S(̃a) = [a1 + b1 + c1 + d1] × 2 + 𝛼ã − 𝜃ã − 𝛽ã (4)
16
And
1 ( )
Accuracy function, A(̃a) = [a1 + b1 + c1 + d1] × 2 + 𝛼ã − 𝜃ã − 𝛽ã (5)
16
The crisp model uses the score and accuracy functions of ã in addition to the
truth membership, indeterminacy membership and falsity-membership functions
using Eqs. (4), (5) and (1), (2), (3), respectively.
Step 3 Carry out the forward pass calculations, and then list the activities of the
project in order of precedence. Add to this list the durations, resource required, the
early start time, the early finish time and slack or float values for each activity.
Step 4 Start scheduling the last activity, the one at the bottom of the bar chart
diagram, by shifting it to the right based on different available floats, and selects the
schedule which gives the minimum total sum of squares of resource requirements of
available for each time unit. Schedule the activities as late as possible based on the
maximum total float when more than one schedule gives the sane minimum sum of
squares of resource requirements.
Step 5 Make the last activity fixed, repeat step 4 on the next last activity in the bar
chart, based on different slacks may have been made available to it by rescheduling
in step 4.
Step 6 Complete the first rescheduling cycles by continuing step 5 until consider
the first activity in the list.
Step 7 Repeat step 4 through 6 to make additional rescheduling cycles until no
additional reduction in the total resource requirements sum of squares is possible.
Step 8 In the previous step, select the best obtained schedule.

13
 M. Abdel‑Basset et al.

Table 1  Project data
Activity Activity time (days) as trapezoidal neutrosophic Predecessors

A About 3 days (1,4,5,6; 0.6,0.1,0.5) –

B About 4 days (1,3,5,7; 0.9,0.4,0.5) –
C About 2 days (1,2,3,4; 0.4,0.3,0.5) –
D About 5 days (1,5,6,8; 0.8,0.2,0.2) C
E About 10 days (4,9,12,15; 0.8,0.7,0.1) D
F About 7 days (3,8,10,11; 0.9,0.6,0.3) A
G About 8 days (4,7,9,12; 0.7,0.5,0.2) B
H About 7 days (2,4,5,9; 0.7,0.1,0.2) B
I About 9 days (3,6,9,14; 0.6,0.2,0.4) F, G
J About 5 days (5,16,18,21; 0.5,0.1,0.8) E, H

Table 2  Crisp values of Act Di Ri Precedence ES LS EF LF TF

activities duration, resource
required and the early start A 2 1 – 0 2 5 7 5
schedule of the project
B 2 4 – 0 2 4 6 4
C 1 2 – 0 1 0 1 0
D 3 4 A. 1 4 1 4 0
E 5 1 B. 4 9 4 9 0
F 4 3 C. 2 6 7 11 5
G 4 2 C. 2 6 7 11 5
H 3 3 D. 2 5 6 9 4
I 4 3 D. 6 10 11 15 5
J 6 2 E, F, I. 9 15 9 15 0

Step 9 Make final adjustments to the schedule chosen in step 8.

4 Illustrative example

Consider the time of project activities are estimated as trapezoidal neutrosophic

numbers and listed in Table 1 as follows:

4.1 Implementation of the proposed method

We estimate the activities durations based on trapezoidal neutrosophic numbers,

and then obtain the crisp values of activities’ time by Eqs.  4 and 5. Then start
with the early start schedule of the project and list the project activities in order
of precedence and add to this order the obtained crisp durations (Di), the data of
resources required by each activity (Ri), the early start (ES), late start (LS), early
finish (EF), late finish (LF) and slack or total float (TF) values for each activity

13
Resource levelling problem in construction projects under…

as given in Table 2, and then we make the initial early start scheduling resource
profile as shown in Fig. 6, that obtained from the bar chart in Fig. 5 with the list
of activities resource, and then calculate the total resource requirements sum of
squares for each time unit as given in Table 3.
Then starting with activity (I), and reschedule it at different times according to
the available total floats on it. Science activity (I) has 5 units of floats, activity (I)
will be rescheduled five times from activity (I) with one unit of float (I1) to (I5)
with five units of float (the max number of available floats on activity I), then cal-
culate the total summation of squares of resource requirements at different float
times and select the schedule that give the lowest total sum of squares of resource
requirements for each time unit. Calculations of resource requirements to the dif-
ferent floats on activity (I) are given in Table 4.
According to the Neutrosophic—Burgess algorithm, holding activity (I) fixed
with the early start schedule of (628) as a minimum sum of squares of resource
requirements. Taking activity (H) in the next steps of our algorithm and also
reschedule it due to the available floats on activity (H) from H1 to H4 to give the
lowest total summation of squares of resources requirements as given in Table 5.

A (1)
A

B (4)
B

C (2)
C

D (4)
D

E (1)
E

F (3)
F

G (2)
G

H (3)
H
I (3)
I
J (2)
J

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Fig. 5  Bar chart of the project in early start schedule

13


13
Table 3  Resource requirements for each time unit and total sum of square of resource requirements
Time per Resource required per day
t=1 R2
∑15
day (T)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day (10) Day (11) Day (12) Day (13) Day (14) Day (15)

Resource 7 9 12 12 9 6 4 4 4 5 2 2 2 2 2 628
Required
(R)
M. Abdel‑Basset et al.
 Table 4  Resource requirements of rescheduling activity (I) at different float times
Activity Resource requirements per day
t=1 R2
∑15
(I)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

I1 7 9 12 12 9 6 1 4 4 5 5 2 2 2 2 634
Resource levelling problem in construction projects under…

I2 7 9 12 12 9 6 1 1 4 5 5 5 2 2 2 640
I3 7 9 12 12 9 6 1 1 1 5 5 5 5 2 2 646
I4 7 9 12 12 9 6 1 1 1 2 5 5 5 5 2 646
I5 7 9 12 12 9 6 1 1 1 2 2 5 5 5 5 646

13


13
Table 5  Resource requirements of rescheduling activity (H) at different float times
Activity Resource requirements per day
t=1 R2
∑15
(H)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

H1 7 9 9 12 9 9 4 4 4 5 2 2 2 2 2 610
H2 7 9 9 9 9 9 7 4 4 5 2 2 2 2 2 580
H3 7 9 9 9 6 9 7 7 4 5 2 2 2 2 2 568
H4 7 9 9 9 6 6 7 7 7 5 2 2 2 2 2 556
M. Abdel‑Basset et al.
 Table 6  Resource requirements of rescheduling activity (H) at different float times
Activity Resource requirements per days
t=1 R2
∑15
(I)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

I1 7 9 9 9 6 6 4 7 7 5 5 2 2 2 2 544
Resource levelling problem in construction projects under…

I2 7 9 9 9 6 6 4 4 7 5 5 5 2 2 2 532
I3 7 9 9 9 6 6 4 4 4 5 5 5 5 2 2 520
I4 7 9 9 9 6 6 4 4 4 2 5 5 5 5 2 520
I5 7 9 9 9 6 6 4 4 4 2 2 5 5 5 5 520

13


13
Table 7  Resource requirements of rescheduling activity (G) at different float times
Activity Resource requirements per days
t=1 R2
∑15
(G)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

G1 7 9 7 9 6 6 6 4 4 2 2 5 5 5 5 508
G2 7 9 7 7 6 6 6 6 4 2 2 5 5 5 5 496
G3 7 9 7 7 4 6 6 6 6 2 2 5 5 5 5 496
G4 7 9 7 7 4 4 6 6 6 4 2 5 5 5 5 488
G5 7 9 7 7 4 4 4 6 6 4 4 5 5 5 5 480
M. Abdel‑Basset et al.
Resource levelling problem in construction projects under…

As our objective of minimizing the total summation of squares of resources

requirement, so we select scheduling activity (H) as late as possible with total
∑
float equal to 4 with 15 t=1
R2 = 556. Then holding activity (H) at late as possi-
ble and reschedule activity (I) for obtaining the lowest possible summation of
resource required as given in Table 6.
According to our algorithm, schedule activities based on max available floats,
as late as possible in all preceding activities when more than on schedule give the
same minimum total of square. As it obvious that scheduling activity (I) with 3, 4
and five units of float give the same total summation of squares equal to 520, then
we select rescheduling (I) at float equal to five units. Then hold activity (I) with
5 units of float and activity (H) with 4 units of float, and then reschedule activity
(G), the next activity relative to our algorithm, to different float times as given in
Table 7.
Select scheduling activity (G) as late as possible with total float equal to five, where
that yielding the minimum total summation of resources requirements equal to 480.
Then hold activity (G) with total float equal to 5 days, activity (H) with total float equal
to 4 days, then reschedule activity (I) again to minimize the total summation of resource
requirements, as given in Table 8.
Hold activity (G) fixed with total float equal to 5 days, activity (I) with total float
equal to 5 days, activity (H) with total float equal to 4 days, then reschedule activity (F)
again to minimize the total summation of resource requirements, as given in Table 9.
Hold activity (F) fixed with total float equal to 1 day, activity (G) with total float
equal to 5 days, activity (I) with total float equal to 5 days, activity (H) with total
float equal to 4 days, then reschedule activity (B) again to minimize the total sum-
mation of resource requirements, as given in Table 10.
Hold activity (B) fixed with total float equal to 4 days, activity (F) with total float
equal to 1 day, activity (G) with total float equal to 5 days, activity (I) with total float
equal to 5 days, activity (H) with total float equal to 4 days, then reschedule activ-
ity (A) again to minimize the total summation of resource requirements, as given in
Table 11.
Hold activity (A) fixed at early start scheduling, (B) fixed with total float equal to
4 days, activity (F) with total float equal to 1 day, activity (G) with total float equal
to 5  days, activity (I) with total float equal to 5  days, activity (H) with total float
equal to 4 days, then reschedule activity (F) again to minimize the total summation
of resource requirements, as given in Table 12.
The minimum total summation of resource requirements in this example by using
the proposed technique is equal to 480, this is obtained from hold all activities fixed
at different scheduling floats times as follows, hold activity (A) fixed at early start
scheduling, (B) fixed with total float equal to 4  days, activity (F) with total float
equal to 1 day, activity (G) with total float equal to 5 days, activity (H) with total
float equal to 4 days, activity (I) with total float equal to 5 days, as given in Table 13.
The final resource profile after applying the proposed procedure is shown in Fig. 7.
The results show that the total summation of resource requirements before and
after applying the proposed procedures, that is reduced from 628 to 480 after apply-
ing the proposed procedure. Also as it obvious from the initial resource profile
(Fig. 6) in early start scheduling and final resource profile (Fig. 7) after applying the

13


13
Table 8  Resource requirements of rescheduling activity (I) at different float times
Activity Resource requirements per days
t=1 R2
∑15
(I)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

I1 7 9 7 7 4 4 4 9 9 7 7 2 2 2 2 552
I2 7 9 7 7 4 4 4 6 9 7 7 5 2 2 2 528
I3 7 9 7 7 4 4 4 6 6 7 7 5 5 2 2 504
I4 7 9 7 7 4 4 4 6 6 4 7 5 5 5 2 492
I5 7 9 7 7 4 4 4 6 6 4 4 5 5 5 5 480
M. Abdel‑Basset et al.
 Table 9  Resource requirements of rescheduling activity (F) at different float times
Activity Resource requirements per days
t=1 R2
∑15
(F)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

F1 7 9 4 7 4 4 7 6 6 4 4 5 5 5 5 480
Resource levelling problem in construction projects under…

F2 7 9 4 4 4 4 7 9 6 4 4 5 5 5 5 492
F3 7 9 4 4 1 4 7 9 9 4 4 5 5 5 5 522
F4 7 9 4 4 1 1 7 9 9 7 4 5 5 5 5 540
F5 7 9 4 4 1 1 4 9 9 7 7 5 5 5 5 540

13


13
Table 10  Resource requirements of rescheduling activity (B) at different float times
Activity Resource required per days
t=1 R2
∑15
(B)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

B1 3 9 8 7 4 4 7 6 6 4 4 5 5 5 5 488
B2 3 5 8 11 4 4 7 6 6 4 4 5 5 5 5 543
B3 3 5 4 11 8 4 7 6 6 4 4 5 5 5 5 504
B4 3 5 4 7 8 8 7 6 6 4 4 5 5 5 5 480
M. Abdel‑Basset et al.
 Table 11  Resource requirements of rescheduling activity (A) at different float times
Activity Resource requirements per days
t=1 R2
∑15
(A)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

A1 2 5 5 7 8 8 7 6 6 4 4 5 5 5 5 484
Resource levelling problem in construction projects under…

A2 2 4 5 8 8 8 7 6 6 4 4 5 5 5 5 490
A3 2 4 4 8 9 8 7 6 6 4 4 5 5 5 5 498
A4 2 4 4 7 9 9 7 6 6 4 4 5 5 5 5 500
A5 2 4 4 7 8 9 8 6 6 4 4 5 5 5 5 498

13


13
Table 12  Resource requirements of rescheduling activity (F) at different float times
Activity Resource requirements per days
t=1 R2
∑15
(F)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day Day Day Day Day Day
(10) (11) (12) (13) (14) (15)

F1 3 5 4 7 8 8 7 6 6 4 4 5 5 5 5 480
F2 3 5 4 4 8 8 7 9 6 4 4 5 5 5 5 492
F3 3 5 4 4 5 8 7 9 9 4 4 5 5 5 5 498
F4 3 5 4 4 5 5 7 9 9 7 4 5 5 5 5 492
F5 3 5 4 4 5 5 4 9 9 7 7 5 5 5 5 492
M. Abdel‑Basset et al.
 Table 13  Resource requirements for each time unit and total sum of square of resource requirements after applying the proposed procedure
Time per Resource requirements per days
t=1 R2
∑15
day (T)
Day (1) Day (2) Day (3) Day (4) Day (5) Day (6) Day (7) Day (8) Day (9) Day (10) Day (11) Day (12) Day (13) Day (14) Day (15)
Resource levelling problem in construction projects under…

Resource 3 5 4 7 8 8 7 6 6 4 4 5 5 5 5 480
required
(R)

13
 M. Abdel‑Basset et al.

R (T)

12

11

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
T

Fig. 6  Initial resource profile of the project

proposed procedure, the max peak is reduced from 12 to 8, and the fluctuation in
day to day of resources used is also minimized as the purpose of best resource lev-
eling. Also the sum of squares of resources requirements is reduced from 628 to 480
as given in Tables 3, and 13, before and after applying the new proposed method.
This indicates the effectiveness of the proposed procedure for leveling resources in
the absence of accuracy and confirmation of the times of project activities.

5 Conclusion and future work

Neutrosophic set is the most comprehensive set, which includes both fuzzy set and
intuitionistic fuzzy set, as it considers the indeterminacy function in addition to
truth-membership and falsity membership, being suitable in analyzing real situa-
tions. Also, in real-life situations, accurate judgments are rarely since ambiguity and
uncertainty surround the decision-making process. Resource leveling is one of the
mechanisms used for handling the renewable resources. In this research, we proposed
Neutrosophic—Burgess heuristic algorithm for the purpose of best resource leveling
to minimize the fluctuations in day-to-day resource usage during the project. When
we compare the resource leveling under the neutrosophic interval method with the
resource leveling under the traditional and fuzzy methods, we find that the most ade-
quate and effective method that provide the parameters under uncertainty and also

13
Resource levelling problem in construction projects under…

R (T)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 T

Fig. 7  Final resource profile after applying the proposed procedure

provide range of parameter in the indeterminacy interval with more real conditions.
This work is done based on trapezoidal neutrosophic number. The numerical exam-
ple indicates the effectiveness of the proposed procedure for leveling resources in the
absence of accuracy and confirmation of the times of project activities. But in the
future, we plan to use the heptagonal neutrosophic number and pentagonal neutro-
sophic number. We also plan to use the neutrosophic model for handling the non-
renewable resource to minimize the cost trade-offs, and also we use the neutrosophic
model for best resource allocation problem as an approach for handling the renewable
resource in the case of limited resources are available. Also we can use neutrosophic
theory for scheduling resource-constrained project problem, and scheduling multi-
mode resource-constrained project problems based on neutrosophic environment.

Compliance with ethical standards 

Conflict of interest  Authors declare that there is no conflict of interest about the research.

References
1. Leu SS, Hung TH (2002) An optimal construction resource leveling scheduling simulation model.
Can J Civ Eng 29(2):267–275

13
 M. Abdel‑Basset et al.

2. Easa SM (1989) Resource leveling in construction by optimization. J Constr Eng Manag

115(2):302–316
3. Neumann K, Zimmermann J (2000) Procedures for resource leveling and net present value prob-
lems in project scheduling with general temporal and resource constraints. Eur J Oper Res
127(2):425–443
4. Hartmann S, Briskorn D (2010) A survey of variants and extensions of the resource-constrained
project scheduling problem. Eur J Oper Res 207(1):1–14
5. Kastor A, Sirakoulis K (2009) the effectiveness of resource leveling tools for resource constraint
project scheduling problem. Int J Project Manage 27(5):493–500
6. Rieck J, Zimmermann J (2015) Exact methods for resource leveling problems. In: Handbook on
project management and scheduling, Springer, Cham, vol 1, pp 361–387‫‏‬
7. El-Rayes K, Jun DH (2009) Optimizing resource leveling in construction projects. J Constr Eng
Manag 135(11):1172–1180
8. Rahman MA, Elazouni A (2015) Devising extended-duration schedules of enhanced resource lev-
eling. Can J Civ Eng 42(8):552–562
9. Rodrigues SB, Yamashita DS (2015) Exact methods for the resource availability cost problem. In:
Handbook on project management and scheduling, Springer, Cham, vol 1, pp 319–338‫‏‬
10. Atan T, Eren E (2018) Optimal project duration for resource leveling. Eur J Oper Res

266(2):508–520
11. Bhaskar T, Pal MN, Pal AK (2011) A heuristic method for RCPSP with fuzzy activity times. Eur J
Oper Res 208(1):57–66
12. Hossein Hashemi Doulabi S, Seifi A, Shariat SY (2010) Efficient hybrid genetic algorithm for
resource leveling via activity splitting. J Constr Eng Manag 137(2):137–146
13. Son J, Skibniewski MJ (1999) Multi-heuristic approach for resource leveling problem in construc-
tion engineering: hybrid approach. J Constr Eng Manag 125(1):23–31
14. Christodoulou SE, Michaelidou-Kamenou A, Ellinas G (2015) Heuristic methods for resource lev-
eling problems. In: Handbook on project management and scheduling, Springer, Cham, vol 1, pp
389–407
15. Ballestín F, Schwindt C, Zimmermann J (2007) Resource leveling in make-to-order production:
modeling and heuristic solution method. Int J Oper Res 4(1):50–62
16. Kreter S, Rieck J, Zimmermann J (2014) The total adjustment cost problem: applications, models,
and solution algorithms. J Sched 17(2):145–160
17. Liao TW, Egbelu PJ, Sarker BR, Leu SS (2011) Meta-heuristics for project and construction man-
agement–A state-of-the-art review. Autom Constr 20(5):491–505
18. Van Peteghem V, Vanhoucke M (2015) Heuristic methods for the resource availability cost problem.
In: Handbook on project management and scheduling, Springer, Cham, vol 1, pp 339–359
19. Ponz-Tienda JL, Yepes V, Pellicer E, Moreno-Flores J (2013) The resource leveling problem with
multiple resources using an adaptive genetic algorithm. Autom Constr 29:161–172
20. Leu SS, Yang CH, Huang JC (2000) Resource leveling in construction by genetic algorithm-based
optimization and its decision support system application. Autom Constr 10(1):27–41
21. Savin D, Alkass S, Fazio P (1996) Construction resource leveling using neural networks. Can J Civ
Eng 23(4):917–925
22. Leu SS, Chen AT, Yang CH (1999) A fuzzy optimal model for construction resource leveling sched-
uling. Can J Civ Eng 26(6):673–684
23. Masmoudi M, Hait A (2013) Project scheduling under uncertainty using fuzzy modeling and solv-
ing techniques. Eng Appl Artif Intell 26(1):135–149
24. Long LD, Ohsato A (2008) Fuzzy critical chain method for project scheduling under resource con-
straints and uncertainty. Int J Project Manage 26(6):688–698
25. Iyer P, Liu Y, Sadeghpour F, Brennan RW (2015) A fuzzy-logic based resource leveling optimiza-
tion tool. IFAC-P Line 48(3):1942–1947
26. Levy FK, Thompson GL, Wiest JD (1963) The ABCs of the critical path method. Harvard Univer-
sity, Graduate School of Business Administration, pp 98–108
27. Burgess AR, Killebrew JB (1962) Variation in activity level on a cyclical arrow diagram. J Ind Eng
13(2):76–83
28. Hezam IM, Abdel-Baset M, Smarandache F (2015) Taylor series approximation to solve neutro-
sophic multi-objective programming problem. Infinite Study, Coimbatore
29. El-Hefenawy N, Metwally MA, Ahmed ZM, El-Henawy IM (2016) A review on the applications of
neutrosophic sets. J Comput Theor Nanosci 13(1):936–944

13
Resource levelling problem in construction projects under…

30. Saaty TL (2006) The analytic network process. In: Decision making with the analytic network pro-
cess, Springer, Boston, pp 1–26
31. Abdel-Baset M, Hezam IM, Smarandache F (2016) Neutrosophic goal programming. Neutrosophic
Sets Syst 11:112–118
32. Mahdi IM, Riley MJ, Fereig SM, Alex AP (2002) A multi-criteria approach to contractor selection.
Eng Constr Architect Manag 9(1):29–37

Publisher’s Note  Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional affiliations.

13