Planning of tunneling projects using computer simulation and fuzzy decision making.
Abdallah, Moatassem ; Marzouk, Mohamed
Introduction
Construction of tunnels has gone through several phases of
development. Decades ago, tunnel construction included mainly manual
excavation and small sized equipment, which are used to carry out
different tasks in the construction operation. Afterward, it has gone
through different phases and developments of techniques until it reached
currently, a high degree of mechanization including: cranes, pile
drilling rigs, trench cutter machines, concrete/bentonite pumps, trucks,
mixers and/or tunnel boring machines (TBM). At the beginning of
construction, tunnel construction included fewer tasks, which were
manageable to be planned with traditional construction techniques.
However, the advancement in tunnel construction has increased the
sophistication, interconnection between activities, and uncertainties in
the construction operation, which can lead to deviation from original
work plans. Traditional planning techniques, such as preceding diagram
method (PDM), critical path method (CPM), Program evaluation and review
techniques (PERT) or line of balance method (LOB) are considered
practical techniques for scheduling construction operations. However,
they do not plan tunnel projects efficiently with the inherited
sophistications and uncertainties in tunnel construction. In addition,
they do not provide the efficient use of available resources in the
construction operation of tunnels. Furthermore, the new techniques in
construction of tunnels have improved constructability through segmental
operation, which made the construction to be carried out in a cyclic
manner (Halpin, Riggs 1992; Banks et al. 2000; Marzouk 2010).
Computer simulation has proved its efficiency in planning
construction projects for the following reasons: (1) the availability of
several probability density functions, which account for uncertainties
that might occur during construction, (2) taking into consideration the
interaction among available resources during the planning phase, and (3)
the application of WHAT-IF analysis and sensitivity analysis in planning
of construction projects to identify factors that might affect the
construction operation. Several efforts have been conducted in planning
of construction projects using computer simulation. These studies
include concrete operations (Hassan, Gruber 2008), bridge decks
construction (Marzouk et al. 2006, 2008a), earthmoving operations
(Marzouk 2002), and tunnel constructions (Al-Battaineh et al. 2006;
Tanaka 1993). Limited studies have been conducted for simulating the
construction of cut-and-cover tunnels with different supporting
techniques. However, few studies have been carried out for modeling
tunnels with circular and horse shoe cross-sections (AbouRizk et al.
1999; Loannou, Likkhitruangslip 2005). A special purpose simulation
template for shielded tunnels was developed using Simphony, where the
user is capable of adding as many tunnel sections to the model based on
soil properties in each section. In addition, the developed simulation
model was used to conduct a comparison between small and large tunnel
diameters in terms of productivity. The developed simulation model was
tested for different soil types and it showed that as the stiffness of
soil increases, the tunnel advancement rate decreases (AbouRizk et al.
1999).
Several techniques have been developed to model multicriteria
decision-making (MCDM). These techniques include utility-theory method
(Keeney, Raiffa 1993), analytical-hierarchy process (Saaty 1980, 1982),
superiority and inferiority method (Marzouk 2008; Xu 2001), ELECTRE III
(Marzouk 2011) and fuzzy trade-off evaluation method (Nishizaki, Seo
1994). Each method has its advantages and drawbacks in dealing with the
problem under consideration. Such techniques lack the ability to model
the uncertainties which are inherited in many of the decision-making
problems. Fuzzy numbers have the ability to model uncertainties in
multicriteria decision-making (MCDM). Several methods have been
introduced for ranking alternatives based on fuzzy numbers (Lee-Kwang,
Lee 1999; Modarres, Shadi-Nezhad 2001; Zhang, Lu 2003). Immense studies
have been conducted in decision making to select the best alternative
that best fits the designated goals. These studies included: assessing
intangible aspects of technical innovation in construction (Skibniewski,
Chao 1992), determination of bid(s) markup using utility-theory model
(Hassan, Gruber 2008), selection of best construction contractor
(Alsugair 1999), selection of finalists for a director position
(McIntyre et al. 1999), obtaining sustainable residential building based
on the acceptable level of environmental impact and socioeconomic
characteristics of residential buildings (Seo et al. 2004), and
selecting the most appropriate contractor for delivering a project
(Singh, Tiong 2005). This paper presents a framework that aids
contractors in planning tunnel projects using computer simulation. It
provides two main functions: (1) estimating time and cost required for
construction of tunnels, and (2) selecting best construction
alternative/technique among a set of different construction options,
considering five tunnel construction techniques. This paper provides an
overview of the developments made in the proposed framework.
1. Construction of tunnel projects
Tunnels are usually constructed in crowded cities for the purpose
of improving transportation networks by decreasing time required to move
from one place to another. Tunnels can be classified into two groups
based on the cross-sectional shape: rectangular cross-section and
circular cross-section. The former group is essentially constructed
using cut-and-cover technique, whereas, the latter group is constructed
using TBM. The following subsections describe these techniques, which
are used in tunnel construction projects.
1.1. Cut-and-cover techniques
Tunnel construction using cut-and-cover techniques offer an
alternative to boring machines, where a trench of required depth and
width can be excavated from the surface. The construction of
cut-and-cover tunnels in its simplest form includes: trench excavation,
building tunnel structure, backfilling tunnel trench, and the surface
restoration (EOT, U.S.DOT 2008). Supporting of ground soil along with
maintaining existing surface, underground facilities, and services
increase the complexity of tunnel projects. The key aspect to the
various cut-and-cover methods lies in supporting the vertical sides of
tunnel construction. Several techniques have been developed in
construction of cut-and cover tunnels, which are categorized based on
ground supporting methods; these techniques include: cut and-cover using
diaphragm walls, cut-and-cover using secant pile walls, cut-and-cover
using soldier piles and lagging, and cut-and-cover using steel sheet
pile walls. Despite the fact that cut-and-cover construction techniques
are one of the oldest techniques that are used in construction of
tunnels, it is still being used in construction due to the following
characteristics (EOT, U.S.DOT 2008):
--it is usually cheaper and more practical than other underground
tunneling, especially for tunnels with small lengths;
--cut-and-cover method is considered as an appropriate tunneling
technique for construction of tunnels with small depths;
--the risk that is taken by the contractor in the construction of
tunnels using this method is considered as small, relative to other
construction techniques;
--it may cause interference with traffic and other urban
activities, but this disturbance is decreased by construction of the
tunnel top slab after excavation or by using temporary decking over the
excavation. This temporary deck is left in place, where construction
activities can be carried out underneath until reaching final
backfilling and surface restoration.
Detail description of cut-and-cover using diaphragm walls and
cut-and-cover using secant pile walls can be found elsewhere (Abdallah
2008; Marzouk et al. 2008b, 2009). Cut-and-cover using soldier piles and
lagging technique is considered to be one of oldest retaining systems
that are commonly used in supporting deep excavations. Soldier piles and
lagging walls are constructed in a cyclic manner by placing soldier
piles at regular intervals (2-4 m), then excavating and installing
lagging between soldier piles. Soldier piles and lagging walls are the
most inexpensive systems compared to other retaining walls (FORASOL
2008). Although, soldier piles and lagging walls are very easy and fast
to construct, they have the following disadvantages (Henery Drilling
2011):
--primarily limited to temporary construction;
--inapplicable when ground water table is near to ground surface;
--significant surface settlements in case of poor backfilling;
--less stiffness than other retaining systems;
--difficult to control soil movements as the flange of soldier
piles are embedded beneath subgrade.
Tunnel construction with cut-and-cover method using soldier piles
and lagging is performed by dividing tunnel length into equal segments
(20 ~ 30 m). It involves eight main processes (Land Transport Authority
2004): (1) segments preparation, (2) installing of soldier piles and
excavation, (3) construction of anchors, (4) excavation, (5)
construction of bottom slab segments, (6) construction of side wall
segments, (7) construction of top slab segments, and (8) segments
backfilling (see Fig. 1). The first process of segments preparation
involves surveying to coordinate the position of each segment according
to the tunnel path, and excavation and soil leveling till the top level
of soldier piles. The second process of installing of soldier piles and
excavation is started by installing soldier piles into the ground, and
then excavating the soil between soldier piles to install timber
lagging. The installation of soldier piles is executed by first drilling
the soldier pile hole using a pile drilling rig, then placing soldier
piles into the holes. Once the soldier pile is installed into position,
concrete can be poured into the bottom part of the hole. Then,
backfilling the rest of the drilled hole is performed. This operation is
repeated to install all soldier piles in the segment.
The third process of construction of anchors is started by drilling
the anchor hole using the anchor drilling rig. Then, a wire bundle is
installed into the anchor hole. As such, the anchor hole can be filled
with cement. Once the cement is settled, the grout can be pumped into
soil behind the anchor to stick it with soil and increase its strength.
After settlement of grout, the anchor can be tensioned to connect the
anchor force to the soldier pile. The fourth process of excavation is
started by excavating soil between each two soldier piles and installing
timber lagging between them. During the excavation between soldier piles
and installation of timber lagging, the excavation in the center of
segment can be executed separately. After finishing the previous tasks,
the excavation for the rest of the segment can be executed.
The fifth process of construction of the bottom slab segments is
started by excavating the soil till the base of the plain concrete.
Then, pouring of plain concrete takes place. After that, the
reinforcement bars, which are either fabricated on-site or at off-site
workshop, are placed to form the reinforcement cage of the bottom slab
segment of the tunnel. Then, forms can be erected and water-stop can be
installed. At this stage, concrete is ready to be poured and cured, and
finally, forms can be removed. The sixth process of construction of side
walls segments is started after removal of forms of the bottom slab. The
walls of a segment are constructed by fixing the steel reinforcements in
the tunnel sides to form the steel cage. Once the steel is fixed, forms
of the tunnel walls can be erected (see Fig. 2). After assembling of
forms, concrete can be poured and cured, and forms can be removed when
concrete gains enough strength that allows the construction of the next
process. The seventh process of construction of top slab segments is
started by shuttering of forms for the top slab to provide support for
concrete, while achieving sufficient strength to support its own weight
and loads. The final process of segments backfilling is executed after
removing the forms of the tunnel top slab. The tunnel may be backfilled
with clean soil or by the excavated soil according to its quality and
specifications. These processes are repeated for each segment.
Cut-and-cover using steel sheet pile walls technique involves the
use of sheet pile walls, which are essentially rows of interlocking
vertical pile segments that are installed to form an efficiently
straight wall with a planned dimension sufficiently large enough to
retain soil. Steel sheet pile walls are used in soft grounds specially
when there is danger of bottom heave in soft clay soil or in the case of
sand. Tunnel constructions using steel sheet pile walls have the
following advantages: suitable for various service conditions, easy
speed driving of sheet pile walls and easy storage and shipping. On the
other hand, steel sheet piling have the following disadvantages: causes
noise and vibrations when the vibratory-hammer is utilized, involves
high cost relative to other retaining methods, can be used only when the
sheet piles are not required to be driven deeply into the ground,
permits large movements in weak soils, and requires effective
de-watering since it cannot provide a watertight boundary (Deep
Excavation 2011).
[FIGURE 1 OMITTED]
The tunnel construction with cut-and-cover method using steel sheet
pile walls is performed by dividing the tunnel length into equal
segments (20 ~ 30 m). It involves eight main processes (Land Transport
Authority 2004): (1) segments preparation, (2) installing of sheet
piles, (3) dewatering and excavation, (4) construction of anchors and
installing of steel anchor connecting beams, (5) construction of the
bottom slab segments, (6) construction of the side wall segments, (7)
construction of the top slab segments, and (8) segments backfilling. All
processes are similar to tunnel construction with cut-and-cover method
using soldier piles and lagging processes except for the second process
of installation of sheet piles. This process starts by locating and
placing of a beam into the ground to set out the position of the sheet
pile wall. Then, the piling rig/vibrator lifts up the first sheet pile
and drives it into the ground, leaving about 1 meter length of the sheet
pile above the ground level. After that, the piling rig/vibrator drives
the second sheet pile into the ground, where the second sheet pile
interlocks with the first one. This process is repeated till
installation of all sheet piles (see Fig. 3).
[FIGURE 2 OMITTED]
1.2. Segmental tunneling using slurry TBM
Circular tunnels are usually constructed using closed/ open face
TBM. A TBM is a complex set of equipments, assembled to excavate a
tunnel. It is manufactured to bore through hard rocks or sand layers and
any type of soil. There are several types of closed face TBM s based on
soil conditions and tunnel lining. In case of hard soil or rock,
machines are built to advance through hard material that is usually
self-supported, and have tools that are made for breaking the hardest
rocks. The excavation is carried out at atmospheric pressure, and the
extraction of material is performed using trains, trucks or conveyor
belts. In case of soft soils, excavation is executed through a turning
cutting wheel. The excavated material is usually handled by a hydraulic
transportation system. This is done by using either water or bentonite
as a transportation medium. The lining of the tunnel may be in-situ
pressed concrete lining which is poured during the excavation of the
tunnel, or pre-cast concrete segments which are installed during
excavation of the tunnel. The segmental tunneling method using closed
face TBM is used for circular cross-section. It is worth noting that,
the construction of tunnels using in-situ pressed concrete lining gives
slow productivity, which is not consistent in urban and crowded areas.
The slurry shield has developed in recent decades for managing
instability of excavation profile in unfavorable geotechnical
conditions. With slurry TBM, the unstable/soft ground at the front is
supported by liquid mixture (bentonite or water) under increased
pressure generating a steady counter pressure. The construction of
circular tunnels using slurry TBMs has the following advantages: high
progress rate, especially in soft ground soil, continuous operation,
less noise and disturbance to surrounding structures, in addition to
being the best way for constructing deep and long tunnels. On the other
hand, slurry TBMs have the following disadvantages: fixed circular
geometry, limited flexibility in response to extremes of geologic
conditions, longer mobilization time, and higher capital costs.
[FIGURE 3 OMITTED]
The construction sequence of circular tunnels using slurry TBM is
divided into four processes: soil injection, TBM setup, tunnel
construction, and TBM dismantle. The first process of injecting soil is
executed for the break in and break out regions of the TBM, as the soil
in these two regions may cause the boring machine to deviate from the
tunnel path in case of weak soil. In addition, segments in the break in
and break out points have to be constructed in stiffer soil. This
process is executed by locating injecting points in the break in region
using survey, and then drilling the located points and injecting them
with cement. This process is repeated till injecting all points in the
break in region are completed. After that, laborers and equipment are
moved to the break out point to inject soil with cement using the same
sequence as of the break in region. It should be noticed that, this
process is not necessary in case of hard or stable soil. The second
process, TBM setup, can be started while injecting the soil of break in
region with cement. This process is started by constructing the buttress
wall that will be used to support the boring machine to get into soil,
and after that, the TBM and its components can be installed. The TBM
setup includes installing of steel shield, back-up tail, separation
plant, feed pump, and slurry pump.
The third process, tunnel construction, starts after the settlement
of cement in the break in region and completion of the second process.
It begins with installing temporary steel segments between the machine
and the buttress wall. Then, the boring machine can start excavation of
soil. During soil excavation, the excavated material is moved to the
shaft using a transportation system, either by using water or bentonite.
The transportation medium is pumped to the tunnel face from the start
shaft by one or several feed pumps in a feed line. The transportation
medium is utilized to stabilize and support the tunnel face and to
facilitate the excavation process. The mixture of soil and
transportation medium is exhausted out of the excavation chamber through
the slurry line and conveyed to the separation plant. To prevent the
slurry line from getting blocked, any major pieces of rock are
pulverized by the cone crusher in the working chamber before being
passed into the slurry line. In the separation plant, the transportation
medium is separated from the loose soil using screens, cyclones and
centrifugal pumps if necessary. Efficient separation means that a large
proportion of the medium can be treated and sent back into
transportation circuit. The separated soil is then loaded into trucks to
be transported to the dumping area.
When the advancement of the machine reaches a distance equal to the
length of a ring, the excavation stops and the pushing jacks are
retrieved. Then a temporary steel ring is installed and the pushing arms
are extended to resume excavation. This sequence is repeated till the
last temporary steel ring is installed in front of the soil. After the
machine reaches a distance equal to the length of a ring, the excavation
stops and the pushing jacks are retrieved. Then, a pre-cast concrete
ring which consists of number of segments is installed. Subsequently,
the pushing arms are extended once again in full contact with the
concrete ring to resume excavation. During excavation, a grout is pumped
to fill the space generated between the precast concrete ring and the
soil. The cycle of excavation and ring erection is repeated as the TBM
advances to form the lining of the tunnel.
[FIGURE 4 OMITTED]
During soil excavation, the pre-cast concrete segments are loaded
into the TBM's train and then transported to the tunnel face. After
installing the segments, the train is transported to the shaft to get
the next segments. The TBM has to follow the pass carefully from the
driving to the receiving shafts using a laser guiding system. The laser
guiding system determines the orientation of the machine head to make
any needed corrections in the tunnel path. During the excavation of the
tunnel, the extension of the articulated jacks allows the TBM to turn
and advance forward in the direction of the tunnel designed axis. After
installing sufficient number of pre-cast concrete rings (10-15 rings),
the buttress wall can be dismantled. As such, the boring machine can be
jacked on the installed pre-cast concrete rings based on the own weight
of the installed rings and grout around them. The final process of TBM
dismantle is executed when the TBM reaches the receiving shaft. It
starts by dismantling the boring machine with its components. Finally,
the tunnel can be cleaned.
2. Proposed simulation framework
The proposed simulation framework aids contractors in planning
tunnel constructions. It performs two main functions: (1) planning and
analyzing tunnel construction and (2) selecting the best construction
alternative based on pre-defined criteria. The framework consists of
three main components: tunnel analyzer module, simulation module, and
decision support module. Figure 4 depicts a schematic diagram for the
proposed framework that shows the interaction between its main
components. Detail description of the framework can be found elsewhere
(Abdallah 2008). The following subsections describe in detail the
components of the framework.
2.1. Tunnel analyzer module
Tunnel analyzer module is considered to be the interface and
coordinator of the planning function for the framework. Tunnels are
broken down into zones taking into consideration the following factors:
--Project is divided into a number of zones, where the construction
method is assigned and defined in each zone;
--The assigned resources in each zone are defined independently of
the other zones (i.e. for a project that is divided into two zones, a
construction method is defined in each zone and resources are assigned
independently for each construction zone even with the same construction
method);
--General sequence of construction is defined for the project. For
example, two construction zones that are constructed with the
same/different construction method and are required to be executed
simultaneously or successively. Each zone is defined separately with a
specific relationship between them (Finish to Start or Start to Start
with lag).
The procedure followed by tunnel analyzer module for planning
tunnels can be summarized as follow:
--Assigning project general data, such as number of working hours
per day, number of working days per week, project start date, number of
zones in the project, and indirect cost required for the project. Also
in this stage, the user is required to set number of simulation runs
which will indicate the accuracy of the simulation output;
--Defining the construction method for each zone. Then, assign the
required data for each zone including: general data of the assigned
construction method, duration of tasks and corresponding probability
density function (e.g. beta, erlang, exponential, gamma, normal, pert,
pertpg, scaled beta, triangular, and uniform), required number of
resources for each task, available resources in each zone, labor and
equipment rates, and materials costs. Then, the module saves the
collected data and sends it to the simulation module;
--Subsequently, simulation module is triggered to estimate the
duration and utilization of resources for each zone. Then, simulation
outputs are sent to tunnel analyzer module in order to calculate project
execution time and costs;
--Finally, tunnel analyzer module presents the estimated data to
the user.
Tunnel costs are classified into direct cost and indirect cost. The
direct cost involves materials, labor, and equipment costs, while
indirect cost involves items which depend on project duration and others
items that are independent of project execution time. Material costs are
calculated by summing up costs of each item used in project
construction, such as concrete quantities, steel bars, forms, pipes,
bentonite, and so on. Labor and equipment costs are calculated using
Eqns (1) and (2). Indirect cost and project total cost are calculated
using Eqns (3) and (4):
TL = [m.summation over (i=1)] [n.summation over (j=1)]
LU[C.sub.j,j] x LU[T.sub.i,j] x Z[D.sub.i]; (1)
TL = [m.summation over (i=1)] [n.summation over (j=1)]
EU[C.sub.j,j] x EU[T.sub.i,j] x Z[D.sub.i], (2)
where TL--total labor costs; TE-total equipment costs, m, number of
zones; n, number of resources in each zone; LUC, labor unit cost/unit
time; EUC, equipment unit cost/unit time; LUT, utilization for specific
labor crew (time elapsed for utilizing a specific labor crew per zone
duration, the developed simulation module calculates a percentage, where
a labor crew is utilized in the construction of each zone); EUT,
utilization for specific equipment (time elapsed for utilizing a
specific equipment per zone duration, the developed simulation module
calculates a percentage, where an equipment is utilized in the
construction of each zone); and ZD, zone duration.
TIC = [summation] TDIC x PD + [summation] TIIC, (3)
where TIC--total indirect cost; TDIC--time dependent indirect
costs; PD-project duration; TIIC--time independent indirect costs.
PC = TM + TE + TL + TIC, (4)
where PC-project cost and TM-total material costs.
2.2. Simulation module
Simulation module is responsible for estimating total duration of
construction and utilization of resources. The proposed module uses
STROBOSCOPE as a general purpose simulation language (Martinez 1996). It
is developed using Microsoft Visual Basic 6.0 language to control and
enter data to STROBOSCOPE simulation engine. Five models have been
developed using STROBOSCOPE simulation language in the simulation module
as listed in Table 1. These simulation models were built based on
construction sequence and available simulation elements to mimic the
different tunnel construction techniques described earlier.
The process followed by the simulation module to estimate
construction duration and resource utilization is summarized as follows:
Simulation module receives project data from tunnel analyzer
module, and then, it depicts the model that represents the construction
technique for each zone from models that are stored in the simulation
module's library. Subsequently, each model is fed by its data
(general data, duration of tasks, required and available resources, and
number of replications) to generate the simulation model that represents
the project described in tunnel analyzer module:
--STROBOSCOPE is triggered to run the generated model(s) in order
to estimate total duration and utilization of resources in each zone;
--The output data are transferred to tunnel analyzer module to
perform total project duration and cost calculations.
Developing a simulation model for construction of tunnels by
cut-and-cover method using diaphragm walls involves the following
assumptions:
--tunnel width is constant along its length;
--the tunnel is divided into zones (inlet, body, and outlet), which
are divided into a number of segments with equal lengths;
--the widths of primary and secondary diaphragm trenches are
constant;
--inlet, body, and outlet zones of the tunnel are represented by
the same model, but with different input data (i.e. for the inlet and
outlet zones, they are represented by the same developed model, but the
durations for tasks of the top slab segments are assigned with zero
values);
--in case of constructing a plug to save the surrounding structures
from dewatering problems, guide walls and diaphragm trenches of the
cutting wall are constructed first (i.e. cutting walls are constructed
before the side diaphragm walls of the tunnel).
Cut-and-cover using diaphragm walls method involves eight
processes: (1) segments preparation, (2) construction of guide walls,
(3) construction of diaphragm walls, (4) construction of plug and
dewatering, (5) construction of top slab segments, (6) construction of
bottom slab segments, (7) construction of un-casted top slab segments,
and (8) segments backfilling. Table 2 lists resources needed for
developing the model of cut-and cover method using diaphragm walls.
Table 3 lists input parameters of cut-and-cover method using diaphragm
walls. Figure 5 depicts elements of the simulation model that represent
cut and cover using diaphragm walls. Resources, listed in Table 2, are
allocated in queues and filled with available number of resources. Each
Combi activity draws the required number of resources from the needed
queues in order to be executed.
The process of segments preparation consists of two tasks: (1)
segment survey, and (2) general excavation and leveling for segment.
This process is repeated for each segments of the tunnel. It should be
noted that: dummy queues, which are utilized in developing the model,
are used to maintain the logic flow and dependency between activities.
After finishing the first process, the dummy queue, named
"L02," and Combi logic activity, named "Logic02"
with zero duration, are used to add the number of guide walls needed for
constructing cutting wall (if needed). The second process, named
construction of guide walls, consists of eight activities: (1)
excavation, (2) steel work, (3) shuttering of forms, (4) concrete
pouring, (5) concrete curing, (6) removal of forms, (7) backfilling, and
(8) paneling and marking. After finishing each segment of the first
process, the last activity in first process sends number of guide walls
in one segment to the first dummy queue in the next process. It should
be noted that, the second process is repeated for each guide wall in a
zone.
[FIGURE 5 OMITTED]
After finishing the second process there is a dummy queue, named
"L03", and Combi logic activity, named "Logic03"
with zero duration, used to add the number of diaphragm wall trenches
needed for construction of cutting wall (if needed). The third process,
named construction of diaphragm walls, consists of four tasks: (1)
preparing and fixing of steel cage, (2) trench excavation, (3) position
steel cage into excavated panel, and (4) concrete pouring. After
construction of each guide wall, second process, it sends number of
diaphragm trenches in one guide wall to dummy queue, named
"Q9", to be the initiation of the third process. The third
process is repeated for each trench of the diaphragm wall in the zone.
After finishing the third process, the dummy queue, named
"L04", and Combi logic activity, named "Logic04"
with zero duration are used to create a lag between the third process
and the fifth one. It should be noted that the lag is captured by
setting a number of segments between those two processes. Similarly, the
remaining fourth to seventh processes are modeled. The simulation model
runs till the number of segments reaches a dummy queue "End"
and it stops where there is no dummy resource to be initiated. More
details about modeling tunnel construction using diaphragm walls, secant
pile walls, steel sheet pile walls, TBM, micro tunneling are available
elsewhere (Abdallah 2008; Marzouk et al. 2009).
An application example is considered to demonstrate the generated
outputs from the Simulation Module. This example considers the
construction of Giza tunnel project to solve a traffic jam problem
through the elimination of a traffic light while converting the other
traffic flow through an underground tunnel, as shown in Figure 6. The
tunnel has a total length of 450 meters and a slightly varying width
with an average of 8.5 meters, and secant pile walls technique was used
in order to support ground while tunnel was being constructed, and also
to act as the side supports of the tunnel. The tunnel has varying
cross-section shapes including: retaining walls at the entrance and
exit, U-section, and a box section in the middle; five zones were
considered in order to account for these varying cross-section shapes
(see Fig. 6). In order to model the Giza tunnel in the simulation
module, project general data should be defined at the beginning, which
includes working hours per day; working days per week and project start
date. After that the project specific data should be fed to the module
in order to simulate the tunnel and provide its results; these data
include tunnel construction technique, task durations, number of
equipment and crews to be used in the project, labor, equipment and
material unit costs, and relationship between zones. Tunnel construction
using secant pile walls include nine processes, while Table 4 shows an
example for the input durations and resources data for three processes
of the Giza tunnel. The relationships between zones can be defined
throughout four relationships including FS, SS, SF, and FF; Figure 7
shows the relationships between these zones for the Giza tunnel. In
order to facilitate data entry and show the output results of the
simulation module, a user friendly tool was developed using Visual Basic
6.0. The output of Giza tunnel is shown in Figure 8. The output results
of the simulation module were validated as they comply with the actual
construction of the tunnel. More details about modeling, input data, and
analysis of Giza tunnel are available elsewhere (Abdallah 2008).
[FIGURE 6 OMITTED]
2.3. Decision support module
Decision support module is responsible for selecting the best
alternative of tunnel construction based on pre-defined criteria and a
group of experts. Fuzzy techniques have been increasingly applied to
construction management research area (Chan et al. 2009). The Fuzzy
Logic tool was introduced in 1965, by Lotfi Zadeh, to provide a
technique to deal with imprecision and information granularity. The
fuzzy theory provides a mechanism for representing linguistic
constructs, such as "many", "low",
"medium", "often", "few". In general, the
fuzzy logic provides an inference structure that enables appropriate
human reasoning capabilities. Particular linguistic assessment terms, so
called fuzzy linguistic variables, are introduced to represent the
underlying fuzzy numbers for criteria evaluations. A fuzzy linguistic
variable is defined as an expression in natural or artificial language
which describes a collection of values (Zadeh 1975; Cakir, Canbolat
2008).
The proposed decision support module uses fuzzy numbers for
preferences of criteria in order to deal with the fuzziness of the
decision maker preferences. Due to the subjective judgment of decision
makers, a belief level was used to express the preferences of decision
makers for each criterion within an alternative. This belief level
belongs to a set of linguistic terms which contain various degrees of
preferences. These linguistic terms and degrees of preferences include
seven levels that are varying from "Very Low" to "Very
High" and represented with triangular fuzzy numbers, as shown in
Table 5. These degrees of preference can be used by decision makers to
evaluate the different criteria under each alternative. This Decision
Support method also aggregates the group decision in a manner that is
most acceptable for the group as a whole The relative importance among
decision makers can be determined based on the number of years of
experience in the field of tunnel construction. Five levels that were
set to account for relative importance among decision makers include
"Less than 5 years", "5 years to 10 years", "10
years to 15 years", "15 years to 20 years", and
"More than 20 years". These levels were represented by a
quantitative scale (1, 2, 3, 4, and 5) respectively, where the
normalized average weight can be determined. Similarly, the criteria
weights can be determined using five level linguistic weights including
"Very Important", "Rather Important",
"Important", "Less Important", and
"Unimportant". These levels were represented with quantitative
scale (1, 2, 3, 4, and 5), respectively, where the relative normalized
weight among criteria can be computed for each decision maker (Zhang, Lu
2003). The procedures that are followed by the proposed framework for
selecting the best alternative for tunnel construction can be listed as
follows:
-- Define selection criteria: experts define the requirements that
each alternative must own, to be better or less important than other
alternatives. The decision support module defines three main criteria:
project cost, project duration, and resources utilization. These three
main criteria are obtained from tunnel analyzer module. Decision support
module can consider another two user-defined criteria to provide
flexibility in selecting alternatives. These two criteria can be defined
separately by the experts, such as utilization of a specific important
resource. Also, criteria weights and expert years of experience are
defined by each expert. Each criterion and expert years of experience
are defined through five levels of importance;
-- Defining alternatives criteria: at this stage, the output data
from tunnel analyzer module can be used to assign values for criteria of
each alternative. These data include project cost, duration, and/or
resources utilization. Also, each expert assigns preferences of criteria
through three ranges low, medium and high;
-- Ranking alternatives: alternatives are categorized based on the
collected data and fuzzy group decision-making method to select the best
alternative. The alternative with the highest preference value is
considered as the best one;
-- Refining selection: experts can refine results by ignoring
alternatives with small far preference values or editing the selection
criteria in the first stage.
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
Planning and analyzing tunnel construction using computer
simulation have been demonstrated using case studies elsewhere (Marzouk
et al. 2008b, 2009, 2010). The decision support module is explained in
this paper via a numerical example to demonstrate its use. A contractor
wants to select the best tunnel construction method among three possible
construction alternatives. The tunnel under consideration has dimensions
of 0.5 km in length and 10.5 m in width. Two experts are represented for
selecting the best construction alternative. The first expert has an
experience of 18 years in tunnel field, whereas the second expert has an
experience of 8 years in the same field. They choose the basic three
criteria, project cost, project duration, and utilization of resources,
as the basic elements in selecting the best alternative. Table 6 lists
the three criteria ranges for the two involved experts.
Criteria weights are leveled and listed by each expert in Table 7.
The criteria values, project cost, project duration, and utilization of
resources, for the three alternatives are listed in Table 8.
To get the most acceptable solution among the three alternatives,
the following steps are carried out (Zhang, Lu 2003).
Step 1. Table 9 lists criteria levels, scale and weights for each
expert.
Step 2. After dividing the three criteria ranges defined in Table 6
into seven ranges, belief level matrixes can be established for
alternatives as shown in Eqn (5):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)
Step 3. Aggregating belief vectors. Table 10 lists belief vectors
and their calculated values. Belief vector values are calculated for
[[??].sup.1.sub.1] and [[??].sup.1.sub.2] as shown in Eqn (6):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)
where: [[??].sup.i.sub.j] is belief vector for a decision maker i
and alternative j; [w.sup.i.sub.k]-weight for a criterion k and a
decision maker i.
Step 4. The weight for each expert can be determined as per Table
11.
Step 5. Determination of fuzzy decision vectors. Table 12 lists the
calculated values of fuzzy decision vector. For example [[??].sub.1] and
[[??].sub.2] are calculated using values of belief vectors (listed in
Table 9), experts' years of experience weight ([v.sup.*.sub.1] and
[v.sup.*.sub.2] which are listed in Table 10) as shown in Eqn (7):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)
Step 6. Calculating the fuzzy positive and negative solution
distances. Table 13 lists the calculated values of solution distances.
Solution positive ([d.sup.+.sub.1]) and negative ([d.sup.-.sub.1])
distances are calculated as shown in Eqn (8).
Step 7. Finally, calculation of closeness coefficient. Table 4
lists values of closeness coefficient.
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)
Closeness coefficient is calculated (e.g.CC1) as shown in Eqn (9):
CC1 =1 (df + (1 - d+))= 1 (0.488 (9) + (1 - 0.564)) = 0.462.
From Table 12, [CC.sub.1] has the highest closeness coefficient,
therefore, the best alternative based on the fuzzy decision-making
method is the first one.
Conclusions
This paper presented a framework that aids contractors in planning
tunnel projects using computer simulation. The proposed tool estimates
time and cost required for construction of tunnels. Furthermore, it aids
contractors to select the best alternative among a set of available
alternatives in construction. This paper presented a framework that is
able to model five techniques that are used in tunnel constructions.
These techniques include: cut-and-cover using diaphragm walls,
cut-and-cover using secant pile walls, cut-and-cover using soldier piles
and lagging, cut-and-cover using steel sheet pile walls and segmental
tunneling using slurry TBM. The proposed framework consists of three
modules: tunnel analyzer module, simulation module, and decision support
module. Tunnel analyzer module is considered as the coordinator of the
decision support tool. It collects needed data for planning construction
of tunnels. Simulation module is responsible for estimating total
duration and utilization of resources in each zone of the tunnel. Five
models for construction of tunnels have been coded in the simulation
module to represent the different construction techniques. Simulation
module receives its input data from tunnel analyzer module, where it
generates an input file for each zone according to the utilized
construction method. Then, it estimates construction duration of zone
and utilization of resources. Decision support module is responsible for
selecting the best construction alternative using a fuzzy group
decision-making method. It selects the best alternative based on
pre-defined criteria and a group of experts. It considers up to five
criteria, three of them are considered main/basic criteria: project
cost; project duration and utilization of resources, while, the
remaining two criteria are defined by the user to allow more flexibility
to the system. The decision support module analyzes alternatives to
obtain the best construction option. A numerical example was presented
in this paper to illustrate the main features of the decision support
module.
doi: 10.3846/13923730.2013.793608
References
Abdallah, M. 2008. Decision support tool for planning tunnels
projects using computer simulation. MSc thesis. Structural Engineering
Department, Cairo University, Egypt.
AbouRizk, S. M.; Er Ruwanpura, J. Y.; Fernando, K. C. 1999. Special
purpose simulation template for utility tunnel construction, in Proc. of
the 1999 Winter Simulation Conference, Phoenix, AZ, 948-955.
Al-Battaineh, H. T.; AbouRizk, S. M.; Tan, J.; Fernando, S. 2006.
Productivity simulation during the planning phase of the Glencoe in
Calgary, Canada: a case study, in Proc. of the 2006 Winter Simulation
Conference, Monterey, CA, 2087-2092.
Alsugair, A. M. 1999. Framework for evaluating bids of construction
contractors, Journal of Management in Engineering ASCE 15(2): 72-78.
http://dx.doi.org/10.1061/(ASCE)0742-597X(1999) 15:2(72)
Banks, J.; Carson, J. S.; Nelson, B. L.; Nicol, D. M. 2000.
Discrete-event system simulation. New Jersey, NJ: Prentice Hall. 640 p.
Deep Excavation. 2011. Sheet pile walls: retaining systems for deep
excavation: sheet pile walls, deep excavation LLC [online], [cited 11
November 2011]. Available from Internet:
http://www.deepexcavation.com/en/sheet-pile-walls.
Cakir, O.; Canbolat, M. S. 2008. A web-based decision support
system for multi-criteria inventory classification using fuzzy AHP
methodology, Expert Systems with Applications 35(3): 1367-1378.
http://dx.doi.org/10.1016Zj.eswa.2007.08.041
Chan, A. P. C.; Chan, D. W. M.; Yeung, J. F. Y. 2009. Overview of
the application of fuzzy techniques in construction management research,
Journal of Construction Engineering and Management ASCE 135(11): 1241
-1252. http://dx.doi.org/10.1061/(ASCE)CO.1943-7862. 0000099
EOT, U.S.DOT. 2008. Technical Tunnel Alternatives summary Report
Urban Ring Phase 2, Circumferential Transportation Improvements in the
Urban Ring Corridor. Massachusetts Executive Office of Transportation
(EOT), USA, U.S. Department of Transportation (U.S. DOT).
FORASOL. 2008. Secant Pile / Berliner Wall, Forasol Travaux
Speciaux [online], [cited 11 August 2010]. Available from Internet:
http://www.forasol.com/.
Halpin, D. W.; Riggs, L. S. 1992. Planning and analysis of
construction operations. New York, NY: John Wiley & Sons. 400 p.
Hassan, M. M.; Gruber, S. 2008. Simulation of concrete paving
operations on Interstate -74, Journal of Construction Engineering and
Management ASCE 134(1): 2-9.
http://dx.doi.org/10.1061/(ASCE)0733-9364(2008) 134:1(2)
Henery Drilling. 2011. Soldier pile and lagging walls. Henery
Drilling: specialized contractor in all types of drilled deep
foundations and drilled earth retention systems [online], [cited 11
November 2011]. Available from Internet:
http://henrydrilling.com/Soldier_pile_walls.html.
Keeney, R. L.; Raiffa, H. 1993. Decisions with multiple objectives:
preferences and the value tradeoffs. New York: Cambridge University
Press. 592 p.
Land Transport Authority. 2004. Construction of contiguous bored
pile walls [online]. Land Transport Authority[cited 11 November 2011].
Available from Internet:
http://www.lta.gov.sg/projects/images/CBP%20Final.pdf
Lee-Kwang, H ; Lee, J -H 1999 A method for ranking fuzzy numbers
and its application to decision making, IEEE Transactions on Fuzzy
Systems 7(6): 677-684. http://dx.doi.org/10.1109/91.811235
Loannou, P. G.; Likhitruangslip, V. 2005. Simulation of
multiple-drift tunnel construction with limited resources, in Proc. of
the 2005 Winter Simulation Conference, Orlando, FL, 1483-1491.
Martinez, J. C. 1996. STOBOSCOPE, State and resource based
simulation of construction processes. PhD thesis. University of
Michigan, USA.
Marzouk, M. 2002. Optimizing earthmoving operations using computer
simulation. PhD thesis. Department of Building, Civil, and Environmental
Engineering, Concordia University, Canada.
Marzouk, M. 2008. A superiority and inferiority ranking model for
contractor selection, Construction Innovation: Information, Process,
Management 8(4): 250-268. http://dx.doi.org/10.1108/14714170810912644
Marzouk, M. 2010. The state of computer simulation applications in
construction, in Abu-Taieh, E. M. O.; El Sheikh, A. A.; (Eds.). Handbook
of research on discrete event simulation environments: technologies and
applications. Pennsylvania, PA: IGI Global, 1554-1575.
Marzouk, M. M. 2011. ELECTRE III model for value engineering
applications, Automation in Construction 20(5): 596-600.
http://dx.doi.org/10.1016/j.autcon.2010.11.026
Marzouk, M.; Zein, H.; El-Said, M. 2006. BRIGE_SIM: framework for
planning and optimizing bridge deck construction using computer
simulation, in Proc. of the 2006 Winter Simulation Conference, Monterey,
CA, 2039-2046.
Marzouk, M.; Said, H.; El-Said, M. 2008a. Special purpose
simulation model for balanced cantilever bridges, Journal of Bridge
Engineering ASCE 13(2): 122-131.
http://dx.doi.org/10.1061/(ASCE)1084-0702(2008) 13:2(122)
Marzouk, M.; Abdallah, M.; El-Said, M. 2008b. Tunnel_ Sim: decision
support tool for planning tunnel construction using computer simulation,
in Proc. of the 2008 Winter Simulation Conference, Miami, FL, 2504-2511.
Marzouk, M.; Abdallah, M.; El-Said, M. 2009. A computer simulation
model for cut and cover tunneling using secant pile walls, Arab
Construction World (ACW) Magazine 27(3): 18-20.
Marzouk, M.; Abdallah, M.; El-Said, M. 2010. Modeling
microtunneling projects using computer simulation, Journal of
Construction Engineering and Management ASCE 136(6): 670-682.
http://dx.doi.org/10.1061/(ASCE)CO.1943-7862. 0000169
McIntyre, C.; Kirschenman, M.; Seltveit, S. 1999. Applying decision
support software in selection of division director, Journal of
Construction Engineering and Management ASCE 15(2): 86-92.
http://dx.doi.org/10.1061/(ASCE)0742-597X(1999) 15:2(86)
Modarres, M.; Shadi-Nezhad, S. 2001. Ranking fuzzy numbers by
preference ratio, Fuzzy Sets and Systems 118(3): 429-436.
http://dx.doi.org/10.1016/S0165-0114(98)00427-8
Nishizaki, I.; Seo, F. 1994. Interactive support for fuzzy
trade-off evaluation in group decision-making, Fuzzy Sets and Systems
68(3): 309-325. http://dx.doi.org/10.1016/0165-0114(94)90186-4
Saaty, T. L. 1980. The analytic hierarchy process: planning,
priority setting, resource allocation. New York: McGraw Hill. 287 p.
Saaty, T. L. 1982. Decision making for leaders: the analytical
hierarchy process for decisions in a complex world. Belmont, CA:
Wadsworth. 292 p.
Seo, S.; Aramaki, T.; Hwang, Y.; Hanaki, K. 2004. Fuzzy
decision-making tool for environmental sustainable buildings, Journal of
Construction Engineering and Management ASCE 130(3): 415-423.
http://dx.doi.org/10.1061/(ASCE)0733-9364(2004) 130:3(415)
Singh, D.; Tiong, R.L.K.2005.Afuzzydecision framework for
contractor selection, Journal of Construction Engineering and Management
ASCE 131(1): 62-70. http://dx.doi.org/10.1061/(ASCE)0733-9364(2005)
131:1(62)
Skibniewski, M. J.; Chao, L.-C. 1992. Evaluation of advanced
construction technology with AHP method, Journal of Construction
Engineering and Management ASCE 118(3): 577-593.
http://dx.doi.org/10.1061/(ASCE)0733-9364(1992) 118:3(577)
Tanaka, Y. 1993. Cycle time simulation of shield-tunneling
operation, in Proc. of the 5th International Conference on Computing in
Civil and Building Engineering, ASCE, Reston, USA, 1386-1389.
Xu, X. 2001. The SIR method: a superiority and inferiority ranking
method for multiple criteria decision making, European Journal of
Operational Research 131(3): 587-602.
http://dx.doi.org/10.1016/S0377-2217(00)00101-6
Zadeh, L. A. 1975. The concept of a linguistic variable and its
application to approximate reasoning, Information Sciences, Part I 8:
199-249. http://dx.doi.org/10.1016/0020-0255(75)90036-5
Zhang, G.; Lu, J. 2003. An integrated group decision making method
dealing with fuzzy preferences for alternatives and individual judgments
for selection criteria, Group Decision and Negotiation 12(6): 501-515.
http://dx.doi.org/10.1023/B:GRUP.0000004197.04668.cf
Moatassem ABDALLAH (b), Mohamed MARZOUK (a)
(a) Department of Civil and Environmental Engineering, University
of Illinois at Urbana-Champaign (UIUC), Illinois, USA
(b) Structural Engineering Department, Faculty of Engineering,
Cairo University, Giza, Egypt
Received 23 Jul. 2011; accepted 8 Dec. 2011
Corresponding author: Mohamed Marzouk
E-mail: mm_marzouk@yahoo.com
Moatassem ABDALLAH. Research assistant at Civil and Environmental
Engineering, University of Illinois at Urbana Champaign, Illinois, USA.
He received his BSc and MSc in structural engineering from Cairo
University in 2005 and 2008, respectively. Currently, he is a PhD
candidate at Civil and Environmental Engineering, University of Illinois
at Urbana Champaign. His research interests include simulation, decision
support systems, construction technology, automation in construction,
life cycle costs, construction optimization, sustainability in
construction, and optimizing upgrade decisions for sustainable
buildings.
Mohamed MARZOUK. Associate Professor at Department of Structural
Engineering, Faculty of Engineering, Cairo University. He received his
BSc and MSc in Civil Engineering from Cairo University in 1995 and 1997,
respectively. He received his PhD form Concordia University in 2002. His
research interests include simulation and optimization of construction
processes, O-O simulation, fuzzy logic and its applications in
construction, risk analysis, and decision analysis.
Table 1. Developed simulation models for tunnels
construction
Model name Description
Cut & cover1 Tunnels construction by cut-and-cover
method using diaphragm walls
Cut & cover2 Tunnels construction by cut-and-cover
method using secant pile walls
Cut & cover3 Tunnels construction by cut-and-cover
method using soldier pile and lagging wall
Cut & cover4 Tunnels construction by cut-and-cover
method using steel sheet pile walls
TBM Segmental tunneling using slurry TBM
Table 2. Resources used in cut-and-cover method using
diaphragm walls model
Labor Concrete crew--formwork crew--reinforcement
crew--chiseler crew-survey
crew--insulation crew.
Equipment Excavator-loader-compactor-crane--concrete
pump--trench cutter--injecting
machine--well point machine.
Materials Guide wall form--top slab segment
form-bottom slab segment form--un-casted
top slab
segment form.
Table 3. Input parameter of cut-and-cover method using
diaphragm walls model
Parameter Description
NS Number of segments in tunnel zone
NGWPS Number of guide walls per segment
NDWPS Number of dewatering wells per segment
NDWPGW Number of diaphragm wall trenches per each
guide wall
NDW Total number of reinforced diaphragm wall
trenches (without cutting wall)
NDP Total number of dewatering well points
NIPPS Number of injecting points per segment
NGWPCW Number of guide walls in each cutting wall
NTSS Number of tunnel top slab segments
NBSS Number of tunnel bottom slab segments
NRTSS Number of un-casted tunnel top slab segments
A Lag between the third and fifth processes
(represented by number of segments)
B Lag between the fifth and sixth processes
(represented by number of segments)
Table 4. Task durations and resources of Giza tunnel
Process Activity Resources
Segment Segment survey Survey crew
preparation
Locating and Survey crew, pile
temporary casing drilling rig
installation
Drilling Pile drilling rig
Plastic concrete Crane, concrete
pouring crew
Construction of Locating and Survey crew, pile
secant pile wall temporary casing drilling rig
(one pile) installation
Drilling Pile drilling rig
Preparing and fixing RFT crew
of steel cage
positioning steel Crane,RFT crew
cage into hole
Concrete pouring Concrete mixer,
concrete crew
Construction of Excavation till the Excavators
one bottom slab base of bottom slab
segment
Pouring plain Concrete mixer,
concrete between concrete crew
the two sides of the
tunnel
Water insulation for Concrete crew
one segment
Steel work for one RFT crews
segment
Shuttering forms Formwork crews
and installing of
water stop (if exist)
Concrete pouring Concrete mixer, concrete
pump, concrete crew
Concrete curing -
Removal of forms Formwork crews
Zone II and IV
Process No. Duration
Segment 1 N(12, 2) h
preparation
1, 1 N (10, 2) m
1 U(15, 20) m
1, 1 u(20, 25) m
Construction of 1, 1 U(1.5, 2) h
secant pile wall
(one pile)
1, 1 U(15, 20) m
1 U(2.5, 4) h
1, 1 N(45, 10) m
1, 1 U(1.5, 2) h
Construction of 2 U(3, 4) d
one bottom slab
segment
1, 1 N(12, 1) h
1 U(4, 5) d
8 U(2, 3) d
4 N(12, 2) h
1, 1, 1 N(10, 2) h
0 N(12, 0) h
2 N(12, 2) h
Zone III
Process No. Duration
Segment 1 N(12, 2) h
preparation
1, 1 N(10, 2) m
1 U(15, 20) m
1, 1 U(20, 25) m
Construction of 1, 1 U(1.5, 2) h
secant pile wall
(one pile)
1, 1 U(15, 20) m
1 U (2.5, 4) h
1, 1 N(45, 10) m
1,1 U(1.5, 2) h
Construction of 2 U(3, 4) d
one bottom slab
segment
1, 1 N(12, 1) h
1 U(4, 5) d
8 U(2, 3) d
4 N(12, 2) h
1, 1, 1 N(10, 2) h
0 N(12, 0) h
2 N(12, 2) h
Zone I and V
Process No. Duration
Segment 1 N (12, 2) h
preparation
Construction of 0 N(0, 0)
secant pile wall
(one pile)
Construction of 1 N(1, 0.5) d
one bottom slab
segment
1, 1 N(6, 1) h
0 N(0, 0)
4 U(1, 1.5) d
2 N(1, 0) d
1, 1, 1 N(8, 2) h
0 N(1, 0) d
2 N(1, 0) d
U[a,b]: Uniform distribution; a is the lower value; b
is the higher value.
N[a,b]: Normal distribution; a is the mean; b is the standard
deviation. d, days; h, hours; m, minutes.
Table 5. Triangular fuzzy numbers for criteria ranges
Project cost and Resources' Triangular fuzzy
duration utilization number
Very high (VH) Very low (VL) (0,0,0.1)
High (H) Low (L) (0,0.1,0.3)
Medium high (MH) Medium low (ML) (0.1,0.3,0.5)
Medium (M) Medium (M) (0.3,0.5,0.7)
Medium low (ML) Medium high (MH) (0.5,0.7,0.9)
Low (L) High (H) (0.7,0.9,1)
Very low (VL) Very high (VH) (0.9,1,1)
Table 6. Criteria ranges for experts
Expert A
Lower limit (L.E.) Upper limit (L.E.)
24,000,000 25,000,000
25,000,000 26,000,000
26,000,000 27,000,000
Lower limit (days) Upper limit (days)
150 200
200 250
250 300
Lower limit (%) Upper limit (%)
10 20
20 40
40 50
Expert B
Lower limit (L.E.) Upper limit (L.E.)
22,000,000 24,000,000
24,000,000 26,000,000
26,000,000 28,000,000
Lower limit (days) Upper limit (days)
175 225
225 260
260 320
Lower limit (%) Upper limit (%)
5 25
25 45
45 65
Table 7. Experts' criteria levels
Expert A Expert B
Very important Rather important
Rather important Rather important
Unimportant Less important
Table 8. Criteria values for the three alternatives
Alternative I Alternative II Alternative III
26,230,395 L.E. 26,916,675 L.E. 25,197,130 L.E.
230 days 215 days 293 days
41.1% 35.7% 32%
Table 9. Criteria weights
Criterion Criterion Criterion Criterion
level scale weight
Expert Project cost Very 5 0.5
A ([W.sup.1.sub.1]) important
Project Rather 4 0.4
duration important
([W.sup.1.sub.2])
Resources Unimportant 1 0.1
utilization
([W.sup.1.sub.3])
Expert Project cost Rather 4 0.4
B ([W.sup.2.sub.1]) important
Project Rather 4 0.4
duration important
([W.sup.2.sub.2])
Resources Less 2 0.2
utilization important
([W.sup.2.sub.3])
[W.sup.j.sub.i], weight for criterion "i'," given by expert "j."
Table 10. Values of belief vectors
Belief vector Value
[[bar.b].sup.1.sub.1] (0.22, 0.42, 0.62)
[[bar.b].sup.1.sub.2] (0.17, 0.27, 0.42)
[[bar.b].sup.1.sub.3] (0.28, 0.4, 0.56)
[[bar.b].sup.2.sub.1] (0.34, 0.54, 0.74)
[[bar.b].sup.2.sub.2] (0.26, 0.42, 0.62)
[[bar.b].sup.2.sub.3] (0.18, 0.34, 0.54)
[[bar.b].sup.i.sub.j], belief vector for criterion "j"
and decision maker "i".
Table 11. Experts' years of experience weight
Years of Expert Expert
Expert experience scale weight
First Expert ([v.sup.*.sub.1]) 18 4 0.67
Second Expert ([v.sup.*.sub.2]) 8 2 0.33
Table 12. Calculated values of fuzzy decision vectors
Decision vector Value
[[??].sub.1] (0.26, 0.46, 0.66)
[[??].sub.2] (0.20, 0.32, 0.49)
[[??].sub.3] (0.25, 0.38, 0.55)
Table 13. Values of solution distances
Belief Positive
vector distances
[d.sup.+.sub.1] d ([[??].sub.1], [r.sup.+]) =0.564
[d.sup.+.sub.2] d ([[??].sub.2], [r.sup.+]) =0.675
[d.sup.+.sub.2] d ([[??].sub.3], [r.sup.+]) =0.62
Belief Negative
vector distances
[d.sup.-.sub.1] d ([[??].sub.1], [r.sup.-]) =0.488
[d.sup.-.sub.2] d ([[??].sub.2], [r.sup.-]) =0.356
[d.sup.-.sub.3] d ([[??].sub.3], [r.sup.-]) =0.412
Table 14. Calculated values of closeness coefficients
Closeness coefficient Value
[CC.sub.1] 0.462
[CC.sub.2] 0.34
[CC.sub.3] 0.397
