Vous êtes sur la page 1sur 13

Computer-Aided Civil and Infrastructure Engineering 31 (2016) 717729

Multicriteria Evaluation of Building


Foundation Alternatives

Zenonas Turskis, Alfonsas Daniunas, Edmundas Kazimieras Zavadskas* & Jurgis Medzvieckas
Research Institute of Smart Building Technologies, Vilnius Gediminas Technical University, Vilnius LT-10223, Lithuania

Abstract: In real conditions, decision makers usually in a much more responsible way, carefully considering
deal with multiple objectives and should make a deci- and weighing all the details. Everyone wants to have a
sion in a state of certainty or uncertainty. The selection quality and durable house at the lowest possible cost.
of the best constructions for a building from a number About one-third of Europeans live in individual houses.
of alternatives is of great importance for owners, con- A lot of works and materials to design and construct
tractors, and stakeholders. Dozens of multicriteria/multi- such houses is consumed. Nevertheless, scientists and
attribute decision-making (MCDM/MADM) models de- researchers pay little attention to examine and investi-
veloped for evaluating the performance of the available gate simple problems. This occurs because they believe
alternatives can be used for selecting the most suitable that implementation of simple construction projects is
alternative from a given finite set of options based on a very easy, and not worth any more attention. In their
set of attributes. A guide to systematic selection among point of view, it is better to focus on optimization of
the available alternatives of building structures is the in- a complex and large project. However, when a simple
tegrated methodology, thoroughly analyzed in the article. problem is solved and is being implemented hundreds of
The article presents a MCDM model for selecting the type thousands of times, the problems solution brings much
of foundation for a single-storey dwelling house based more benefit than an optimization of a single complex
on the WASPAS-G (Weighted Aggregated Sum Prod- project. In fact, only expenditures will be determined
uct Assessment) method and Analytic Hierarchy Pro- bigger depending on the building size, whereas ranking
cess (AHP) approach. The aggregate criteria weights are of possible discrete alternatives will remain the same or
determined by using the AHP and experts judgement similar. The construction scheme of an economic single-
methods. storey dwelling house is usually simple. It consists of
masonry (brick, ceramic, etc.) walls installed on shallow
1 INTRODUCTION foundation or piles. It is clear that engineering and tech-
nological projects are also business projects. They in-
The increasing competition and stakeholders demands, volve such aspects associated with business as customer
as well as the scale of production and consumption satisfaction, resource utilization, deadlines, costs, prof-
cause a lot of construction problems. In the report of its, etc. The variables, such as the environment, terrain,
European Foundation for the Improvement of Living soil conditions, zoning laws, labor market, and local
and Working Conditions (Eurofound, 2013), the main utilities, make every project unique. This requires doing
factor, reflecting a negative effect on life satisfaction of something different from what has been done before.
people which is associated with housing, is the feeling One of the most important activities in management is
of insecurity in their houses. In the European countries, the performance measurement-based decision making
living in cities and medium-to-large towns has been (Li et al., 2011; Zavadskas et al., 2012c). A great ma-
associated with some loss of life satisfaction compared jority of numerous papers published on optimization of
to living in the countryside. This has been caused by structures deal with multicriteria optimization (Sarma
housing problems. Now, people build or buy houses and Adeli, 2000). The total structural optimization of
the building material cost involves many components,
Towhom correspondence should be addressed. E-mail: edmundas. including foundation, slabs, beams, columns, bracings,
zavadskas@vgtu.lt.


C 2016 Computer-Aided Civil and Infrastructure Engineering.

DOI: 10.1111/mice.12202
718 Turskis, Daniunas, Zavadskas & Medzvieckas

Fig. 1. The decision-making scheme of foundations.

a
4150 d

50 kN/m
120 kN/m b
60 kN/m
80 kN/m

180 kN/m

350 kN b
120 kN/m 50 kN/m

b
65 kN/m 350 kN 110 kN/m
160 kN/m

c
160 kN/m 180 kN/m

40 kN/m
50 kN/m
b
40 kN/m 65 kN/m

Fig. 2. The plan of walls with the applied loads and foundation alternatives: a shows shallow foundations deepened to 1.0 and
1.5 m; b denotes 0.5 and 0.6 m diameter short bored piles; c denotes 0.3 m diameter bored piles.

belt or outrigger trusses, and shear walls (Aldwaik and capabilities (Oloufa and Ikeda, 1995; Lepre et al., 2010;
Adeli, 2014). A foundation is a very important element Tezcan and Ozdemir, 2012). Usually, the installation
of a building, which does not only ensure the stability cost, materials, the excavated soil, the equipment used,
of building structures, but also accounts for a significant etc. determine the foundation type selection. The
part of the total house construction cost. In many cases, model predicts the expected cost and the duration of
the idea of selection is based on subjective preferences the project, depending on the input parameters, such
and scope of knowledge of individuals. The choice as the production rate, the scope of the work, time
is objective and seldom systematic. The foundation schedule, bonding conditions, maximum and minimum
types are usually chosen based on the designers and deviations from the volume of work. Clients are able
builders opinions and experience, regardless of the to make proper decisions concerning the time and cost
specific engineering-geological conditions and features schedules of their investments (Beran and Hromada,
of the house structure (Mangushev, 2004; Susinskas 2008). One of the ways to find the most suitable so-
et al., 2014). When selecting a type of foundation, it is lution is associated with multicriteria decision-making
necessary to take into account engineering-geological or multi-attribute decision-making methods (MCDM
conditions, climate impact, ground water levels, specific or MADM) (Hwang and Yoon, 1981; Ishizaka and
features of the structure, and builders technological Nemery, 2013; Xu, 2015). The procedure of selecting
Multicriteria evaluation of building foundation alternatives 719

the best alternative of the foundation type is given in Table 1


Figure 1. Densities and geotechnical properties of sands

Soil properties

2 THE SELECTION OF THE FOUNDATION FOR qc


A SINGLE-STOREY DWELLING HOUSE Type Soil (kN/m3 ) (MPa)
1 Loose sand (LS) 17.50 3132 3.05.0
The foundation alternatives for a single-storey dwelling 2 Average density sand (AS) 18.50 3436 6.010.0
house were selected for multiple criteria evaluation. 3 Dense sand (DS) 19.50 3739 12.020.0
The foundations were designed according to a standard
single-storey house project. The plan of the loads on the
building wall foundations and their alternatives are pre-
sented in Figure 2.
Rationality of construction works and the appropri- civil engineering practice, the compressed sand cushions
ate choice of technologies make the most challenging (Usmanov, 2013) and soil reinforcement are used in
tasks in real life (Zavadskas et al., 2012a). The de- loose and high compressibility soils (Kolay et al., 2013).
sign and construction of foundations require the pro- Foundations deepened to 1.0 and 1.5 m are designed for
found knowledge of the mechanical behavior of soils various types of sandy ground. The expenditure of ma-
and their spatial variability (Sivilevicius et al., 2012). terials for shallow foundations is presented in Table 2.
Sandy ground having different geotechnical properties Short bored piles are also often used as building foun-
( is the unit weight; is the internal friction angle; qc dations. The top layer of the soil throughout most parts
is the cone resistance) is one of the most common con- of Lithuania consists of quaternary sediments which are
struction sites. Sandy ground has been chosen for eval- shaped by the receding ice shield. Naturally formed
uating the foundation alternatives. Geotechnical prop- glacial sand and clay derivatives cover 79% of the coun-
erties of sandy ground are presented in Table 1. Sands trys territory (Dundulis, 1997). These derivatives are
of various densities cover 32% of Lithuanian territory relatively strong and suitable for the foundation base.
(Dundulis, 1997), therefore, the foundations are often Short piles are easy to install by using the effective
installed in these soils. drilling machines. The analysis of piles of various forms
In most cases for one-storey houses, standard shal- has shown that maximum efficiency was achieved un-
low foundations are constructed. The city plan can show der the minimum depth of 2 meters, and gradually de-
the areas whose potential for urban development leads creased later (Bulatov and Kolosova, 2013).
to occupation suitable for the construction of shallow Researchers have shown that quantitative values of
foundations in one-storey residential buildings (Lorandi pile efficiency are approximately the same for cross-
and Mendes, 2010). Shallow foundations can be mono- sections of various forms. The increase in the bearing
lithic or prefabricated. The reinforcement amount for capacity of piles is now often achieved by special treat-
shallow foundations of one-storey building is negligi- ment of pile bases after concrete placement, as well
ble. The bottom of foundation is placed on the lev- as by increasing the pile length and diameter. The re-
elled ground. The bottom part may be wider and it dis- sults obtained show that the postgrouting to the shaft
tributes the load to the base. The walls or foundation and tip of the pile can increase bearing capacity of de-
blocks are installed on pads. Wall units can be placed fected bored piles two times and by about 20% com-
directly on the levelled ground if the base bearing ca- pared to that of the normal bored piles (Nguyen et al.,
pacity is sufficient to withstand the loads. The calcu- 2013). Short piles are rigid, and when the transverse
lations of base bearing capacity are regulated by LST forces and/or bending moments are acting, the foun-
EN 1997-1 (2006). The foundation installation technol- dation does not deform. This is their main difference
ogy is simple. A trench is excavated to the foundation from the conventional bored piles. Compared to shal-
pad level on the site, then, concrete monolithic founda- low foundations, short piles have many advantages, in-
tions or prefabricated foundations are laid. The width cluding more effective use of materials and sub-grade
of the pit should be sufficient to smooth out and com- because the load on the base is transmitted not only by
pact the soil before installing the foundation. When the the tip, but also by the lateral surface. The settlement
foundations are installed, the excavation is filled with of these foundations is lower. Moreover, the extrac-
soil which is compacted layer after layer. A base of the tive work during installation is reduced because only
foundation can be strengthened to improve its proper- the ground required for the foundation is excavated.
ties. This reduces the amount of the required material, The installation process is quick: the holes are drilled
but not necessarily the foundation installation cost. In on the site and filled with concrete after inserting the
720 Turskis, Daniunas, Zavadskas & Medzvieckas

Table 2
Expenditure of materials for shallow foundations

Depth, Concrete Reinforcement Excavated soil


Alternatives Soil d (m) (m3 ) kG (m3 )
A1 LS 1.0 37.939.2 232.1257.3 205.3216.8
AS 36.237.3 199.9226.5 189.7200.4
DS 35.335.8 190.7196.3 178.4185.7
A2 LS 1.5 48.048.9 201.9225.6 314.0328.4
AS 46.947.4 190.7196.3 294.3306.2
DS 46.446.5 186.2187.1 279.3287.9

reinforcement. On the top of bored foundations, a re- Table 3


inforced beam is installed. Bored foundations are de- Expenditure of materials for pile alternatives
signed in accordance with the National Building Code
Expenditure of materials for 0.3 m diameter bored piles
RSN 91-85 (1987). The design strength of the ground
is defined by the cone resistance qc determined by the Pile Pile cap
cone penetration test (CPT). In DS 0.6 m diameter piles
are obtained too short and cannot be used. Reinforce- Reinforce-
For a single-storey dwelling house, the cavity is more Concrete ment Concrete ment
often formed during drilling. Soil is raised to the surface. Soil (m3 ) kG (m3 ) kG
In weak and water-saturated soils, the protective tube A3 AS 9.618.3 207.0367.2 14.0 727.6
is used in drilling. Bored pile installation embraces the DS 5.28.1 127.1177.5
procedures of preparing the borehole cavity, inserting
the reinforcement, and removing the protective tube. Expenditure of materials for 0.5 m diameter bored piles
At present, continuous flight auger piles known as CFA
Pile Pile cap
piles are often installed. This technology ensures fast
and reliable installation of piles (Zayed, 2005). The rein- Reinforce- Reinforce-
forcement is placed after placing the concrete. In weak Concrete ment Concrete ment
soil, displacement piles are usually installed (Kelevisius Soil (m3 ) kG (m3 ) kG
et al., 2014). Using a vibrator, an inventory tube with a
A4 LS 29.953.0 226.1378.0 14.0 727.6
cap at the end is pressed into soil. Tube is pulled out AS 14.823.4 127.8184.0
after it filled up with concrete. A reinforced beam is DS 12.8 113.4
installed on the piles. Pile foundations are designed by
LST EN 1997-1 (2006) and LST EN 1997-2 (2009) rec- Expenditure of materials for 0.6 m diameter bored piles
ommendations. Piles of 0.3 m in diameter are designed
for the selected type of soil. LS is assumed to be inap- Pile Pile cap
propriate for pile installation. The expenditure of mate-
Reinforce- Reinforce-
rials for piles is presented in Table 3.
Concrete ment Concrete ment
Soil (m3 ) kG (m3 ) kG
A6 LS 34.159.0 186.1301.3 14.0 727.6
3 MULTI-ATTRIBUTE EVALUATION OF AS 19.228.0 118.8162.4
FOUNDATION ALTERNATIVES

MADM (also called multi-objective decision making


with finite alternatives) is an important component Benjamin Franklin (MacCrimmon, 1973) in 1772 pro-
of modern operations research science. Essentially, posed a procedure and rules for decision making, pro-
MADM involves making preference decisions over the viding moral or prudential algebra for coordinating
available set of alternatives that are characterized by multiple attributes in everyday decisions, and suggested
multiple, usually conflicting, attributes. The problem of a workable solution. There are lots of MCDM methods
making the appropriate decisions by individuals and that are suitable for multi-attribute problems solution
institutions was discussed by Aristotle (1990). Even (Jato-Espino et al., 2014).
Multicriteria evaluation of building foundation alternatives 721

Most of multi-attribute decision-making methods are three types of grey model construction, which
share a common number of steps as follows: are as follows (Liu and Lin, 2006, 2010):

1. Grey rules of situation: Grey rules of situation deal


1. Define the relevant goals;
with strategy making based on multi-objects that
2. Determine the alternatives to be evaluated;
are contradictory in an ordinary way.
3. Identify the relevant attributes for evaluating the
2. Grey decision-making group: This field includes
alternatives;
grey relational space, grey statistics, grey cluster-
4. Determine weights (relative importance) of the
ing, and grey prediction.
attributes;
3. Grey programming.
5. Score each attribute of each alternative;
6. Select a scoring method and calculate the multi- In the following four general cases, system informa-
attribute utilities of various options; and tion manifests itself as incomplete or grey information
7. Perform postevaluation analysis. (Fang et al., 2009):

In real life, decision makers are usually faced with 1. Incomplete information of elements (parameters);
multiple objectives and need to make a decision in a 2. Incomplete structural information;
state of uncertainty (Xu and Yao, 2011). Grey systems 3. Incomplete information of boundaries;
theory, probability and statistics, and fuzzy mathemat- 4. Incomplete information of systems behavior.
ics are the three most frequently used methods deal-
ing with unascertained problems. Due to the increasing There are dozens of studies and problem solution
complexity and uncertainty of objects and the fuzziness models dealing with the application of the grey system
of human thought, much more attention has been paid theory in construction. For example, Zavadskas et al.
to the investigation of MADM under uncertain environ- (2008) presented the selection of the effective dwelling
ments, and fruitful research results have been achieved house walls based on grey attribute values (the main
over the last decades (Xu, 2015). As people were getting ideas of a new COPRAS-G method were presented).
more knowledge about uncertain events, the methods of Tserng et al. (2015) integrated the grey system theory
research into the problems of random multi-objective with all available firm-year samples during the sample
decision making, fuzzy multi-objective decision making, period and provided a new method for predicting the
rough multi-objective decision making, and two-fold probability of the construction firm default.
uncertain multi-objective decision making were gradu- The Weighted Aggregated Sum Product (WASPAS)
ally developed (Xu and Zhou, 2011). A great number of method was originally described and presented by
researchers have been applying intuitionistic fuzzy sets Zavadskas et al. (2012c). There is a limited number
to multi-attribute decision making in various situations of the WASPAS method with crisp values applications
(Xu, 2014). to solving the construction problems. Recently, An-
Furthermore, global optimization problems also in- tucheviciene et al. (2015) have developed and presented
volve multiple criteria. Cheng et al. (2015) stated that the WASPAS-G method.
integrating several strategies into one algorithm and
determining the application rate of each strategy as
3.1 Operations with grey numbers
well as the associated parameter values online had be-
come an ad-hoc research topic. They proposed a novel Let a grey number G be defined by two parameters
differential evolution (DE) algorithm related to mul- and , which are called the lower and upper limits
ticriteria adaptive DE (MADE) for global numerical of a grey number (interval number) G = [G , G ] =
optimization. {x|G x G , G and G R}. Let +, , , and
Arce et al. (2015) pointed out that one of the prob- denote the operations of addition, subtraction, multipli-
lems with analyzing a MADM problem is a lack of cation, and division, respectively. According to the in-
data, especially, in empirical works. To address this terval of confidence (Kaufmann and Gupta, 1985), some
constraint, grey relational methods could be used in basic operations with positive grey numbers G 1 and
research works. The grey system theory and method G 2 can be expressed as follows:
applicable to the study of unascertained problems with
few data and (or) poor information were first proposed G 1 + G 2 = (G 1 + G 2 , G 1 ) + (G 2 ) Addition (1)
by Deng (1982). Grey-related analysis is a technique
that can be applied to both fuzzy and crisp data. There G 1 G 2 = (G 1 G 2 , G 1 G 2 ) Subtraction (2)
722 Turskis, Daniunas, Zavadskas & Medzvieckas

G 1 G 2 = (G 1 G 2 , G 1 G 2 ) Multiplication (3) ues xi j of the normalized decision-making matrix


X :
 
G 1 G 1
G 1 G 2 = , , only if G 1 , G 2 , G 1 , x01 x0 j x0n
G 2 G 2 .. .. .. .. ..
. . . . .
and G 2 does not contain 0 division (4)
X =
xi1 xi j xin

. .. .. .. .. (8)
.. . . . .
xm1 xm j xmn
k (G 1 ) = (kG 1 , kG 1 ) Grey number multiplied
by positive real number k (5) i = 0, m; j = 1, n

The attributes, whose preferable values are maximal,


  are normalized as follows:
1 1
(G 1 )1 = , ; (G 1 )k = ((G 1 )k , (G 1 )k )
G 1 G 1 xi j
xij = max xi j
k
if G 1 and G 1 1 , and k  1 ; (G 1 ) = (1, 1) , i
(9)
xi j xi j
G 1 if and G 1 1 and k = 0 . Exponentiation by a or xija = max xi j
and xi jb = max xi j
i i

natural power k (6)


The attributes, whose preferable values are minimal,
are normalized as follows:

min xi j
i
3.2 Grey extension of WASPAS method: WASPAS-G xij =
xi j
method
(10)
The Weighted Sum Model (WSM) is one of the best min xi j min xi j
i i
known and commonly used MADM methods for evalu- or xij = and xi j =
ating the number of alternatives in terms of the number xi j xi j
of decision attributes. It is based on the assumption that
the phenomena of the complicated world could be un- When the dimensionless values of the attributes are
derstood based on simple comparisons. known, all the attributes, originally having different di-
The first stage is associated with the formation of grey mensions, can be compared.
decision-making matrix (GDMM). In the GMADM of a Usually, the performance values xi j and the at-
discrete optimization problem any problem to be solved tribute weights w j (denotes the relative significance
is represented by the following DMM of preferences for of the attribute) are viewed as the entries of a DMM.
m reasonable alternatives (rows) rated on n attributes The third stage is defining normalized-weighted
(columns) describing each alternative as follows: matrix - X . It is possible to evaluate the attributes
with weights 0 < w j < 1. Only well-founded weights

x01 x0 j x0n should be used because weights are always subjective
.. .. .. .. .. and influence the solution. The values of weight wj are
. . . . .
usually determined by the expert evaluation method.

X = xi1 xi j xin The sum of weights w j would be limited as follows:
. .. .. .. .. (7)
.. . . . .
xm1 xm j xmn

n
wj = 1 (11)
j=1
i = 0, m; j = 1, n

xi j = (xi j , xi j ) is a grey value representing the The values of all the normalized attributes are
performance of the ith alternative in terms of the jth at- weighted by defining the values xi j of the normalized-
tribute. weighted decision-making matrix X :
At the second stage, the initial values of all
the attributes are normalized by defining the val- xi j = xi j w j ; i = 0, m (12)
Multicriteria evaluation of building foundation alternatives 723

Table 4
Foundation alternatives considered for LS

Criteria

Concrete Excavated Duration of Installation Reinforcement Mechanisms


Alternatives Depth (m) (m3 ) soil (m3 ) works (hours) cost () (kg) (hours)
x1 x2 x3 x4 x5 x6


Weight (w) 0.1304 0.2376 0.0652 0.1699 0.0539 0.1330 0.2464 0.4388 0.1823 0.2568 0.0303 0.0652
w (without x4 ) 0.1958 0.3092 0.0447 0.2055 0.0789 0.3603 0 0 0.1983 0.4444 0.0520 0.1983
A1 1.0 37.9 39.2 205.3 216.8 249.3 257.9 7584.8 7845.0 232.10 257.30 28.9 29.9
A2 1.5 48.0 48.9 314.0 328.4 315.8 321.7 9606.1 9786.2 201.90 225.60 32.4 33.1
Diameter (m)
A4 0.5 43.9 67.0 52.7 80.4 252.4 385.3 12638.0 19288.1 953.7 1105.6 65.9 100.5
A5 0.6 48.1 73.0 57.7 87.6 276.6 419.8 13847.1 21015.4 913.7 1028.9 72.2 109.5

Table 5
Foundation alternatives considered for AS

Criteria

Concrete Excavated Duration of Installation Reinforcement Mechanisms


Alternatives Depth (m) (m3 ) soil (m3 ) works (hours) cost () (kg) (hours)
x1 x2 x3 x4 x5 x6


Weight (w) 0.1304 0.2376 0.0652 0.1699 0.0539 0.1330 0.2464 0.4388 0.1823 0.2568 0.0303 0.0652
w (without x4 ) 0.1958 0.3092 0.0447 0.2055 0.0789 0.3603 0 0 0.1983 0.4444 0.0520 0.1983
A1 1.0 36.2 37.3 189.7 200.4 238.2 245.4 7244.3 7464.5 199.90 226.50 27.6 28.4
A2 1.5 46.9 47.4 294.3 306.2 308.6 311.8 9385.6 9485.7 190.70 196.30 35.8 36.1
Diameter (m)
A3 0.3 23.6 32.3 28.32 38.76 135.7 185.7 6794.0 9298.6 934.6 1094.8 35.4 48.5
A4 0.5 28.8 37.4 34.56 44.88 165.6 215.1 8291.0 10766.8 855.4 911.6 43.2 56.1
A5 0.6 33.2 42 39.84 50.4 190.9 241.5 9567.3 12103.2 846.4 890 83.0 105.0

where w j is the weight (significance) of the jth attribute where Q i is the value of the additive optimality func-
and xi j is the normalized value of the jth attribute. tion for ith alternative.
According to the Weighted Product Model (WPM),
x01 x0 j x0n the total relative importance of the alternative i denoted
.. .. .. .. ..
. . . . . by Pi is defined as follows:

X = x i1 x ij x
in
n
. (13) w j
.. . .. .
.. . .. .
.. Pi = xi j , or
j=1
xm1 xm j xmn

n
w w
i = 0, m; j = 1, n Pi = 0.5 xi j j + xi j j (15)
j=1
The following task is the determining of the optimal-
ity function values: An attempt was made to use an integrated multi-
attribute utility value for determining the total perfor-

n
mance of the alternative Kt (Zavadskas et al., 2012b)
Qi = xi j , or
j=1
according to the WASPAS method, assuming equal
contribution of WSM ( Qi ) and WPM ( Pi ) to the

n
total score:
Q i = 0.5 (xi j + xi j ) , i = 0, m (14)
j=1 K i = 0.5 Q i + 0.5 Pi (16)
724 Turskis, Daniunas, Zavadskas & Medzvieckas

Table 6
Foundation alternatives considered for DS

Criteria

Concrete Excavated Duration of Installation Reinforcement Mechanisms


Alternatives Depth (m) (m3 ) soil (m3 ) works (hours) cost () (kg) (hours)
x1 x2 x3 x4 x5 x6


Weight (w) 0.1304 0.2376 0.0652 0.1699 0.0539 0.1330 0.2464 0.4388 0.1823 0.2568 0.0303 0.0652
w (without x4 ) 0.1958 0.3092 0.0447 0.2055 0.0789 0.3603 0 0 0.1983 0.4444 0.0520 0.1983
A1 1.0 35.3 35.8 178.4 185.7 232.2 235.5 7064.2 7164.3 190.70 196.30 26.9 27.3
A2 1.5 46.4 46.5 279.3 287.9 305.2 305.9 9285.6 9305.6 186.20 187.10 35.4 35.5
Diameter (m)
A3 0.3 19.2 22.1 23.04 26.52 110.4 127.1 5527.3 6362.2 854.7 905.1 28.8 33.2
A4 0.5 26.8 26.8 32.16 32.16 154.1 154.1 7715.2 7715.2 841 841 40.2 40.2

Based on the previous research and supposing the in-


crease of ranking accuracy, as well as the effectiveness
of decision making, it was offered to determine K i as
follows (Zavadskas et al., 2012b):

n
n
w j
Ki = xi j + (1 ) xi j
j=1 j=1

= 0, . . . , 1 (17)

The value could be determined in different ways.


In this case, is determined as a lever based on the as-
sumption that the total WSM performance score of all
alternatives should be equal to the total of WPM per-
formance score:
m
i=1 Pi
= m m (18)
i=1 Q i + i=1 Pi

Archimedes explained levers mechanical advantage,


or leverage, in his work On the Equilibrium of Planes
and is noted for his claim Give me a place to stand
and a long enough lever and I can move the Earth.
Archimedes explanation of the theory of the lever
is based on the principle of balancing the input and
output torques about the fulcrum of the device so
that, the input force multiplied by its distance from
the fulcrum, is equal to the weight (or downward
force) of the load multiplied by its distance from the
fulcrum.
Fig. 3. Determining of criteria weights.
The calculated values K i are in the interval [0; 1] and
can be arranged in the ascending order, which is the
wanted order of precedence. teria describing the foundation installation are calcu-
lated using standard tariffs. According to the load pre-
sented in Figure 2, five alternatives of foundations are
3.3 Problems solution
possible: A1 is a shallow foundation deepened to 1.0 m;
The foundation selection problem is solved by apply- A2 is a shallow foundation deepened to 1.5 m; A3 de-
ing the above described model. The values of the cri- notes the 0.3 m diameter bored piles; A4 denotes a 0.5 m
Multicriteria evaluation of building foundation alternatives 725

Fig. 4. Entering input data into MCDSS WEAR for one expert, using the Analytic Hierarchy Process (AHP) method.

Table 7
Weight coefficients based on experts judgement, taking into account the installation cost

diameter bored short pile; A5 is a 0.6 m diameter bored the installation of foundations in LS, AS, and DS are
short pile. The alternative A3 , representing the 0.3 m di- presented in Tables 46. Because the cost of materials,
ameter bored piles, was considered to be inappropriate mechanisms, and work duration make the basic cost, the
for LS. For DS, the alternative A5 , representing a 0.6 alternatives given in Tables 46 were calculated with-
m diameter bored pile, is too short and, therefore, can- out taking into account the installation cost criterion
not be treated as a short pile. Therefore, it was not con- (weight of x4 is equal to 0).
sidered as an alternative. Six criteria were selected for Expert systems have been employed for solving var-
evaluating foundations: x1 denotes the amount of con- ious complicated problems and providing advice (Lee
crete; x2 is the excavated soil; x3 denotes the duration et al., 2011). These solutions have been adopted to man-
of works; x4 denotes the installation cost; x5 denotes the age civil engineering problems (Ahmadlou and Adeli,
reinforcement amount; x6 denotes the working time of 2010; Reuter and Moller, 2010; Hasanzadehshooiili
mechanisms. The considered alternatives representing et al., 2012).
726 Turskis, Daniunas, Zavadskas & Medzvieckas

Table 8
Weight coefficients based on experts judgement, not taking into account the installation cost

Table 9
The integrated ranks of foundation alternatives obtained using WASPAS-G method.

Type of sand

Loose Average density Dense

Expenditures are included

R R R
Alternatives Q P K (Rank) Q P K (Rank) Q P K (Rank)
A1 0.873 0.778 0.821 1 0.753 0.591 0.649 1 0.206 0.564 0.390 4
A2 0.758 0.688 0.720 2 0.645 0.511 0.559 2 0.682 0.465 0.570 3
A3 0.747 0.480 0.575 3 0.578 0.576 0.577 2
A4 0.572 0.421 0.489 3 0.634 0.453 0.517 4 0.760 0.489 0.620 1
A5 0.528 0.403 0.460 4 0.550 0.394 0.449 5
= 2.731 2.290 = 0.456 3.329 1.838 = 0.356 2.226 2.095 = 0.485
Expenditures are not included
A1 0.895 0.803 0.844 1 0.726 0.661 0.687 1 0.197 0.596 0.393 4
A2 0.814 0.682 0.741 2 0.650 0.547 0.588 4 0.686 0.480 0.585 3
A3 0.709 0.584 0.634 2 0.609 0.660 0.634 1=2
A4 0.626 0.446 0.526 3 0.606 0.530 0.560 3 0.712 0.522 0.634 1=2
A5 0.581 0.423 0.494 4 0.512 0.471 0.487 5
= 2.915 2.353 = 0.447 3.203 2.131 = 0.400 2.203 2.288 = 0.509
Final ranks

A1 1 1 4
A2 2 3 3
A3 2 2
A4 3 4 1
A5 4 5

Not all the criteria are of equal importance. The role three main approaches to determining the relative sig-
of the weights is to reflect relative importance of each of nificance of attributes (Hwang and Yoon, 1981):
the attributes with respect to other attributes. In prac-
tice, it is often difficult to determine the relative im-
portance of the criteria. Different theories of criteria 1. The subjective approach: Delphi Method, The
weight determination have been developed. There are Analytic Hierarchy Process (AHP method),
Multicriteria evaluation of building foundation alternatives 727

Additive Normalization Method, Expert Judge- 4 CONCLUSION


ment Method, SWARA method, etc.
2. The objective approach (which deals only with The suggested problem solution model offers higher
the criteria values in the defined set of alter- prediction accuracy, as well as simpler and clearer pro-
natives and criteria, but does not show the real cedures for analyzing building foundation alternatives
importance of criteria and preferences of decision and ranking the attributes in terms of their significance
maker): Entropy method (generates the criteria level. The contrast between the results show the ap-
weights directly from the value of the criterion plicability of this model as a valuable tool to be used
for each alternative and eliminates the problem by construction work managers, as well as for decision-
of subjectivity or lack of decision makers); sta- making activities and performance evaluation, which
tistical methods (CRITIC CRiteria Importance contains multiple alternatives and multiple criteria. Ap-
Through Intercriteria Correlation: determination plying the MCDM is automated selection from several
of objective weights of criteria based on the level alternatives. MCDM methods provide a well-structured
of contrast and conflict observed in the structure procedure (sequence of steps) to select the type of struc-
of the decision problem; Chi-square test: calcu- ture. The algorithm can be easily modified to include
lating whether there is a statistically significant new variables, changing their weights, depending on the
correlation between the frequencies of the two interested persons and the environment in which deci-
attribute characteristics or between the observed sions are made.
frequencies and the frequencies expected in a The suggested model includes only six ranked vari-
particular hypothesis). ables (attributes). As a result, practitioners may directly
3. The integrated method, which combines the two use the model as a means of fast and convenient analy-
previous approaches. sis of building foundation alternatives by applying sim-
ple and clearly defined procedures. Another important
Seven experts (E1 . . . E7 ) assessed the importance of managerial implication is that the obtained data will
the criteria. The criteria weights were determined ac- vary, depending on a particular case.
cording to the algorithm presented in Figure 3. Researchers and stakeholders have found a way to
For that purpose, the modified software MCDSS evaluate and rank the alternatives by applying grey val-
WEAR was used (Zavadskas et al., 2012c). Weight co- ues and to compare their scores with that of the ideal
efficients (Figure 4) determined when the installation alternative.
costs were assessed are presented in Table 7 and those Weighting results show that stakeholders are more
obtained when these costs were not evaluated are pre- concerned with the installation cost and reinforcement
sented in Table 8. They were determined by using the than with the consumption of concrete, mechanisms, or
AHP (Analytic Hierarchy Process) approach (Saaty, duration of works.
1980). Finally, it should be noted that the WASPAS-G
Multicriteria evaluation was performed using the method has a promising future in the decision mak-
WASPAS-G method. The ranks assigned in the evalua- ing because it offers a methodological basis for decision
tion of expenditures for the case of LS, for AS, and for support.
DS are shown in Table 9. Besides this, in Table 9, the
integrated evaluation of alternatives is presented. Cal- REFERENCES
culations made with and without the installation cost
assessment give similar results. For LS, the alternative Ahmadlou, M. & Adeli, H. (2010), Enhanced probabilistic
A1 , representing shallow foundations deepened to 1.0 neural network with local decision circles: a robust clas-
m is best suited, whereas the alternative A2 , represent- sifier, Integrated Computer-Aided Engineering, 17(3), 197
ing shallow foundations deepened to 1.5 m is ranked 210.
second. Aldwaik, M. & Adeli, H. (2014), Advances in optimization
of highrise building structures, Structural and Multidisci-
For the average density sand, the alternative A3 , plinary Optimization, 50, 899910.
representing 0.3 m diameter bored piles, is the best, Antucheviciene, J., Zavadskas, E. K. & Turskis, Z. (2015),
whereas the alternative A1 , representing shallow foun- Evaluating and selecting a contractor for a construction
dations deepened to 1.0 m, is ranked second. For DS, company by using novel method for multi-attribute anal-
the alternative A4 , representing 0.5 m diameter bored ysis: Weighted Aggregated Sum Product ASsessment with
Grey values (WASPAS-G) method, Studies in Informatics
pile foundations is the best, whereas the alternative and Control, 24(2), 14150.
A3 , representing 0.3 m diameter bored piles, is ranked Arce, M. E., Saavedra, A., Mguez J. L. & Granada, E.
second. (2015), The use of grey-based methods in multi-criteria
728 Turskis, Daniunas, Zavadskas & Medzvieckas

decision analysis for the evaluation of sustainable energy Liu, S. & Lin, Y. (2006), Grey Information: Theory and Prac-
systems: a review, Renewable and Sustainable Energy Re- tical Applications, Springer-Verlag, London.
views, 47(2015), 92432. Liu, S. & Lin, Y. (2010), Grey Systems: Theory and Applica-
Aristotle (1990), Ethica Nicomachea, Oxford University tions, Springer-Verlag, Berlin, Heidelberg.
Press, Oxford. Originally published in 350 BC, English edi- Lorandi, R. & Mendes, R.M. (2010), Geospatial analysis of
tion by I. Bywater. geotechnical data applied to urban infrastructure planning,
Beran, V. & Hromada, E. (2008), Dynamic simulations in cost Journal of Geographic Information System, 02(01), 2331.
and time estimation of the construction process, Acta Poly- LST EN 1997-1 (2006), Eurokodas 7. Geotechninis projektavi-
technica, 48(1), 3035. mas: 1 dalis. Pagrindines taisykles. (Eurocode 7. Geotechni-
Bulatov, G. Y. & Kolosova, N. B. (2013), Efficiency of piles of cal design Part 1: General rules), Lietuvos standartizacijos
various cross-sectional forms, Magazine of Civil Engineer- departamentas, Vilnius, 138 p.
ing, 42(7), 6776. LST EN 1997-2 (2009), Eurokodas 7. Geotechninis projektavi-
Cheng, J., Zhang, G., Caraffini, F. & Neri, F. (2015), Multi- mas: 2 dalis. Pagrindo tyrinejimai ir bandymai. (Eurocode
criteria adaptive differential evolution for global numeri- 7. Geotechnical design Part 2: Ground investigation and
cal optimization, Integrated Computer-Aided Engineering, testing), Lietuvos standartizacijos departamentas, Vilnius,
22(2), 10307. 159 p.
Deng, J.-L. (1982), Control problems of grey systems, Systems MacCrimmon, K. R., (1973), An overview of multiple ob-
and Controls Letters, 1(5), 28894. jective decision making, in J. L. Cochrane and M. Ze-
Dundulis, K. (1997), The Lithuanian Unified Soil Classification leny (eds.), Multiple Criteria Decision Making, University
System, Lithuanian Geological Society Publ, Vilnius. of South Carolina Press, Columbia, SC, pp. 1843.
Eurofound (2013), Third European Quality of Life Survey Mangushev, R. (2004), Analysis of foundation installation for
Quality of life in Europe: Subjective Well-Being, Publica- old buildings in downtown Saint Petersburg, Soil Mechanics
tions Office of the European Union, Luxembourg. and Foundation Engineering, 41(5), 16871.
Fang, Z., Liu, S., Shi, H. & Lin, Y. (2009), Grey Game Theory Nguyen, V. L., Nie, L. & Zhang, M. (2013), Method ce-
and Its Applications in Economic Decision-Making, CRC ment post-grouting to increase the load capacity for bored
Press, Boca Raton, FL. pile, Research Journal of Applied Sciences, Engineering and
Hasanzadehshooiili, H., Lakirouhani, A. & Medzvieckas, Technology, 5(19), 472732.
J. (2012), Superiority of artificial neural networks over sta- Oloufa, A. A. & Ikeda, M. (1995), An automated environment
tistical methods in prediction of the optimal length of rock for soils- and terrain-dependent applications, Automation
bolts, Journal of Civil Engineering and Management, 18(5), in Construction, 4(2), 13946.
65561. Reuter, U. & Moller, B. (2010), Artificial neural networks for
Hwang, C. L. & Yoon, K. S. (1981), Multiple Attribute Deci- forecasting of fuzzy time series, Computer-Aided Civil and
sion Making-Methods and Applications: A State-of-the-Art Infrastructure Engineering, 25(5), 36374.
Survey, vol. 186, Springer, Berlin. RSN 91-85 (1987), Respublikines statybos normos. Greztini
Ishizaka, A. & Nemery, P. (2013), Multi-Criteria Deci- pamat projektavimas.(Lithuanian construction regulation.
sion Analysis: Methods and Software, Wiley, Chichester, Bored foundation design.), Vilnius, 55 p.
UK. Saaty, T. L. (1980), The Analytic Hierarchy Process, McGraw-
Jato-Espino, D., Castillo-Lopez, E., Rodriguez-Hernandez, Hill, New York.
J. & Canteras-Jordana, J. C. (2014), A review of application Sarma, K. & Adeli, H. (2000), Fuzzy discrete multicriteria cost
of multi-criteria decision making methods in construction, optimization of steel structures, Journal of Structural Engi-
Automation in Construction, 45, 15162. neering, 126(11), 133947.
Kaufmann, A. & Gupta, M. M. (1985), Introduction to Fuzzy Sivilevicius, H., Daniunas, A., Zavadskas, E. K., Turskis, Z. &
Arithmetic: Theory and Applications, Van Nostrand Rein- Susinskas, S. (2012), Experimental study on technological
hold, New York. indicators of pile-columns at a construction site, Journal of
Kelevisius, K., Gabrielaitis, L., Amsiejus, J., Norkus, A. & Civil Engineering and Management, 18(4), 51218.
Sikora, Z. (2014), Study of bearing capacity of vibratory pile Susinskas, S., Turskis, Z., Stasiskis, A., Valickas, J. &
applying acceleration record, Journal of Civil Engineering Rudzionis, Z. (2014), The efficiency analysis of foundations
and Management, 20(1), 14248. for buildings with bearing walls for Lithuania ground condi-
Kolay, P. K., Kumar, S. & Tiwari, D. (2013), Improve- tions, Journal of Civil Engineering and Management, 20(1),
ment of bearing capacity of shallow foundation on geo- 14248.
grid reinforced silty clay and sand, Journal of Construction Tezcan, S. S. & Ozdemir, Z. (2012), Allowable bearing pres-
Engineering, 2013, Article ID 293809 10 pp. sure in soils and rocks through seismic wave velocities,
Lee, D.-E., Lim, T.-K. & Arditi, D. (2011), An expert sys- Earth Science Research, 1(1), 98108.
tem for auditing quality management systems in construc- Tserng, H. P., Ngo, T. L., Chen, P. C. & Tran, L. Q. (2015),
tion, Computer-Aided Civil and Infrastructure Engineering, A grey system theory-based default prediction model for
26(8), 61231. construction firms, Computer-Aided Civil and Infrastruc-
Lepre, P., Rocha, C. G. & Dos Santos, A. (2010), Barriers and ture Engineering, 30(2), 12034.
opportunities in developing do-it-yourself (DIY) products Usmanov, R. A. (2013), Improvement of efficiency of appli-
for low-income housing, Journal of Construction in Devel- cation of condensed soil cushions to loose soils, Vestnik
oping Countries, 15(1), 2943. MGSU, 5, 6979.
Li, H., Adeli, H., Sun, J. & Han, J.-G. (2011), Hybridizing Xu, J. & Yao, L. (2011), Random-Like Multiple Objective De-
principles of TOPSIS with case-based reasoning for busi- cision Making, Springer-Verlag, Berlin Heidelberg.
ness failure prediction, Computers & Operations Research, Xu, K. & Zhou, X. (2011), Fuzzy-Like Multiple Objective De-
38(2), 40919. cision Making, Springer-Verlag, Berlin Heidelberg.
Multicriteria evaluation of building foundation alternatives 729

Xu, Z. (2014), Intuitionistic Preference Modeling and Interac- Zavadskas, E. K., Turskis, Z., Antucheviciene, J. & Zakare-
tive Decision Making, Springer-Verlag, Berlin Heidelberg. vicius, A. (2012b), Optimization of weighted aggregated
Xu, Z. (2015), Uncertain Multi-Attribute Decision Mak- sum product assessment. Elektronika irelektrotechnika, 6,
ing: Methods and Applications, Springer-Verlag, Berlin 36. http://erem.ktu.lt/index.php/elt/article/download/1810/
Heidelberg. 1468.
Zavadskas, E. K., Kaklauskas, A., Turskis, Z. & Tamosaitiene, Zavadskas, E. K., Vainiunas, P., Turskis, Z. & Tamosaitiene,
J. (2008), Selection of the effective dwelling house walls by J. (2012c), Multiple criteria decision support system for
applying attribute values determined at intervals, Journal assessment of projects managers in construction, Interna-
of Civil Engineering and Management, 14(2), 8593. tional Journal of Information Technology & Decision Mak-
Zavadskas, E. K., Susinskas, S., Daniunas, A., Turskis, Z. & ing, 11(2), 50120.
Sivilevicius, H. (2012a), Multiple criteria selection of pile- Zayed, T. M. (2005), Productivity and cost assessment for con-
column construction technology, Journal of Civil Engineer- tinuous flight auger piles, Journal of Construction Engineer-
ing and Management, 18(6), 83442. ing and Management, 131(6), 67788.

Vous aimerez peut-être aussi