An object-oriented model for construction method selection in buildings using fuzzy information

Accepted Manuscript

An Object-Oriented Model for Construction Method Selection in Buildings Using Fuzzy Information M. Ebrahiminejad , E. Shakeri , A. Ardeshir , MH Zarandi PII: DOI: Reference:

S0378-7788(18)30194-4 https://doi.org/10.1016/j.enbuild.2018.08.002 ENB 8738

To appear in:

Energy & Buildings

Received date: Revised date: Accepted date:

29 January 2018 28 June 2018 3 August 2018

Please cite this article as: M. Ebrahiminejad , E. Shakeri , A. Ardeshir , MH Zarandi , An ObjectOriented Model for Construction Method Selection in Buildings Using Fuzzy Information, Energy & Buildings (2018), doi: https://doi.org/10.1016/j.enbuild.2018.08.002

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT An Object-Oriented Model for Construction Method Selection in Buildings Using Fuzzy Information M. Ebrahiminejada, E. Shakeria,* , A. Ardeshira,b, MH Zarandic a

Department of Civil and Environmental Engineering b Environmental Research Center c Department of Industrial Engineering & Management Systems Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

CR IP T

Abstract:

A critical decision to be made once a building project is defined is the selection of a convenient construction method (CM) for various building elements. To deal with varied decision criteria and the classified nature of building elements, this paper

AN US

implements the decision-making process in an object-oriented model, in which each building element is modeled as a class with various possible alternative construction methods as subclasses. Although the subclasses inherit all the main class attributes, each subclass is specified by additional attributes. The uncertainty and vagueness of expert knowledge on the performance of each construction method regarding the

M

attributes are mapped with fuzzy numbers in this multi-criteria-decision-making problem to deal with the inherent imprecision of subjective judgment. To

Keywords:

PT

ED

demonstrate the use and capability of the model, a case study is presented.

AC

CE

Building, Construction method, object-oriented modeling, fuzzy, decision-making

*

Corresponding Author Email: [email protected]

1

ACCEPTED MANUSCRIPT

1. Introduction The global apprehension over climate change issues is driving the study of energy and carbon emissions through national and global reduction goals. Thus carbon footprint (CF) reduction is perceived as a global priority and an important step towards sustainability [1][2]. The building sector (residential and commercial) is responsible for around 20% of the world’s total delivered energy while about 60% of

CR IP T

the energy related carbon dioxide (CO2) emissions are ascribed to developing countries outside the Organization for Economic Cooperation and Development (OECD), many of which rely heavily on fossil fuels [3].

The carbon emissions from a building lifecycle are composed of (1) the embodied carbon (EC) which is incurred through material extraction, processing and

AN US

transportation and the construction phase and (2) the operational carbons (OC) incurred in the operation phase [4]&[5]. Significant efforts have been devoted to developing building energy codes, energy-efficient equipment, green material and education of clients [6]. Such focus has not been extended to reducing embodied

M

energy of the construction phase in the building sector for further reduction. During the construction phase, the construction process of each building element

ED

(e.g. structure, roofing, etc.) can be carried out in many different ways. These various ways in which materials and other resources are transformed into

PT

constructed products are known as construction methods [7]. Different construction methods require different combinations of labor, tools, equipment and material resources, despite their advantages and disadvantages in terms of the decision-

CE

making criteria [8]. Non-labor resources which release CO2 emissions are commonly powered by petrol, diesel and electricity of which diesel fuel and electricity produce

AC

the greatest total emissions [9]&[10]. The selection of an appropriate construction method according to project conditions, and regulatory, client, or self-dictated constraints is a critical issue for construction decision makers. It is a complex decision-making process which considers multiple attributes. The selection of construction methods can be viewed as a multi-criteria decision-making problem (MCDM) usually based on expert preference rather than numerical approaches [11]. MCDM is applied where an optimal decision is made in the presence of conflicting objectives: the trade-off between minimizing costs while 2

ACCEPTED MANUSCRIPT maximizing the quality, and minimizing emissions while maximizing productivity are few examples of the MCDM problems. The outcome of this decision-making process is an alternative which represents the most acceptable trade-off among the considered criteria. For solving MCDM problems, AHP has been used in various areas of construction [12], [13], [14], [15]. However, the inherent subjectivity and ambiguity of mapping an expert’s knowledge to an exact number is a shortcoming of the AHP approach [12]. Information used to make decisions in construction are not

CR IP T

always precise and experts prefer to express their knowledge using linguistic terms such as the “average”, “maximum” and “minimum” values. To address this issue, fuzzy approach methods have been widely developed. The fuzzy approach is applied to many construction problems, e.g. bridge construction methods [12], sustainable material selection for buildings [11], construction site selection [16] and

AN US

building performance assessment [17]. Thus, triangular fuzzy numbers (TFN) are used in this model to capture the fuzzy nature of information in construction. According to the UniFormat, building construction involves various building elements. However previous studies focus on selecting an appropriate construction method for

M

one building element under specific project conditions. For example [14] compares three different variants of a specific floor construction method using analytical

ED

hierarchy process (AHP) and [18] compares two construction methods for the building structure. A holistic model capable of focusing on multiple building elements

PT

each with varied selection criteria is missing in the literature. Also various decision tools, selection techniques and types of input information exists which might affect

CE

the final decision outcome. Such a situation poses a need for a model, compatible with the classified nature of building elements while allowing for modularity of the key decision components. The methodology to overcome the previously presented

AC

shortcomings is drawn from the object-oriented modelling (OOM) approach used in software engineering. The proposed object-oriented approach is a natural metaphor to implement real world decision-making problems in practice. The key decision-making factors and the fuzzy nature of input information can be best implemented in an object-oriented model. Concepts such as modularity, encapsulation, extensibility, abstraction and reusability, which are well defined in the object-oriented approach, help better meet business modelling requirements. To the best of the author’s knowledge, this is the 3

ACCEPTED MANUSCRIPT first implementation of a building construction method selection problem in an OOM. Also no fuzzy application was found regarding the selection of building element construction method. Using inherent OOM features, the model can easily be expanded to any extensions in terms of building elements, alternative construction methods, criteria, project conditions and input information including the fuzzy nature of information in construction. The model output presented as If-Then rules, presents the most suitable construction method for each building element based on possible

CR IP T

project conditions.

This paper presents a fuzzy-based approach for selecting construction methods using an object-oriented model.

It deploys object-oriented model to classify

attributes and functions of building elements and fuzzy analysis in order to deal with

AN US

the qualitative and subjective nature of expert judgments.

To reach the proposed objective, this research is presented as follows. In section 2, a background on decision-making in construction is presented to better understand the decision-making components in the construction industry. This is followed by the

M

presentation of the main concepts of OOM and MCDM. In section 3, model development is described by presenting the objects and classes used to build the

ED

model. To demonstrate the capabilities of the model, a case study is presented in

PT

section 4 which includes remarks on validation of the model.

2. Background

CE

The research model consists of a decision-making component implemented in an object-oriented model (OOM). The decision-making component is part of the OOM

AC

which is used to select the optimum construction method for various building elements according to the given decision criteria. The OOM provides the overall structure of this research’s methodology. This section first reviews the key problems and components of decision-making in construction. This helps identify the problem context and key decision-making factors to be implemented in OOM. The main OOM concepts are then presented followed by the MCDM process which is the decision making context used to solve the CM selection in this study.

4

ACCEPTED MANUSCRIPT 2.1.

Decision-making in construction: A review

Decision-making in construction is applied to an extensive domain: Project management (planning and scheduling among others) [19], equipment fleet selection for large projects [20], path selection for material delivery [21], supplier selection [22] and technology selection [23]. Building construction methods (CM) have also been the focus for decision-making but compared to building material selection [11], these studies are rather scarce [11]. Previous works considering building CM, have

CR IP T

focused on selecting a suitable CM for one building element. Some considered various structures[24] [25] and few investigated alternative floor CMs [26].

Three primary components of a decision problem are the decision makers, the decision tools and the selection techniques [27]. Selecting a CM is a decision-

AN US

making problem in which an alternative is selected among various alternatives based on a set of criteria using a selection technique. The selection is done based on an assessment of the consequences of each alternative which is carried out by a decision tool. The selected alternative represents a satisfactory trade-off among

M

problem criteria.

Decision makers aim to achieve their objectives under a set of criteria. As

ED

indicated in the literature, cost, time and technical criteria such as safety and quality have traditionally been the fundamental criteria when selecting a CM [18] [15]. With the ever growing global concerns over sustainability, decision-making criteria in

PT

construction like other businesses developed from merely economical and engineering criteria to encompass social and environmental aspects of construction

CE

as well [27]. Although [18] identified 33 sustainable performance criteria under Triple Bottom Line (TBL) but also showed that some criteria are inter-related and can be

AC

grouped into a smaller set of primary factors. Through factor analysis, seven latent factors were generated from the initial criteria. However, the study shows that time and cost continue to be the most important criteria for choosing a CM while social and environmental concerns are increasingly important [18]. Among environmental concerns, global warming is the focus of this study. Under the 1997 Kyoto Protocol, six greenhouse gases (GHG) related to global warming namely carbon dioxide (CO2),

methane

(CH4),

nitrous

oxide

(N2O),

perfluorocarbons

(PFCs),

chlorofluorocarbons (CFCs) and hydrofluorocarbons (HFCs) must be reduced [28]. Carbon dioxide (CO2) is the primary contributor to global warming among the GHG 5

ACCEPTED MANUSCRIPT is the atmosphere [29], [30], comprising approximately 83% of all GHG emissions [31]. In this study, CO2 is the focus regarding environmental impacts of a CM. The other primary component of a decision problem is the decision tool. The consequences of an alternative based on input information and assumptions are evaluated using a decision tool. The decision tools used by researchers to evaluate the outcome of an alternative range from the most primitive such as mathematical

CR IP T

regressions to optimization techniques including linear programming and queuing theory and meta-heuristic models to complex trained systems such as multi agent simulation and knowledge based expert systems [25][27].

Selection techniques are the other primary component of decision-making. The selection of an alternative among others is usually thorough comparing the

AN US

consequence of alternatives under the decision criteria. Again, selection techniques can be as simple as ranking and pros/cons or as complex as the Delphi method and analytical hierarchy process (AHP).

2.2.

Object-oriented (OO) concept

M

The object-oriented methodology became a major area of attention in the 1980s. It originated from the first object-oriented language programming language Simula-67

ED

developed in Norway, in the late 1960s [32]. Several efforts have extended the software system engineering constructs to decision-making problems: supply chains

PT

[33], production scheduling decisions [34] and MCDM modelling [32]. An object is the fundamental element in OOM: It encapsulates different properties of

CE

the entities such as attributes and relations with other objects. A set of object instances with the same nature is known as a class. The concept of object class is

AC

relevant in the classification of building elements, where specific building elements exist, but a wide variety of alternative construction methods may be available in those elements. Objects intercommunicate with other objects and its environment through message transfer between each other. Objects respond to messages by selecting a corresponding method (also called operations or behaviors) to execute the received message [32][35]. Separation of the external aspects of an object from the internal aspects is known as encapsulation or the black-box paradigm. Also with the encapsulation, the boundary 6

ACCEPTED MANUSCRIPT and the interactions cross the boundary are defined which supports stepwise process refinement and allows changes be applied to internal detailed activities of a class until necessary. Modularity, as Booch [36] defines, allows for decomposition of a system into a set of cohesive and loosely coupled modules [35]. The emergence and use of design patterns as a best practice has enabled the development of robust and highly reusable objects. The strategy pattern and the mediator pattern [33] are

CR IP T

two important examples [37]. Several relations can be defined between objects. An “association” which is the most general of all relationships between objects, provides a pathway for communication. Two types of associations are recognized among objects: inheritance (IS-A) and aggregation (HAS_A). The number of independent system components can be

AN US

reduced using the Inheritance feature of object-orient modelling thus simplifies the system. Through inheritance, objects acquire characteristics from other objects [32]. Inheritance represents generalization and specialization. For example, while “structure” IS-A building element, the “concrete structure” as a subclass of “structure” inherits all the attributes of the “structure” class, but has additional attributes that

M

specializes “concrete structure” from other subclasses of the “structure” class. Aggregation is the ability to create objects from other objects, with each, a part of the

ED

aggregate object (a building is composed of a sub-structure, a structure and finishing). Composition is a strong type of aggregation where each part only belongs

2.3.

PT

to one aggregate object [35].

Multi-criteria decision-making

CE

Construction managers constantly encounter situations where they have to make a choice among two or more alternatives through the evaluation and assessment of

AC

the consequences of different options based on the criteria set by the decision maker to achieve an objective [27]. The selection of construction methods is a multi-criteria decision-making problem (MCDM) usually based on expert preference rather than numerical approaches [11]. MCDM is applied where an optimal decision is made in the presence of conflicting objectives: The trade-off between minimizing the cost while maximizing the quality and minimizing emissions while maximizing productivity are few examples of the MCDM problems. The outcome of this decision-making process is an alternative 7

ACCEPTED MANUSCRIPT which represents the most acceptable trade-off among the considered criteria. [14] is an example of a MCDM in roof material selection. MCDM problems are classified in two main categories: the multi objective decisionmaking (MODM) and the multi attribute decision-making (MADM) problems. In MODM, the best choice is designed according to system constraints, different objectives and decision makers’ optimum value. While in MADM models the

CR IP T

optimum selection, is considered based on the decision maker’s attributes. MADM models are in fact selection models and are used to select the most suitable option among m options. The construction method selection is in fact a MADM problem. MADM is usually formulated using a decision matrix ( ), in which Ai represents the ith alternative, xj represents the jth attribute and rij represents the value of the jth attribute

AN US

for the alternative i. The aim in MADM is ranking and ultimately selecting the best alternative.

Table 1. A sample decision matrix used in MADM

Attributes

ED

M

Alternatives X1 A1 r11 A2 r21

rm1

Xn r1n r1n

rm1

rmn

PT

Am

X2 r12 r22

Various methods are used to solve compensatory MADM problems including the

CE

weighted sum model (WSM) as the earliest method, the weighted product model (WPM) which seeks to overcome the weaknesses of the earlier model. ELECTRE

AC

and TOPSIS are other widely used methods [38]. In most MADM problems, the relative weight of the attributes is required. This issue is more important in compensatory models. To determine the weight of the attributes, besides the expert judgment, methods such as the Entropy method and the Analytical Hierarchy Process (AHP) are available(Triantaphyllou, n.d) The weight of attributes in the MADM problem in this paper is determined by the decision maker (expert judgment) and the problem is solved using WSM. This all is implemented within an object-oriented model. For acquaintance, the following section provides more information on the concepts of object-oriented modeling. 8

ACCEPTED MANUSCRIPT

3. Model development 3.1.

Model overview for construction method selection

The identification of model objects is the initial step in OOM. Figure 1 presents an overview of the modelling stages. The main model objects in this study are the building elements and alternative construction methods, the key MCDM components and the external entities and the project description. These objects and classes are

CR IP T

categorized under structural objects and policy objects. Separating the policy objects from structural objects enables flexible modeling capability and the option for tracking varying decision-making behaviors. The following two sections give insight to the structure objects and policy objects of the model.

Decision Context: MCDM

AN US

Problem: CM selection

Define Objects and Classes

M

Structure Objects and Classes

Policy Objects and Classes

ED

Define Communication Mechanism

3.2.

CE

PT

Figure 1. Object-Oriented problem modelling and analysis

Structure objects

The structural objects of this model represent the building elements and its

AC

alternatives construction methods. Elements are the building blocks of a building structure. To describe and manage building elements, a consistent elemental classification framework for buildings is required as a reference. The ASTM E1557 standard (UniFormat) and the MasterFormat are the two main organization standards of building content. While MasterFormat is a material-based organization standard, UniFormat arranges construction information based on functional elements, regardless to the materials used to accomplish them. Utilizing the UniFormat is more congruent with decision-making and modeling at element level 9

ACCEPTED MANUSCRIPT and thus the approach is used in this research. Based on UniFormat, three building elements: substructure, structure and roofing, are included in this study. Each element (based on UniFormat), is considered and modeled as a “class” and its alternative CMs are modeled as “subclasses” to their relative classes. Each class consists of a name, specific attributes and a method or operation (Figure 2). The decision-making criteria values ( ), (where

denotes the criteria and

denotes a

CR IP T

sequential number of criteria from 1 to k) are presented as attributes. Element classes have both common and element specific criterion which specializes each class from other element classes. The amount of required material and energy for

AN US

each CM is also modelled as attributes.

Roof

Class name

Class attributes including: Decision criteria and required common resources for all the alternatives(Equipment and material)

Class operations / Methods

M

-time -cost -quality -. -resourcesEquipment -resourcesMaterial +callWeight() +calculateIndex()

JoistClayblock

JoistCementblock

ED

SteelDeck

JoistPolyesterneBlock

-clayBlocks

-cementBlocks

+calculateJoistClayCF()

+calculateJoistCementCF() +calculateJoistPolyesterneCF()

ConcreteSlab

-polyesterneBlocks +calculateConcretSlabCF()

PT

-galvanizedSheets -shearStuds +calculateSteelDeckCF()

CE

Figure 2. Object-oriented presentation of the roof element class and roof alternatives presented as its subclasses.

Each subclass obtains a unique operation as the method for calculating its CF (CO2-

AC

e). A given construction method’s CF is composed of the sum of CF from all sources (Eq. 1) where i denotes an alternative (i=1,…,m) and j denotes energy consuming resources (j=1,…,n): (

)

∑

(

)

(1)

Three sources of energy are considered in this model: embodied energy of material (Eq. 2), the energy used to transport the material to the site and the energy

10

ACCEPTED MANUSCRIPT consumed for the placement / erection of material during the construction phase (Eq. 3). Both fuel and electricity consumption are considered for non-material resources. (

)

(

(

)

(

)

(2)

) ( (

)

(3)

)

CR IP T

For fuel consuming (petroleum and diesel) resources, the EPA NONROAD model provides emission rates by time (g/h) or by horse-power (g/hp-h). The emission rates (g/h) are multiplied by the working time of non-material resources to calculate total emissions for each fuel consuming resource [40]. For calculating the CO2 emissions from electricity powered resources, the electricity consumption is multiplied by , an

AN US

electricity emission coefficient (kgCO2/kWh) [31]. The electricity emission coefficient ( ) values vary according to the average share of energy sources in different countries [41].

3.3.

Policy objects

M

The policy objects encompass the business logic used by decision-makers for selecting the suitable CM among various alternatives. The criteria (decision-makers’

ED

objectives) the selection techniques and decision tools are some of the main entities modeled by the policy objects. Taxonomy of the policy objects of this model is provided in Figure 3. The following section describes the policy objects used in this

Aggregator

Criteria Weights

Decision-making criteria

Constants

Policy objects

Suppliers

Emission factors

Resources

AC

Project bill of quantities

CE

PT

model.

quantity (BOQ): A BOQ provides an inventory for the amount of work estimated in various disciplines and activities in order to complete a project. It reflects the physical properties of the building including the land dimension, the floor area ratio (FAR), the number of stories above and below the ground and the height of each story. The amount of resource required for each CM is derived from the BOQ. The BOQ which 11

ACCEPTED MANUSCRIPT is presented as an instance in our OOM provides the project specific data in our model. Resources: The resource object, provides the amount of energy consumed through running tools and equipment and embodied in materials used in each CM. This class communicates the input data for CF calculation with other classes in the model. Emission conversion: Consumption of a unit of energy (diesel, petroleum and

CR IP T

electricity) and material (steel, concrete and water) is converted to CF using specific conversion rates. This object provides the emission factors (material), emission rates (fuel) and electricity emission coefficient (electricity) required for CF estimation of CM resources.

The Supplier object, holds the essential information regarding the

AN US

Suppliers:

suppliers. Material transportation to the site is a significant contributor to CF. Supplier name, distance from the site, and the required transport vehicle description (capacity, power) are stored in the supplier object. The CF due to material

object (Eq. 4-6). (

ED

(

M

transportation is calculated using the unique method designated in the supplier

⁄ ⁄

(4)

)

(5)

)

(6)

PT

(

)

Constants: Assumptions used in the model such as the average speed of delivery vehicles and the height of each floor are stored in this object.

CE

Decision tools: To assess the advantages and disadvantages of an alternative CM, a decision tool is required. The weighting method, among various methods, is the

AC

decision tool used for solving the MCDM problems in this research. objective function

The linear

is stored as the method or operation of every CM is

considered as the CM index (priority score) (Eq. 7) where wi >0 and ∑

.

Figure 3. Taxonomy of policy objects of the model

Bill of ∑

(7)

12

ACCEPTED MANUSCRIPT Weight: As decision makers’ preferences change in various project conditions, a separate class is defined to capture the importance of various decision criteria. The weight (w) or importance of a criteria (objective) depends on the project conditions which is provided by the decision-maker. Selection techniques: Ranking fuzzy numbers is used in this paper to make the final selection. The final decision is based on selecting the maximum index which is a

CR IP T

triangular fuzzy number (TFN). To have a crisp preference of indices we require a method for ordering fuzzy numbers. Although various methods of ranking fuzzy numbers exist but they all don’t induce the same result and some TFNs are not simply comparable. Base on the results of [42], we use the extension principle in this

AC

CE

PT

ED

M

AN US

paper to rank TFNs [43].

Figure 4. Class and object presentation of the model

The rational rose® is used to design the model using the object-oriented concepts (Figure 4) and then programmed is carried out using Eclipse for Java®. 13

ACCEPTED MANUSCRIPT

3.4.

Fuzzy input information

Data used to make decisions in construction are not always precise. However, decision-makers can easily express their assessment on the CM alternatives using linguistic terms such as the “average”, “maximum” and “minimum” values rather than numerical ratings. As illustrated in Figure 5 the symmetric triangular and half

CR IP T

trapezoidal membership functions are used to represent the linguistic terms. 1 TFN1 Very Low

TFN2 Low

TFN3 Average

TFN4 High

0 0

1

2

3

4

AN US

0.5

5

6

7

TFN5 Very High

8

9

10

Figure 5. Membership function for linguistic values

Triangular fuzzy numbers (TFN) denoted as (a1, a2, a3) are based on the minimum

M

value, the average value and the maximum value that describes an event. Its

( {

(

)⁄( )⁄(

)

(8)

)

PT

( )

ED

membership function can be defined as (Eq. 8):

The outcome of operations on fuzzy numbers is in the shape of fuzzy set thus the

CE

result is expressed in membership functions. However, operations on TFNs which are used in this model (Eq. 9-15) are of more importance to us. The results from addition or subtraction between TFN are also TFN. But the results from

AC

multiplication, division, maximum or minimum operations are not TFN but it is often assumed to be TFN as approximation values. Because fuzzy numbers are used instead of crisp numbers, the solution methods are different from crisp numbers. Consider TFN A = (a1, a2, a3) and TFN B = (b1, b2, b3). Addition, subtraction and symmetric image operations on A and B are as follows (refer to section 4.4 and 4.5 of [43] for details) (Eq. 9-11): A(+)B = (a1+b1, a2+b2, a3+b3)

(9)

A(-)B = (a1-b3, a2-b2, a3-b1)

(10)

14

ACCEPTED MANUSCRIPT -(A) = (-a3, -a2, -a1)

(11)

For multiplication and division, the α-cuts of the two TFN are used (Eq. 12-15): Aα=[(a2-a1)α+a1,-(a3-a2)α+a3],

(12)

Bα=[(b2-b1)α+b1,-(b3-b2)α+b3]

(13)

Aα (●) Bα

[a1b1, a2b2, a3b3]

(14)

Aα (/) Bα

[a1/b3, a2/b2, a3/b1]

(15)

CR IP T

A fuzzy decision matrix, representing the performance value of the alternatives in terms of the criteria as TFNs is constructed. The decision-maker uses TFNs to express the importance of the criteria. The sum of the modal values of the TFN weights which represent the criteria weight should be equal to 1.

AN US

The input information in this model can be categorized into two sets: (1) the available information to experts (i.e. the duration, the cost, the quality, the limitations and the technical requirements of various construction methods) (2) the unknown information to experts (environmental impacts of material and equipment) which has to be calculated.

M

One set of input information for the model is the available information which construction professionals are familiar when making decisions (i.e. the duration, the

ED

cost, the quality, the limitations and the technical requirements of various construction methods).

PT

However, another set of data is required to run the model which reflects the environmental impacts of material and equipment and is new to contractors. This

CE

later data in the scope of this research is limited to GHG emissions and can generally be obtained either through direct field measurement or indirect estimation.

AC

In direct measurement, the amount of studied emissions is measured using the sensors of a portable emission measuring system (PEMS) placed into the tailpipe vehicles. Indirect estimation which is more common, estimates the amount of emissions using specific emission factors (Eq. 1-6). The sources used to provide the model’s input information data are an important part of the decision-making model. Figure 7 provides the sources of input information for various model objects.

15

ACCEPTED MANUSCRIPT

4. Case study To demonstrate the application of the object-oriented model in optimum construction method selection a numeric example based on a case study project is presented. The project is a nine story height (including three underground floors),

AN US

plan and section is provided in Figure 6.

CR IP T

commercial and office building with a total of 3000 m2 area in Tehran. The building

M

Figure 6. Plan (right) and section (left) of the case study building.

The objective is to select the most appropriate construction method for three building

ED

elements: deep excavation and soil stabilization, structure and roof. The overall system presentation of the model is presented in Figure 7.

PT

An expert group is formed based on practical experience and acceptable level of knowledge, involving eight experienced engineers in building construction. Two

CE

members are from the university and the other six are project managers with 20-30 years of experience who are the main decision-makers in their companies regarding

AC

the selection of construction methods. In this study the traditional criteria for selecting a CM is used in the selection procedure for all building elements: time, cost and quality. In addition, the carbon footprint is considered as an image of environmental impacts of CM. These criteria can be characterized by their associated sub-criteria: Cost (material cost, transportation cost, labor cost), time (construction time, lead time), quality (defects and damages) and carbon footprint (embodied carbon of material and carbon form energy consumption). However, for each building element a specific criterion is also considered: A decision criterion which affects the selection of a CM for the 16

ACCEPTED MANUSCRIPT substructure is the legal constraints such as the permit to use nailing under the land of a neighbor’s house. Required initial capital is another criterion which influences the selection of an appropriate CM for the structure. Roofing CM selection is influenced by the design capability of each alternative. Total work quantity calculations for the three building elements, based on the input data is presented in Table 2. Source of Information

Project Description

MCDM-WSM

Land Perimeter and Area, Constructed ratio of land Number of Stories Below & Above Ground,

Supplier Information

Material Supplier Information

Alternative Performance Values in terms of CF

Industrial References

Required Resources for the Construction of One Unit of Each Building Element

Equipment & Tools

Material

Alternatives’ Assessment

CM Selection

Emission Factors/ Rates

M

International References

AN US

Delivery Vehicle Information

Selection Tool Fuzzy Ranking

CR IP T

Project Documents and People

Decision Tool

Type of Information

Fuzzy Input Information

ED

Experts

Fuzzy Input Information

Alternative Performance Values in terms of the criteria (except for CF).

Legend:

Importance of Criteria (Weight)

Model Input Information

Model Predetermined Information

PT

Figure 7. Overall system presentation

CE

Table 2. The BOQ for the basement excavation, structure and roof based on the building description coded in JAVA and used in the model. Basement excavation (m2):

AC

buildingCoEfficient = BillOfQuantities.i().getLandPerimeter() * BillOfQuantities.i().getStoriesUG() * Constants.FLOOR_HEIGHT * BillOfQuantities.i().getOccupiedArea() / 100; Structure (m2):

if(elementName.toLowerCase().contains("roof") || elementName.toLowerCase().contains("structure")) { buildingCoEfficient = BillOfQuantities.i().getLandArea() * BillOfQuantities.i().getOccupiedArea() / 100 * (BillOfQuantities.i().getStoriesAG() + BillOfQuantities.i().getStoriesUG()); Roof (m2): Land area * (# Stories UG + AG ) * %Occupied area/100

17

ACCEPTED MANUSCRIPT According to ASTM-E, (1) basement excavation and walls, (2) structural elements and (3) floor construction are considered as the focused building elements. Based on author’s previous study [44], these three selected elements are the most energy consuming elements during construction (Figure 8). A set of alternatives for the three

AN US

CR IP T

main building elements is also considered.

Figure 8. Building elements studied in this paper based on ASTM E.

Concrete retaining walls, nailing and anchorage and concrete piles are the main methods used to stabilize the cuts in Tehran while the basement excavation process

M

using backhoes and trucks for excavation and material hauling is a similar process in all the previously mentioned methods. Despite various structural designs, the type of

ED

material can either be steel or concrete. According to reports, more than 97% of the constructed structures over the past 5 years in Tehran, were either concrete or steel

PT

[45] [46]. The connections in steel structure buildings are traditionally welded but recently bolt and nut connections are also used in Tehran. Thus three alternatives for

CE

structure is considered: steel (welded connections and bolt & nut connections) and concrete structure. The main floor systems considered in this study are steel deck,

AC

concrete slabs and block joisted systems with various types of blocks. Selecting a suitable deep excavation method however is more technical than the structure or floor system. It depends on critical issues such as: soil type, the depth of the cut and the surrounding conditions. The Tehran Construction Engineering Organization (TCEO) reports show that thee construction truss, nailing and the continuous piles are the three alternative major deep stabilizing methods used in Tehran. Thus they are considered as alternatives in this study (Figure 9).

18

CR IP T

ACCEPTED MANUSCRIPT

Figure 9.Truss, nailing and piles (Left to right) studied as deep stabilizing methods in this study.

Various criteria are used when making decision on construction method selection. However, in this study seven decision criteria are considered through the literature and interviews with experts. It should be noted that the selected criteria

AN US

may not be absolute. The judgment result for the alternatives of each element with respect to the criteria is extracted from expert knowledge, except for the carbon footprint which is new to contractors and needs to be estimated. The criteria and an expert’s judgment for the alternatives with respect to the criteria are given in Table 3

M

as an example.

A. DEEP EXCAVATION

CE

B. STRUCTURE

Alternatives

AC

C. ROOF

A1 A2 A3 B1 B2 B3 C1 C2 C3 C4 C5

PT

Building elements

ED

Table 3. Decision matrix based on expert opinion

A A H L L VH A A L A H

A VH H A H VL L A A H A

A H H A H A L L A H VH

H VH VL A L VH A H H L VH

VH VL VL N/A N/A N/A N/A N/A N/A N/A N/A

N/A N/A N/A H VH L N/A N/A N/A N/A N/A

N/A N/A N/A N/A N/A N/A L L L A H

Criteria: Z1 = Time, Z2 = Cost, Z3 = Quality, Z4 = CO2-e, Z5 = Legal constraints, Z6 = Required initial capital, Z7 = Design capability TFNi: TFN1 = Very low (VL), TFN2 =Low (L), TFN3 =average (A), TFN4 =High (H) and TFN5 =Very high (VH) ; N/A = not applicable. Alternatives: A1=Nailing & Anchorage, A2=Continuous concrete piles, A3=Truss B1=Steel-welded, B2=Steel- Bolt & Nuts B3= Concrete C1=Joist-Concrete, C2=Joist-Clay C3=Joist-polystyrene, C4=Steel deck, C5=Concrete slab

The carbon footprint of an alternative is strongly related to the distance of the supplier from the site and the total amount of required material. The importance of 19

ACCEPTED MANUSCRIPT each criteria is defined by the decision maker as an input for the model based on linguistic terms previously described. It is emphasized that the number of building elements, the alternatives and the criteria are easily expandable using the proposed object-oriented model. The following sections describe the estimation procedure used in the model.

Carbon estimation

CR IP T

4.1.

A detailed analysis of the carbon footprint for all building element alternatives is performed considering the embodied carbon of material, delivery of material



AN US

(transportation) and the construction phase as described below. Embodied carbon of material

Each building is a complex assembly of many elements composed of different materials. The generic term of embodied carbon or the specific cradle to gate embodied carbon term refers to carbon emissions incurred during extraction,

M

production and process of material. The embodied carbon and energy factors for a number of building materials used in this study are presented in table 4. In this study,

ED

the quantity take offs for a unit of a building element is applied to estimate the required material (such as steel, concrete, water and reinforcement bars). The embodied carbon (EC) of each material is then estimated by multiplying its quantity

PT

by the corresponding carbon footprint value of the material. The total EC of an element incurred through its material is then calculated by adding up the EC of all



CE

the materials used in the specific element. Construction

AC

The material described above are handled and placed using construction equipment and tools which consume liquid fuel or electricity. In this study, industrial references are used to estimate the equipment/tool working time for one unit of each of the alternative elements during the construction phase. The usage time is influenced by various factors such as operator experience, project conditions and weather. To deal with the uncertainties, the carbon footprint during construction is presented as TFNs. The carbon footprint due to construction equipment and tools usage for the construction of one unit of roof element is presented in Table 5. 20

ACCEPTED MANUSCRIPT



AN US

CR IP T

Table 4. Embodied carbon of major construction materials Embodied carbon* Material Unit (kgCO2-e/ unit) Cement Tone 845a Clay Kg 0.36 a 2 Cotton Sq.m (m ) 0.22 a 3 Sand Cu.m (m ) 7.9 a 3 Gravel Cu.m (m ) 11.97 b 3.84 b Polystyrene Kg 2.89 a Steel Kg 3.01 a Steel (galvanized) Kg 3 0.81 b Water Cu.m (m ) Wood Cu.m (m3) 25 a *Accessed through http://www.carbonfootprint.com/factors.aspx a Bath University, UK b Ecoinvent 2.2

Transportation

Transportation in construction includes both on-site and off-site transportation of material, equipment and personnel. The transportation vehicle features such as the

M

type, the capacity, the fuel type and the fuel consumption rate vary. In this study, only material transportation to the site is considered and an appropriate vehicle is

ED

assigned for the delivery of each material based on expert knowledge. It is also assumed that the material supplier is selected among various suppliers.

PT

In Table 6 supplier and delivery vehicle information for a sample material is presented. When the decision maker selects a supplier from a predefined set of

CE

suppliers, according to the supplier’s distance from the project site, the number of trips required for delivering the total amount of material to the site is calculated and based on the emissions factor of the vehicle, the carbon footprint due to material

AC

transportation to site is estimated. These calculations are carried out as a method for each alternative within its related class.

21

ACCEPTED MANUSCRIPT

Table 5: Carbon footprint (gram-CO2) for various roof alternatives due to equipment and tool usage during construction

d

e

c

CR IP T

b

AN US

a

h

M

g

f

i

ED

C1 to C5 are roof alternatives EQPi = equipment/tool Id. used in modelling

PT

Table 6. Embodied energy and carbon factors for a sample construction material Supplier description Distance (km)

CE

Name

A 100 B 143 C 247 Average vehicle speed is assumed 60km/h

AC

Material 1

Delivery vehicle description Capacity EF Name (tonne) CO2-e(kg/h) Dump truck 1 16 100.8 Trailer 1 35 210 Trailer 2 40 280

5. Results and discussions The comprehensive combination of weights for different criteria defines the

project status. Considering n criteria (n=5) each with a weight of w, where wi ∈ {TFN1, TFN2, TFN3, TFN4, TFN5}, the selection problem falls in one of the wn = 625 single objective linear programs which represent 625 different project status. Each alternative element is suitable for a specific project status. 22

ACCEPTED MANUSCRIPT The decision maker has to select a construction method based on his/her project status. For example, in one project you might be tight in time and short on budget but there are no constraints on the alternatives regarding legal issues. So you can probably select nailing as an alternative construction method for deep excavation and deep stabilization. That would be one project status. While if neighbors’ written approval is difficult to obtain, because of strict legal issues, the new project status

CR IP T

might exclude nailing. Solving each of the 625 single objective linear programs provides the solution for a construction method selection problem for one specific project status. In brief it provides a rule for each project status. The project status (input) and the suitable construction method selected for the three building elements (output) compose an

AN US

input-output table which retrieves the suitable construction method (output) for any input (project status). A set of generated rules which represent some of the project boundary status are provided as an input-output table in Table 7.

3 4

AC

5 6 7 8 9

PT

2

element

w(Time) = Very Low Other criteria same weight w(Cost) = Very Low Other criteria same weight w(Quality) = Very Low Other criteria same weight w(CO2-e) = Very Low Other criteria same weight All criteria same weight w(Time) = Dominant criteria w(Cost) = Dominant criteria w(Quality) = Dominant criteria w(CO2-e) = Dominant criteria

CE

1

Selected construction method for the building

Input: Project status

ED

Rule #

M

Table 7. A set of rules generated for the most suitable construction method according to the given project status. Output:

Substructure

Structure

Roof

A1

B2

C4, C5

A1

B2

C4

A1

B2

C4

A1

B2

C4, C5

A1 A1, A2 A1 A2, A3 A3

B2 B1, B2 B3 B2 B2

C4 C3 C1 C5 C4

The selection technique used in this study to conclude the final decision is a ranking method adopted to rank the TFNs based on their priority score. The final decision is effected by the adopted ranking method, as different ranking methods yield to different rankings among TFNs [42]. The TFN representing the CM index for rules 1

23

ACCEPTED MANUSCRIPT to 5 are presented in table 8. In some cases such as rules no. 1 and 4 the selected alternative might change if the selected ranking method is altered. Table 8. Index TFNs for rules 1-5 based on project status described in Table 7 for roof and structure. Structure Alternatives

AN US

CR IP T

Roof Alternatives

Rule 2

Rule 3

AC

CE

PT

ED

M

Rule 1

Rule 4

24

ACCEPTED MANUSCRIPT

Rule 5

For two main reasons, it is difficult to compare the results of this study with previous

CR IP T

studies: First, this study presents the results for various project conditions (different weights for criteria), while to the knowledge of the authors, except one study [47], other studies represent one specific set of weights for criteria and represent one project condition. Therefore, the criteria are not comparable. Second, the selected alternatives for the three building elements considered in this study are not the same

AN US

as the subject of previous studies. Therefore, the alternatives are different to compare. The objective of model validation is to ensure that, when provided with a set of legal inputs, the system output is equivalent to that provided by the best human experts [48].Therefore to validate the model the focus can either be on

M

individual components (rules) or by examining the system operation as a whole [49]. Since, each case study represents only one of the 625 possible project conditions

ED

presented as If-Then rules by the model, exhaustive testing of model’s behavior for all possible input values (Project conditions) is impractical. Therefore, as a

PT

component validation approach, one can manually determine which rules fire the most. In this study this means to determine which project conditions are more

CE

dominant so as to assess the accuracy, representativeness and validity of those individual rules [49]. In this study, the model is applied in a real case project to compare the model outputs (selected construction method under specific project

AC

conditions) with the results in the real case. In this regards the project decisionmakers were asked to model the project conditions through determining the importance (weight) of each of the decision-making criteria for each of the building elements. The construction method selected by the decision-makers are then compared with the model output. Based on the results shown in Table 9, the results from the model are similar to the CM selected by the project decision makers in practice.

25

ACCEPTED MANUSCRIPT

Table 9. Comparison of model output with the CM selected by the project decision makers in practice. Project Condition

Building elements A. Deep Excavation B. Structure C. Roof

H VH VH

H VH H

H VH H

L L L

VH

N/A

N/A

N/A

VH

N/A

N/A

N/A

H

Model Output

Selected CM in the case study

A1 B2 C4 or C5

A1 B2 C4

For validating the system as a whole, it is supposed that the shell (OOM and MCDM)

CR IP T

are already verified by the developers or previous literature. Thus, only the knowledge and its representation needs verification [49]. Content validity determines the content representativeness of an instrument. Index of content validity (CVI) is the most widely used quantification of the content validity. In this regards responses of experts in the field is reviewed. This involved a presentation of the OOM (including

AN US

the identified objects, the defined methods and assumptions), a structured 4-point scale rating questionnaire of the content relevance and open questions for comments on areas of improvement. The CVI is derived from the results of the questionnaire. The proportion of experts whose endorsement is required to establish

M

content validity beyond the 0.05 level of significance for different number of experts is presented in [50]. Based on the results, the system validation as a whole in this

ED

study is also achieved.

The carbon intensity of energy sources used in each country affects the local EF.

PT

Thus for using EF values, the priority is the local /domestic information compared to foreign data. However, when local data is not available, using EF of non-local

CE

databases is inevitable. The emission factors used in this study are not local which effects the estimation of the carbon footprint in various stages. This is purely a limitation to this study. There are techniques to convert emission factors obtained

AC

from international database to local emission factors based on the energy efficiency of the industries [41] but this is not in the scope of this study and has not been applied. However, it should also be considered that in this study a set of alternatives are compared and all use the same set of EF to calculate their carbon footprint. Therefore, for the sake of comparison, the effect of foreign EF is negligible. A final issue is the sensitivity analysis of the weight of the criteria in MCDM problems. When criteria (such as cost, time, quality, etc.) are expressed in qualitative terms, it is difficult to represent the importance of each criterion, accurately. 26

ACCEPTED MANUSCRIPT Identifying the critical criteria and the accurate evaluation of the weight of these criteria improves the decision-making process. The most critical criterion is not always the criterion with the highest weight. The change in the best alternative and the change in the ranking of any alternatives are the interest regarding the most critical criterion. This issue is detailed in [39]. However, it should be noted that the subject of this study is the selection of the most suitable alternative under specific project conditions. Therefore, change in the ranking (other than the best alternative)

CR IP T

is not of any importance. On the other hand, the weights are also given by the decision maker itself, therefore the above mentioned sensitivity analysis is not

AN US

carried out in this study.

6. Conclusion

Construction method selection is a frequent problem facing decision-makers in the beginning of construction projects. This study presents the methodology of object-

M

oriented decision-making modelling as an initial step in the automated modelling of a common decision-making problems in construction namely the selection of the most

ED

suitable construction method for building elements. This approach is useful for many reasons. First, object-orientated models can model the real world naturally and in a

PT

simple way. Classes (objects) are fit for classification of attributes of different building elements. Secondly, existing knowledge and past experience can easily be used by

CE

the inherent features of object-oriented modelling. The problem context, key decision factors and expert knowledge can be reused for future decision problems with similar features. The results from this study presented as a set of rules, helps decision

AC

makers choose a construction method suitable to the conditions of their project. Although the results show that environmental issues were less important to decision makers in this study compared to environmental issues but it makes decision makers conscious towards the impact of their selections on the environment. Future recommendations for research can include applying different ranking methods for TFNs and comparing the cost of applying different project conditions.

27

ACCEPTED MANUSCRIPT

7. Acknowledgments The authors would like to thank all the experts that spent their valuable time to provide the required information. Without their co-operation, this research would not have been possible. We are also grateful to the anonymous reviewers for their helpful and constructive comments.

[2]

[3] [4]

[5]

[9]

[10]

AC

[11]

PT

[8]

CE

[7]

ED

M

[6]

K. R. Piratla, S. T. Ariaratnam, and A. Cohen, “Estimation of CO2 Emissions from the Life Cycle of a Potable Water Pipeline Project,” J. Manag. Eng., vol. 28, no. 1, pp. 22–30, 2012. M. Ebrahiminejad, E. Shakeri, and A. Ardeshir, “ESTABLISHING ABATEMENT ALTERNATIVES IN,” in 3oth Annual ARCOM Conference, 2014, no. September, pp. 163–172. EIA, “International Energy Outlook 2016,” Washington DC, 2016. Z. S. M. Nadoushani and A. Akbarnezhad, “Effects of structural system on the life cycle carbon footprint of buildings,” Energy Build., vol. 102, pp. 337–346, 2015. G. Kang, T. Kim, Y. Kim, H. Cho, and K. Kang, “Statistical analysis of embodied carbon emission for building construction,” Energy Build., vol. 105, pp. 326–333, 2015. P. Nejat, F. Jomehzadeh, M. M. Taheri, M. Gohari, and M. Z. Abd. Majid, “A global review of energy consumption, CO2 emissions and policy in the residential sector (with an overview of the top ten CO2 emitting countries),” Renew. Sustain. Energy Rev., vol. 43, pp. 843–862, Mar. 2015. C. B. Tatum, “Classification System for Construction Technology,” J. Constr. Eng. Manag., vol. 114, no. 3, pp. 344–363, 1988. D. W. Halpin and L. S. Riggs, Planning and Analysis of Construction Operations. Wiley Interscience. P. J. Davies, S. Emmitt, and S. K. Firth, “On-site energy management challenges and opportunities: a contractor’s perspective,” Build. Res. Inf., vol. 41, no. 4, pp. 450–468, Aug. 2013. A. L. Sharrard, H. S. Matthews, and M. Roth, “Environmental Implications of Construction Site Energy Use and Electricity Generation,” J. Constr. Eng. Manag., vol. 133, no. 11, pp. 846–854, 2007. P. O. Akadiri, P. O. Olomolaiye, and E. a. Chinyio, “Multi-criteria evaluation model for the selection of sustainable materials for building projects,” Autom. Constr., vol. 30, pp. 113–125, Mar. 2013. N.-F. Pan, “Fuzzy AHP approach for selecting the suitable bridge construction method,” Autom. Constr., vol. 17, no. 8, pp. 958–965, 2008. A. Ebrahimian, A. Ardeshir, I. Zahedi Rad, and S. H. Ghodsypour, “Urban stormwater construction method selection using a hybrid multi-criteria approach,” Autom. Constr., vol. 58, pp. 118–128, Oct. 2015. B. Reza, R. Sadiq, and K. Hewage, “Sustainability assessment of flooring systems in the city of Tehran: An AHP-based life cycle analysis,” Constr. Build. Mater., vol. 25, no. 4, pp. 2053–2066, 2011. M. Medineckienė, Z. Turskis, and E. K. Zavadskas, “SUSTAINABLE

AN US

[1]

CR IP T

8. References

[12] [13]

[14]

[15]

28

ACCEPTED MANUSCRIPT

[22] [23]

[24]

[25]

[26] [27]

AC

[28]

CR IP T

[21]

AN US

[20]

M

[19]

ED

[18]

PT

[17]

CE

[16]

CONSTRUCTION TAKING INTO ACCOUNT THE BUILDING IMPACT ON THE ENVIRONMENT,” TVARIOJI Statyb. Vertin. Statyb. POVEIKĮ Aplink., vol. 18, no. 2, pp. 118–127, 2010. K. B. M. Errington, A. Ardeshir, N.Mohseni, “Selection of a bridge construction site using fuzzy analytical hierarchy process,” Geogr. Inf. Syst. SpringerArab.J. Sci.Eng., vol. 39, no. 6, pp. 4405–4420, 2014. Nang-fei Pan, “Evaluation of building performance using fuzzy FTA,” J. Constr. Manag. Econ., vol. 24, no. 12, pp. 1241–1252, 2006. Y. Chen, G. E. Okudan, and D. R. Riley, “Sustainable performance criteria for construction method selection in concrete buildings,” Autom. Constr., no. October, 2010. J. Zeng, M. An, and N. J. Smith, “Application of a fuzzy based decision making methodology to construction project risk assessment,” Int. J. Proj. Manag., vol. 25, no. 6, pp. 589–600, 2007. R. Akhavian and A. H. Behzadan, “Evaluation of queuing systems for knowledge-based simulation of construction processes,” Autom. Constr., vol. 47, pp. 37–49, 2014. K. Koo and S. Park, “GA-Based Fuel-Efficient Transfer Path Selection Model for Delivering Construction Materials,” J. Constr. Eng. Manag., vol. 138, no. 6, pp. 725–732, 2012. W. L. Ng, “An efficient and simple model for multiple criteria supplier selection problem,” Eur. J. Oper. Res., vol. 186, no. 3, pp. 1059–1067, May 2008. S. T. Ariaratnam and S. S. Sihabuddin, “COMPARISON OF EMITTED EMISSIONS BETWEEN TRENCHLESS PIPE REPLACEMENT,” J. Green Build., vol. 4, no. 2, pp. 126–140, 2009. A. B. Robertson, F. C. F. Lam, and R. J. Cole, “A Comparative Cradle-to-Gate Life Cycle Assessment of Mid-Rise Office Building Construction Alternatives: Laminated Timber or Reinforced Concrete,” Buildings, vol. 2, no. 4, pp. 245– 270, Jul. 2012. Y. Chen, G. E. Okudan, and D. R. Riley, “Decision support for construction method selection in concrete buildings: Prefabrication adoption and optimization,” Autom. Constr., vol. 19, no. 6, pp. 665–675, Mar. 2010. N. J. B. Idrus, “Construction related factors influencing the choice of concrete floor systems,” Constr. Manag. Econ., vol. 20, no. 1, pp. 13–19, 2002. M. N. Bakht and T. E. El-Diraby, “Synthesis of Decision-Making Research in Construction,” J. Constr. Eng. Manag., vol. 141, no. 9, pp. 1–18, 2015. D. Pandey, M. Agrawal, and J. S. Pandey, “Carbon footprint: current methods of estimation.,” Environ. Monit. Assess., vol. 178, no. 1–4, pp. 135–60, Jul. 2011. EPA, “Energy Star Program, in: Environmental Protection Administration, R.O.C.T. (Ed.).” Institute for Environment and Development Energy Foundation, Taiwan, 1997. Y. Ji, K. Li, G. Liu, A. Shrestha, and J. Jing, “Comparing greenhouse gas emissions of precast in-situ and conventional construction methods,” J. Clean. Prod. J. Chou, K. Yeh, J. Of, and C. Production, “Life cycle carbon dioxide emissions simulation and environmental cost analysis for building construction,” J. Clean. Prod., vol. 101, no. AUGUST, pp. 137–147, 2015. D. Liu and T. J. Stewart, “Integrated object-oriented framework for MCDM and DSS modelling,” Decis. Support Syst., vol. 38, no. 3, pp. 421–434, 2004. S. Biswas and Y. Narahari, “Object oriented modeling and decision support for 29

[29]

[30]

[31]

[32] [33]

ACCEPTED MANUSCRIPT

[40]

[41]

[42]

[43] [44] [45] [46] [47]

[48]

[49]

AC

[50]

CR IP T

[39]

AN US

[38]

M

[37]

ED

[36]

PT

[35]

CE

[34]

supply chains,” Eur. J. Oper. Res., vol. 153, no. 3 SPEC. ISS., pp. 704–726, 2004. S. N. Pillutla and B. N. Nag, “Object-oriented model construction in production scheduling decisions,” vol. 18, pp. 357–375, 1996. A. H. Levis and D. M. Perdu, “Object Oriented Design of Decision Making Organizations,” Fairfax, VA, 1994. G. Booch, “Object-Oriented Development,” IEEE Trans. Softw. Eng., vol. 2, no. 2, pp. 211–221, 1986. J. Seemann and J. W. von Gudenberg, “UML- Unified Modeling Language,” Informatik-Spektrum, vol. 21, no. 2, pp. 89–90, 1998. S. D. Pohekar and M. Ramachandran, “Application of multi-criteria decision making to sustainable energy planning - A review,” Renew. Sustain. Energy Rev., vol. 8, no. 4, pp. 365–381, 2004. E. Triantaphyllou, Multi-Criteria Decision making methods: A Comparative Study, vol. 44. Florida: Springer Science+ Business Media Dordrecht, 2000. W. Rasdorf, M. P. Lewis, S. K. Marshall, I. Arocho, and H. C. Frey, “Evaluation of On-Site Fuel Use and Emissions over the Duration of a Commercial Building Project,” J. Infrastruct. Syst., vol. 18, no. 2, pp. 119–129, 2012. T. Hong, C. Ji, M. Jang, and H. S. Park, “A Conversion Method for Obtaining Emission Data from the Life Cycle Inventory Database of Foreign Countries,” J. Manag. Eng., p. 04014059, Oct. 2013. M. Brunelli and J. Mezei, “How different are ranking methods for fuzzy numbers? A numerical study,” Int. J. Approx. Reason., vol. 54, no. 5, pp. 627– 639, 2013. G. J. Klir and B. Yuan, Fuzzy sets and fuzzy logic : theory and applications. Prentice Hall PTR, 1995. M. Ebrahiminejad, E. Shakeri, and A. Ardeshir, “Emission Estimation from Building Construction in Tehran Metropolitan,” 2014. “Tehran Municipality.” [Online]. Available: statistics.tehran.ir. “Statistics Centre of Iran.” [Online]. Available: amar.gov.ir. G. Ozcan-Deniz and Y. Zhu, “A system dynamics model for construction method selection with sustainability considerations,” J. Clean. Prod., vol. 121, pp. 33–44, May 2016. R. Knauf, A. J. Gonzalez, and T. Abel, “A framework for validation of rulebased systems,” IEEE Trans. Syst. Man, Cybern. Part B Cybern., vol. 32, no. 3, pp. 281–295, 2002. D. E. O’Leary and R. O’Keefe, “Expert system verification and validation,” Artif. Intell. Rev., 1993. M. R. Lynn, “Determination and Quantification of Content Validity,” Nurs. Res., vol. 35, no. 6, pp. 282–314, 1986.

30