Building Information Modeling–Based Analysis to

Minimize Waste Rate of Structural Reinforcement

Atul Porwal1 and Kasun N. Hewage2

Abstract: Building information modeling (BIM) is an emerging tool in architecture/engineering/construction (A/E/C) industry that is used
to design, document, and enhance communication among all the project stakeholders. Trim loss of rebar can be minimized with the use of
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

discrete bars. To achieve this goal, a model to analyze reinforced concrete structure with one-dimensional (1D) cutting waste-optimization
technique, integrated with BIM, is proposed. Building information modeling is selected as the hub in communicating project information
among diverse design teams. This process permits project teams to utilize BIM models to simulate architectural and structural design require-
ments, and compare results speedily to make necessary changes in the designs to minimize rebar waste. The BIM rebar optimization analysis
approach also supports cost-effective decision making during the design process. The proposed approach was validated with a two-story
reinforced concrete structure, and the results indicated a high potential for budgetary savings. The proposed approach is also applicable for
complex reinforced concrete construction projects, with repeated structural elements, and cost saving increases with the increase in the
diameter of rebar. DOI: 10.1061/(ASCE)CO.1943-7862.0000508. © 2012 American Society of Civil Engineers.
CE Database subject headings: Building information models; Structure reinforcement; Construction materials; Communication;
Optimization.
Author keywords: Building information modeling; Waste management; Material management; Communication; Optimization.

Introduction between these two design processes, as architectural design defines

Environmental sustainability has become a key focus for many
industries, and sustainable development practices are well recog-
nized and enforced by all governments (local and federal) in
Canada. The construction industry, like many other industries,
needs to adopt advancements in technologies to overcome adverse
effects of construction waste. Leadership in energy and environ-
mental design (LEED) is one of the most accepted and widely
recognized sustainable building rating systems (Syal et al. 2007).
Construction waste management often focuses on reuse, recycle,
and proper disposal of waste materials at landfills. Waste reduction
at the source is the best and most efficient method for minimizing
waste. It eliminates many of the waste disposal-related problems
(Begum et al. 2006). However, the focus on reduction of waste
at the source was negligible in the past years. Waste reduction
at the source may still be the best approach toward minimizing
the intensity of the construction waste problem.
Traditionally, building design process consists of two different
objectives, i.e., architectural and structural designs. Architectural
designing focuses on planning, designing spaces, and ambience
that consider functional, technical, social, and aesthetic aspects
of the utility, whereas structural designing aims on the mechanics
of the structural elements of a building. There is a close relationship
1
Ph.D. Candidate, School of Engineering, Univ. of British Columbia–
Okanagan Campus, BC, Canada. E-mail: Atul.Porwal@ubc.ca
2 structural modeling software programs, it was noted that these pro-
Assistant Professor, School of Engineering, Univ. of British
Columbia–Okanagan Campus, BC, Canada (corresponding author).
E-mail: Kasun.Hewage@ubc.ca
Note. This manuscript was submitted on April 12, 2011; approved on
November 2, 2011; published online on November 4, 2011. Discussion
period open until January 1, 2013; separate discussions must be submitted
for individual papers. This paper is part of the Journal of Construction
Engineering and Management, Vol. 138, No. 8, August 1, 2012. ©ASCE,
ISSN 0733-9364/2012/8-943–954/$25.00.
the geometrics of the building elements and that information
becomes key input in structural designs. On construction projects,
reinforcement bars are generally purchased in standard lengths and
required bar lengths are normally shorter than the standard lengths
with large length variations. The success of a sustainable design
process thus becomes highly dependent on efficient analysis among
the two diverse design teams, to optimize trim loss in structural
elements of a building. The coordinated modeling of these two de-
sign processes typically does not occur in the early design stage.
Poor integration leads to uneconomical design and higher material
waste (Chen et al. 2005). This difficulty is primarily because of lack
of coordination and modifications in building models to cater ar-
chitectural and structural needs. Building information modeling
(BIM) makes it possible to coordinate among stakeholders in
the building design process to explore different design alternatives
more efficiently by avoiding the time-consuming and error-prone
method of recreating all the building geometry attributable to a
change in structural analysis and design. This challenge can be
overcome by creating structural and architectural objects from
the same underlying database. It facilitates faster repetitive modi-
fications between the architectural and structural designs, and fur-
ther analysis for rebar trim losses with an optimization tool. This
approach enables a design team to recursively modify the building
model to analyze best cutting combinations of the available market
lengths and special lengths of rebar. While exploring different BIM
grams have limitations in producing abstract of rebar schedule with
a count of bar for a length of specific type (diameter, for example).
Further, to obtain an accurate rebar schedule, reinforcement place-
ment in the BIM model has to exactly match with the actual field
work. For example, if trim loss is calculated by considering con-
tinuous beam over multiple spans, then the main reinforcement bar
should also be laid as a continuous member in the model.

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012 / 943

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 This paper presents a case study that investigates rebar trim-loss
optimization process using an optimization algorithm with struc-
tural BIM model. Algorithm makes it possible to use leftover rebar
cutoff lengths available at site in combination with the market
length and special rebar length.
Background
The U.S. Green Building Council (USGBC) LEED reference
guide states that a typical North American commercial construc-
tion project generates 12 kg of solid waste per square meter of
floor space (USGBC 2011). According to the report on waste
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.
management for construction industry by the Canadian Con-
struction Association (CCA), 8.1% of total construction waste
generated consists of metal whereas 5.2% metallic waste is gen-
erated from demolition wastes (CCA 1992). Metro Vancouver,
British Columbia also estimated the metallic construction waste
as 0:09 t∕1;000 sq ft for high-rise buildings and up to
0:21 t∕1;000 sq ft for institutional low-rise constructions (Metro
Vancouver Construction Waste Management Disposal 2008).
Poon et al. (2004) found that out of total construction waste,
reinforcement steel waste was 3–5% in public housing and up
to 8% in private residential construction as metallic waste. This
metallic waste has recycling potentials; however, with an addi-
tional cost. In addition, rebar waste has a direct effect on the
project cost (Salem et al. 2007) and is much higher if construc-
tion type is a reinforced concrete frame structure, involving ma-
jor structural parts of the building as reinforced members.
In case when rebar shaping is done using computerized numeri-
cally controlled (CNC) machines, for rebar supplied in coils, the
process produces few scraps with almost 0% trim loss. However,
CNC can be used for bar shaping up to a bar diameter of 16 mm
(Kim et al. 2004). Generation of waste is inevitable when rebar is
supplied in straight market lengths for on-site fabrication. Rebar is
produced in Canada in accordance with the National Standard of
Canada CAN/CSA-G30.18-M92 for concrete reinforcement (CSA
2000a). The national rebar standard is approved by the Standards
Council of Canada. The most common bar designation numbers,
with the corresponding diameter in millimeters in brackets, used
in Canada are 10 (11.3), 15 (16.0), 20 (19.5), 25 (25.2), 30
(29.9), 35 (35.7), 45 (43.7), and 55 (56.4) (Canada Border Service
Agency 2011). Rebar sizes are commonly referred to as the bar
designation number combined with the letter “M.” Thus, 10 M
referred to a bar designation number of 10 and a diameter of
11.3 mm. The standard lengths for rebar available in the market
are 6 m (20 ft), 12 m (40 ft), and 18 m (60 ft). Residential markets
primarily use rebar of smaller diameter, whereas complex construc-
tion and fabrication markets use most of the larger sizes of rebar
(Canada Border Service Agency 2011). The standard market
lengths produce relatively more scraps after rebar is cut off as
per the design schedule. Percentage cost ratio of waste increases
with the increase in diameter of the bar, even for a small trim length
(Kim 2002).
Interoperability
Interoperability is the ability to manage and communicate elec-
tronic product and project data among collaborating firms. The
construction industry consists of most fragmented supply chain
of the architecture/engineering/construction (A/E/C) industry.
Interoperability is an important factor, as reentering data from a
BIM into another application creates wasteful and costly duplica-
tion. Interoperability between various software applications can
944 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012
J. Constr. Eng. Manage., 2012, 138(8): 943-954
be achieved in a number of ways; three of the most common
are (buildingSMARTAlliance 2011)
1. Software applications developed by one software vendor that
directly reads the appropriate file format contained in the a
BIM software application;
2. Using a BIM software that incorporates an application pro-
gramming interface (API), providing a well-developed inter-
face between software from different providers; and
3. Using a software that supports data exchange standards having
industrywide acceptance.
Data Exchange Standards
The steel industry’s CIMSteel Integration Standards (CIS/2) [Steel
Construction Institute (SCI) 2003] is an example of a successful
application of a data exchange standard. The International Alliance
for Interoperability (IAI) was established in 1995 by American and
European A/E/C firms to promote interoperability in the industry,
and successfully pioneered a technical effort across 19 countries to
define a single building model as semantic definition of building
elements, their properties, and relationship as industry foundation
class (IFC) model. This is now endorsed as a draft ISO standard
(buildingSMARTAlliance 2011).
Optimization Algorithms
In the past, optimization of rebar trim loss has been addressed as
one-dimensional cutting stock problem (1D-CSP), and several
algorithm-based solutions have been proposed. At present, CNC
machines use tools that are developed on the basis of these algo-
rithms, to produce the required cut lengths of rebar in the shop. The
following are the widely used algorithmic approaches.
Linear Programming (LP)
This approach was first used by Gilmore and Gomory (1961) to
solve the trim-loss problem by generating cutting patterns. A LP
problem may be defined as the problem of maximizing or minimiz-
ing a linear function subject to linear constraints. The constraints
may be equalities or inequalities. For a simple example to find
numbers x1 and x2 that maximize the sum x1 þ x2 subject to the
constraints x1 ≥ 0, x2 ≥ 0, and x1 þ 2x2 ≤ 4; 4x1 þ 2x2 ≤ 12;

x1 þ x2 ≤ 1.
The function to be maximized (or minimized) is called the ob-
jective function. In the example, the objective function is x1 þ x2
(e.g., combination of rebar piece lengths).
The solution to the problem is composed of two steps (Fig. 1).
The first is the search for a set of feasible combinations to be used in
the second step, which is the optional choice of the number of these
combinations needed to satisfy the requested number of pieces.
The idea is simply to compile a large array of piece lengths in a
certain order and use this array to choose feasible combinations.
This choice is carried out by using the array to add one length
to the next until the accumulated length exceeds the maximum stan-
dard length. Then, the number of pieces in the combination is
obtained by excluding the last piece.
Integer Programming (IP)
The first step to solve this problem is to generate all feasible cutting
patterns. The procedure adapted from Pierce and Lasky (1973) can
be used to generate all the efficient feasible cutting patterns. The
next step after generating the cutting pattern is to formulate the IP
model as follows (Goulimis 1990):
1. Decision variables: The next step is to assign a decision vari-
able for each pattern that is equal to the number of times this
pattern will be used in the final solution. At the end of the so-
lution, the final values of decision variables would tell how
 operate on the series methods, there are heuristics that can provide approximate
in sequence solutions to some optimization problems. Gradisar et al. (1999)
proposed an SHP for optimizing 1D stock cutting when all stock
lengths are different. The proposed algorithm was developed on a
step-by-step basis. The number of basic steps equals to the number
add the next length of stock lengths necessary for fulfillment of an order. Initially, all
Outer
in the series stock lengths belong to the set of unprocessed stock lengths. The
Loop
set of processed stock lengths is empty. At each step, the set of
Inner unprocessed stock lengths is reduced by one and the set of proc-
Loop essed stock length increases by one. Also, the number of cut pieces
of particular order lengths changes, and the length of the processed
decrement series stock length, which becomes equal to trim loss. Algorithm has the
NO if accumulated YES by one and print
length > following steps:
feasibility
1. Select order lengths.
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

standard length
combination
Fig. 1. LP combination generating algorithm for 1D-CSP
many times each of the generated patterns should be used to
give the minimum possible waste.
The decision variables are denoted as X i , where X i = num-
ber of times pattern number i is used.
For example, if in the final solution the value of X 4 ¼ 5 and
the value of X 9 ¼ 8, the cuts that would minimize the gener-
ated waste are to use pattern No. 4 a total of five times, and
pattern No. 9 a total of eight times.
2. Objective function: The objective is to minimize the trim loss,
which could be written as [Eq. (1)]
! sible to examine each one of them to find an optimal solution.
X I X
I
Minimize Lsi X i
Ci X i ð1Þ
i¼1 i¼1
where I = total number of patterns generated; C i = utilized
length of pattern number I; and Lsi = standard length of pattern
number i is cut.
The utilized length of a pattern is the actual useful length of
the pattern in which any excess or unused length would end up
as scrap when using this pattern in the final solution. Thus, if
for example, the pattern that has a composition of 2–3 m (uti-
lized length ¼ 5 m) is used three times in the final solution and
if the standard length used is 6.10 m, then the trim loss would
be computed as ð6:10
5Þ  3 ¼ 3:3 m that would probably
end up as scrap.
3. Constraints: After setting up the objective function, some con-
straints must be fulfilled. The constraints are simply to satisfy
the demand of each length needed, which could be formulated
by setting up N constraints (one for each demanded length) as
follows [Eq. (2)]
! ics, specifically with the way that liquids freeze and crystallize, or met-
X I
Subject to Aij X i ¼ d j for ðj ¼ 1
NÞ ð2Þ
i¼1
where Aij = number of bars of length lj that are present in pat-
tern number i; and d j = number of bars of length lj that are
needed to satisfy the demand.
Additional constraints should be set up to ensure that all the
decision variables X i are nonnegative, and integers X i ≥ 0 and
integer i ¼ 1
I.
Sequential Heuristic Procedure (SHP)
Haessler (1975) and many other solved CSP using SHP approach.
Beside (finitely terminating) algorithms and (convergent) iterative
2. Select stock length and cut it with chosen order lengths.
3. If all stock lengths are not cut yet and the requirements for
order lengths are not fulfilled, then go back to step 1, else stop.
Genetic Algorithm (GA)
The GA model was developed by Salem et al. (2007), and they
compared it with the LP and IP models. Genetic algorithm is an
optimization algorithm developed by Holland (1970) on the basis
of the theories of genetics and natural selection. In GA,
a population of strings (called chromosomes), which encode can-
didate solutions (called individuals) to an optimization problem,
evolves toward better solutions. Traditionally, solutions are repre-
sented in binary as strings of 0s and 1s. Genetic algorithms have
been used to find optimal solutions for a wide spectrum of complex
combinatorial problems in civil engineering in which the possibility
of a huge number of combinations or alternatives makes it infea-
Binary Search Algorithm
Kim et al. (2004) used a binary search algorithm for rebar combi-
nation on the basis of reading rebar data file generated by automatic
rebar detailing algorithm. They extracted rebar geometry data from
structural design data file to estimate precise cutting lengths and
quantities.
In computer science, a binary search or half-interval search
algorithm locates the position of an item in a sorted array binary
search works by comparing an input value to the middle element of
the array. The comparison determines whether the element equals
the input, less than the input, or greater. When the element being
compared to equals the input, the search stops and typically returns
the position of the element.
Simulated Annealing Approach
The method of simulated annealing (Kirkpatrick et al. 1983) is a tech-
nique for optimization problems of large scale, especially ones in
which a desired global extreme is hidden among many, poorer, local
extreme. At the heart of the method is an analogy with thermodynam-
als cool and anneal. Metropolis et al. (1953) first incorporated
annealing principles into numerical calculations [Eq. (3)]. They as-
sumed a simulated thermodynamic system to change its configuration
from energy E 1 to energy E2 with probability
Pr ¼ exp½
ðE2
E1 Þ∕kT ð3Þ
Quantity k is Boltzmann’s constant of nature that relates temper-
ature to energy.
The following elements are provided to make use of the
Metropolis algorithm:
1. A description of possible system configurations;
2. A generator of random changes in the configuration, and these
changes are the “options” presented to the system;

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012 / 945

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 3. An objective function E (analog of energy) whose minimiza-
tion is the goal of the procedure; and
4. A control parameter T (analog of temperature) and an annealing
schedule that tells how it is lowered from high to low values.
An objective function E is created to minimize rebar waste in 1D
optimization algorithm.
Algorithm Comparison
There are multiple combinations of different patterns, and the
solution may become computationally intractable. It is difficult
to reliably state which of the previously discussed methods is best
for which problem.
The LP technique is, although very efficient, still might not
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.
made anywhere in the model updated all the time, which makes
reanalysis and revisions of designs fast and simple. Because BIM
represents virtual true space, its clash detection process can check
intersecting volumes. A four-dimensional (4D) BIM scheduling
application can link critical path method (CPM) schedule activities
to 3D objects. Cost data can be associated with each element,
resulting in a detailed cost schedule (Popov et al. 2010). Three-
dimensional objects can also be linked to a variety of source docu-
ments through hyperlinks that enable the model to function as
graphical information system (GIS) for the building (Isikdag et al.
2008; Li et al. 2005). Process simulation creates a virtual feedback
loop such that design and construction coordination challenges and
sequence can be identified before commitment of field process.

result in the optimal amount waste attributable to rounding the re-

laxed fractional solutions to integer values. As a result, the next step
Proposed Methodology
would be to use IP techniques to reach an optimal integer solution.
This optimization approach is not widely used because of its com- The virtual environment of BIM enables identification and imple-
plexity and higher computational efforts (Goulimis 1990). When mentation of desired changes, through the increased visibility and
using IP techniques such as “branch and bound,” an excessive com-
putational effort is required when the number of cutting patterns
becomes too large. The approach is based on the principle that
the total set of feasible solutions can be partitioned into smaller
subsets of solutions. These smaller subsets are then evaluated sys-
tematically until the best solution is found. In case of reinforcement
cutting stock problem, for example the supply length of bar is
12.20 m and 500 different lengths ranging from 0.5 to 10.0 m
are to be cut, then the number of different patterns may easily
exceed 10 million or even 100 million. In that case, the solution
may not be tractable. Genetic algorithms are simple to implement,
but their behavior is difficult to understand, specifically for large
combinatorial problems. Binary search algorithms are generally Fig. 2. BIM models: (a) architectural model; (b) structural model
suitable for large combinatorial problems because of more compu-
tational time. However, it is possible to use the simulated annealing
approach method that consistently produces better results in case of Owner
large-scale optimization problems. The results (Jahromi et al. 2011)
show the efficiency of the proposed SA method in finding cutting Architect

plans with good average values of trim losses for randomly gen- Fabricator
erated problems instead of high computational times for finding Architectural
BIM model Structural
global optimum with the assumptions that (1) all used stock lengths drawings
must be cut to the end in as much as it is possible; and (2) all stock
lengths are identical, and there is no difference between the lengths Final
of them. BIM model Structural
BIM model

Building Information Modeling as a Coordinated Construction

Design Analysis Tool Structural Analysis document
and Design (STAAD) Bar bending
schedule Civil Engineer
Building designers consider reanalysis and revising the designs as a
time-consuming process. Most of the time, it leads to a model that
might be impressive in architectural aspects; however, lack in in-
tegration, coordination, and collaboration between the various Rebar
Analysis
functional disciplines of a project. Building information modeling
is a richer repository than a set of drawings, capable to store
objects, spaces, and facility characteristics in a digital database
(Eastman 1999), and helps achieve levels of coordination far in ex- Design with feedback
cess of the current norms (Schlueter and Thesseling 2009). These from structural engineer
BIM-enabled qualitative and quantitative analyses significantly Multidisciplinary
enhance the efficiency of structural design, energy consumption, coordination
and other simulations of the building process. It has a superior
Design with
three-dimensional (3D) modeling environment compared with architect revision
two-dimensional (2D) computer-aided design (CAD) (Fu et al.
2006), and the ability to parametrically capture design intent that
reduces the likelihood of coordination errors and rapidly perform
Fig. 3. Workflow of early rebar waste-optimization approach
iterative changes. The parametric modeling property keeps changes

946 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 H
PB3

1800

PB14

PB14
PB4 PB4
G

1500 PB3

PB11X
F

PB12

PB14
PB5
1850
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

PB6

PB14

PB4
E
630 PB3 PB7
D
100 THK. SLAB
PB11

WITH 8 Ø C/C B/W

PB13
2110

PB13
PB8 PB4 PB4 +600
C
1730

+450
PB4

PB3
B
BR1
400

PB8
A
+300
PB4

930 1430 760 1470 4460

1 2 3 4 5 6

Fig. 4. Structural arrangement of plinth beams and columns for initial analysis

predictability, earlier in the design process. This in effect encour-

Initial structural Analysis F ages experimentation and collaboration to simulate free exchange
E
E
of ideas to exercise better control over the project cost, and to
Preparation of
Structural BIM Model D achieve significant improvements to reduce waste at source.
B Fig. 2 shows the structural model created out of the underlying
A architectural BIM database without putting any extra efforts.
Initial Structural Structural
C Traditionally, a structural engineer starts the design process by
Analysis and Design member
(STAAD) optimization K
interpreting architectural drawings, as a separate step, with focus on
T Final analyzing the mechanical properties of building elements and struc-
O
Trim Loss Analysis BIM ture. The analytical models for structural analysis are then repro-
model duced with respect to general framing layout, and section properties
For Rebar Market Length Used
A suggested in architectural designs.
R In the proposed approach, structural members of the building
For Special Rebar Length Used C
H are created out of the same architectural model by assigning struc-
For Structural Member I tural property to the members, and then analyzed for trim-loss
Dimensions modification T optimization. The engineer shares the same underlying building da-
E tabase for structural design to make changes quickly. Design teams
C
Integration with MEP, HVAC T explore design changes, develop, and study several design alterna-
tives to make key design decisions. Each option is substituted into
the model for visualization and quantity takeoffs to help the
Fig. 5. Steps for early rebar waste optimization
designer work with rebar optimization analysis. Fig. 3 shows

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012 / 947

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 the writers’ suggested approach of collaborative building design
with the BIM models. First, the architect proposes the building
design concept on the basis of the requirements of the owner, build-
ing bylaws, and code of practice. An integrated BIM model is then
created by incorporating structural elements (i.e., beams and col-
umn positions). An interference check of structural and architec-
tural objects is performed before structural analysis (integration
of building spaces and structural elements). The structural elements
are then checked for possible modifications in the dimension of the
members and reduction in the total number of structural members,
such as column, beam, and footing. The architect then reviews and
incorporates the changes in the model. Structural analysis is then
performed on the same model database and a structural BIM model
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.
If 6.10-m market length is used, then 2 × 3 ¼ 6 bar of length
3.86 m will only be required as cutoff length.
2. Special ordered length: Then, similar analysis (as of step 1) is
performed for the special rebar lengths available with the man-
ufacturer. Percentage waste is calculated for combinations of
standard market length and special lengths.
3. Modify the design: As a third step, the designer reviews the
geometry of the structural members. For example, continuous
beams may result in less number of cutoff lengths. Similarly,
repeated structural members (such as beam repeated in an
apartment for 10 apartments on a floor in a 20-floor construc-
tion) are checked for rebar waste optimization. Structural ar-
rangement of column footings is also reviewed for any
special structural needs, such as combined footings that need

is developed. The detailed bar bending schedule generated by BIM

is then used for rebar optimization analysis, with 1D stock optimi-
zation tool, to identify possible trim losses related to available
market lengths. In this research, 1DNest software, developed by
using a mix of simulated annealing heuristic algorithm, was used
to analyze best combination rebar lengths.
Under the proposed approach, rebar waste analysis should be
conducted by the design engineer in three steps:
1. Available market length: First, a detailed structural BIM model
is created by considering the reinforcement bar lengths avail-
able in the market. Building information modeling facilitates
the designer to detail bar placement with splices, hooks, and
bends in a 3D environment. For example, beam lengths along
5 and 6 for lengths A to H [Fig. 4] are provided with two
numbers of 20-M bars at the bottom for three beam lengths.
more reinforcement than usual.
In addition, the design engineer should also check for any minor
modifications in the arrangement of spaces and facility, by which
material requirement for construction can be reduced. The architect
may work on the model on the basis of feedback from mechanical/
electrical/plumbing (MEP) and HVAC design teams. This saves
repetitive modeling and designing tasks among the design teams,
and avoids errors attributable to manual coordination between struc-
tural engineers, architects, and drafters (Geodert and Meadati 2008).
The previously discussed steps are repeated until the design
team achieves the final optimum building model. Finally, structural
drawings for fabricator and construction documents for the engi-
neers are generated from the final building model for the use at
the construction site.

H H
PB3
2280
2280

OPEN

PB14
W1 BEDROOM

PB14
4490X2080
4260X3200

W4 V PB4 PB4
G D3 G
920
920

F PB3
TOILET F
D2
PB14

PB14
KITCHEN 1220X2130
PB11X
1289

PB12

2240X3150 D3 STORE
D3
PB5 PB6
2480

E
E
PB4
1191

PB7
D2
D SLIDING DOOR D
PB3 100 THK. SLAB WITH
8 Ø C/C B/W
PB13

ARCH OOPENING
PB11

DINING
2690

2690

PB13

3460X2740
DRAWING
4260X3840
W1 +600
PB8 PB4 PB4
C C +600
1050

1050

D1
+450 +450
PB3
PB4

B W3 B
1080

BR1
1080

PORCH PB8
A A
+300

+300
PB4
1600

930 1430 760 1470 4460

930 1430 760 1470 4460

6 1 2 3 4 5 6
1 2 3 4 5
(a) (b)

Fig. 6. (a) Revised building plan; (b) revised plinth beam plan

948 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012

J. Constr. Eng. Manage., 2012, 138(8): 943-954

Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Screenshot of structural BIM for plinth beam and column

Case Study
A case study to explore rebar trim-loss optimization was conducted
on a four-unit multifamily, semidetached two-story building of
reinforced concrete structure.
As shown in Fig. 5, rebar waste-optimization process has
three steps.
1. The first step was to prepare the structural BIM model and
analyze it for possible reduction in number and size of struc-
tural members that resulted in minimum quantity of material
(steel and concrete) required in the construction.
2. The second step was to analyze the trim-loss percent for market
rebar lengths, special rebar lengths, and adjustment of dimen-
sions of the structural members.
3. Finally, the integrated optimal BIM model was prepared. This
model was further checked for interference collision for build-
ing elements with MEP and HVAC elements.
During the process in each step, the structural engineer commu-
nicated his feedback to the architect through the BIM model, and
in turn the architect quickly submitted his feedback with changes in
the BIM model. The initial proposed building plan as per the owner
needs is as shown in Fig. 4. This building plan resulted in repetitive

Fabrication
Bar bending Structural
schedule Drawings
BIM model

Final
MS Excel model
data file Market
Length

Minimum
1D Loss Cutofff waste
Optimizer Analysis Length ratio

Special
Length

Fig. 8. Structural modeling for trim-loss optimization

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012 / 949

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 L
TOP BARS (THROUGH)
0.25 L 0.3L
EXTRA AT EXTRA AT
LEFT SUPP RIGHT SUPP

EXTRA AT MIDDLE
0.1 L 0.15L
BOTTOM BARS (THROUGH)

STIRRUPS NEAR SUPP STIRRUPS MIDDLE STIRRUPS NEAR SUPP

2d 2d
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 9. Reinforcement detailing of plinth beams

beam layout for plinth, ground-floor roof, and first-floor roof lev- bearing area, created a combined footing for columns at A4, C5,
els. An irregular shift in the four units was proposed to provide and B4. This needed higher depth and reinforcement compared
aesthetic beauty to the building and to obstruct vision through win- with isolated footing, and led to uneconomical section (Bangash
dow [e.g., kitchen window W4, Fig.6(a)]. 2003). The savings in steel and concrete quantity was achieved
by shifting the column positions to get an unoverlapped bearing
Initial Structural Analysis area of footings. The feedback was then reanalyzed in the architec-
The initial layout plan of the reinforced concrete frame structure tural BIM model and modified the building plan by arranging main
building described in this paper was prepared using traditional door entry to the west instead from the south, omitting one column
2D CAD. Structural analysis and design process was performed C4 out of three columns, which created combined footing. The
after the finalization of all architectural design works and comple- BIM’s ability to coordinate changes and maintain consistency,
tion of the bidding process. In this approach, the BIM model was called bidirectional associativity, updated the structural model
prepared using Autodesk Revit Architecture because of its interop- components automatically.
erability with existing data with AutoCAD. A structural model was
created from the underlying BIM database using Autodesk Revit Trim-Loss Analysis
Structure. The next step in the process was to reimport updated analytical
The initially suggested architectural plan by the architect model from architectural BIM, with the architect’s feedback, into
resulted in the following structural layout (Fig. 4): the structural BIM (Revit structures) and analyze it for structural
• Column and beam arrangement, repeated for four units; requirements. After the structural design, a detailed reinforcement
• Columns shown in hatch continued until ground-floor roof model of the structure in BIM was developed (Fig. 7). The rebar
level, whereas columns shown in dark continued to the schedule created automatically in BIM was then exported to MS
second-floor roof level;
Excel for trim-loss optimization process. Fig. 8 shows the process
• Columns around the stairwell went to even higher level and sup-
of optimizing BIM model for rebar waste. Trim-loss quantity of
ported the overhead water tank above the second-floor roof level; and
• Plinth beams were to support the load of 200-mm-thick walls
and plinth slab. Table 1. Rebar Schedule for Plinth Beam Plan
On the basis of this architectural model, the first structural
10-M bar 15-M bar 20-M bar
model was developed with the following features:
• Discontinuous beams (along grid 5, 6) with the beams resting over Bar length Bar length Bar length
other beams creating point loads on the main beams (PB5, PB14), (mm) Count (mm) Count (mm) Count
• Depth and reinforcement of beam PB3 were more attributable to 763 4 1,395 6 865 4
increased bending moment and shear force (attributable to point 950 1,016 1,540 10 1,640 8
load from beam PB14),
975 4 1,640 8 1,736 6
• Size and reinforcement of beam PB14 could not be reduced
1,150 545 1,754 6 2,240 6
because of minimum requirements of structural code of practice,
and 1,847 4 1,770 16 2,516 24
• Rebar available in 6.1- and 12.2-m market lengths could not be 1,830 4 2,756 21
used as single bar in most of the beams (for top and bottom 2,066 6 3,195 7
reinforcement) because of shorter beam lengths. This further 2,288 4 3,740 6
increased the number of 90° hook length at the ends of beams 2,400 4 3,860 6
and rebar trim loss. 3,860 6 4,070 10
With the structural engineer’s feedbacks, a new architectural 4,070 10 4,120 8
model was proposed to address these rebar waste issues shown 4,120 8 5,890 4
in Fig. 6(a), and a structural BIM model was modified and analyzed 4,610 24 6,100 44
for the feedback and architectural modifications [Fig. 6(b)].
5,890 4
As a next step, structural analysis was then performed and
6,100 44
column footing resulted in cluster of footings with overlapped soil

950 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 rebar for available market lengths and special lengths used were
calculated using 1D Cutting Optimizer software. Alternatives to
optimize waste were worked out and further checked for possible
changes in the layout of the structural members and their longitu-
dinal arrangements. The whole analysis was performed quickly the
same day.
As the last step, a structural BIM model, which generated mini-
mum waste ratio, was proposed and fabrication drawings were
created from the latest structural model to work at site.
Rebar optimization analysis was performed for each potential
construction phase (plinth, ground-floor, first-floor phase) in the
sequence of the construction work to be carried out. This facilitated
coordinated purchase of the rebar and planned inventory manage-
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.
for beam reinforcement depending on bar types (bent bar, straight
bar) and number of spans (single span and multiple). In this case
study, trim-loss analysis was done for plinth beams to test the fea-
sibility of the proposed approach. After working on the structural
feedbacks, a final building model was proposed and optimization
analysis was performed. The general reinforcement arrangement
for the beams is shown in Fig. 9.
Main bar sizes used in this case study were 15 and 20 M. The
10-M bars were used for stirrups throughout. Comparison of trim
loss for each design alternative is made for the conditions when
beam main reinforcement is placed (1) column to column and
(2) continuous over supports. For high yield strength deformed
bars, 90° hooks were considered at the ends, and basic development
length [Eq. (4)] was calculated as per structural detailing code of

ment in different stages of construction. A structural BIM model practice of the CSA (CSA A23.1) (CSA/National Standard of
was created to generate rebar schedule for the following stages: Canada 2000b) using the relation
1. Footings and columns up to plinth height,
2. Plinth beam and plinth slab, p
3. Ground-floor columns and stair, lhb ¼ 100 d b ∕ f 0c ð4Þ
4. Ground-floor roof beams and roof slab,
5. First-floor column and stair, and where db = diameter of bar; and f 0c = compressive strength of
6. First-floor roof beams and roof slab. concrete.
Each structural member needs different rebar detailing depend- Length of splices was taken as 40 times the diameter of bar.
ing on the rebar placement arrangement and type of the structural For the case when beams are considered as continuous over
member. For example, column reinforcement has fixed length support, the number of rebar lengths [Eqs. (5)–(7)] required was
according to story height while many alternatives were available calculated by the relations (Table 1)

end bar length = 800+2760+300

= 3860 mm

300 300

5000 4500 4500 2760

800 800 800
(9420x2 + 160x2)
= 19160

Fig. 10. Calculation of rebar schedule quantities

Table 2. Trim-Loss Waste Produced When 6,100-mm Market Length Was Used
10 M 15 M 20 M
Waste Waste Waste
Bar Quantity Cutting pattern (mm) Quantity Cutting pattern (mm) Quantity Cutting pattern (mm)
B01 2 1,847 3 × 1;150 2 × 763 40 44 6,100 0 44 6,100 0
B02 2 1,847 2 × 1;150 2 × 975 3 4 5,890 210 4 5,890 210
B03 5 5 × 1;150 350 6 4,610 1,395 95 6 4,120 1,736 244
B04 254 2 × 1;150 4 × 950 0 18 4,610 1,490 2 4,120 865 865 250
B05 1 2 × 1;150 2 × 763 2,274 8 4,120 1,770 210 8 4,070 1,640 390
B06 4 4,070 1,830 200 2 4,070 2,030
B07 6 4,070 1,770 260 6 3,860 2,240 0
B08 6 3,860 2,066 174 6 3,740 2,360
B09 1 2,400 1,770 1,770 160 7 3,195 2,756 149
B10 3 2,400 1,754 1,754 192 7 2,756 2,756 588
B11 4 2,288 1,640 1,640 532 12 2,516 2,516 1,068
B12 3 1,540 1,540 1,540 1,480
B13 1 1,540 4,560
Total 264 108 104
Material used: 1,606.3 m Material used: 613.6 m Material used: 592.3 m
Material to order: 1,610.4 m Material to order: 658.8 m Material to order: 634.4 m
Waste: 4.10 m (0.256%) Waste: 45.2 m (7.36%) Waste: 42.1 m (7.11%)

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012 / 951

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 Waste (mm)
L ¼ S
2c
2h ð5Þ

0
0
220
150
270
18
72

21
3
16
1,525
2,615
N ¼ L∕l ð6Þ

2 × 865
Le ¼ ðN
iÞ × l þ ði × sÞ ð7Þ

1,736
where L = total lengths required; Le = cutoff length at the end;

2 × 2;756 2 × 1;736
N = number of market lengths required; l = market length of rebar;
2,240

3,860
3,740
2,516

2,240
1,640
h = hook length (i.e., lhb ); lhb = basic hook length; c = clear cover to

Material to order: 597.8 M

Cutting pattern

Material used: 592.3 M

Waste: 5.5 m (0.93%)

reinforcement; S = beam length (continuous or support to support);
20 M

s = splice length; and i = integer value of L.

3 × 2;756

2 × 2;516
2 × 2;516
Structural BIM is capable of handling most of the code of prac-
4,070
3,740
4,070
4,070
3,860

3,195 tice in use. Table 1 shows the results of calculations done by using
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

Eqs. (5)–(7) for the plinth beam plan. Calculation of rebar quan-
tities for beams PB13 and PB14 at grids 5 and 6 is shown in Fig. 10.
2 × 6;100

2 × 4;120

2 × 3;740
3 × 3;195

2 × 2;756
Hook length as per CSA A23.1 (CSA 2000b) was calculated as
5,890

4,120
4,120
4,070
3,860

3,195
3,195
300 mm and splice length as 800 mm. Total length of 20-M main
bars came to 19,160 mm for which three lengths of 6,100 mm and
end bar length of 3,860 mm were required. Total number of bar
Waste (mm) Quantity

lengths required is shown in Table 1.

22
4
3
1
1
4
1
1
1
1
2
8

49
Considering the practical aspect of laying reinforcement cage, it
was assumed that 20 ft (6,100 mm) market lengths are used at site,
and trim-loss comparison was compared for all the available market
6,084
0
60
160
9
0
12
190
10
2
200
232
430

lengths of 12,200 and 18,300 mm. Table 1 shows the rebar schedule
obtained on the basis of the calculations as per Eqs. (5)–(7), and
created in the BIM model accordingly. Rebar lengths of 20-M bar
type and their total count used were extracted from the BIM model
1,395
1,640

1,640
1,640

for optimization of trim loss when 300 pieces of 12,200-mm-long

rebar purchased from the market. Cost-/Meter length was used for
2 × 1;754
2 × 1;770
2 × 1;770
2 × 1;830

assigning weight to the source bar when different source rebar

1,640
1,540
2,066
1,830

1,540

lengths were used. For example, when cutoff length (rebar from
Material to order: 622.2 M
Cutting pattern

Material used: 613.6 M

waste) available at site was used, it was assigned a “cost” value

Waste: 8.6 m (1.40%)
15 M

less than 100 (say 75) to give less priority to this cut length to
5 × 1;540

be used in optimization calculations and not to generate another

4,610
4,610
4,120
4,120
4,070
3,860
2,400
2,288
3,860
2,288

1,540

waste piece from a fresh length. Cut allowance for cutter blade
thickness was assumed negligible compared with total length of
the bar. Trim length longer than 300 mm was assumed to be reused
2 × 6;100

2 × 4;070
2 × 4;070

2 × 2;288
5,890
5,890
4,610
4,610
4,610
4,610
4,610
4,610

4,070

in further construction works at the site (known as cutoff length).

Table 3. Trim-Loss Waste Produced When 12,200-mm Market Length Was Used

Tables 2 and 3 show the results of cutting patterns for 6,100 mm

(20 ft) and 12,200 mm (40 ft) rebar lengths, and 10, 15, and 20 M
Waste (mm) Quantity

type bars, which are most frequently used in the project. Table 2 is
22
1
3
6
2
3
4
4
1
2
1
1
1
51

combination list for 6,100-mm bar length. Rebar combination gen-

erates a maximum material loss of 0.256, 7.36, and 7.11% for 10,
15 and 20 M bars, respectively. Table 3 shows maximum 0.256,
1.4, and 0.93% material loss with 12,200-mm-long rebar are used,
3,774
6
3
324

respectively, for 10, 15 and 20-M bars. All the percent material loss
includes reusable cutoff lengths greater than 300-mm long also.
These lengths will not be recycled and will be reused in the work.
Material to order: 1,610.4 M
2 × 1;847 4 × 1;150 4 × 975

Fig. 11 shows the comparison of estimated trim loss of the three

Material used: 1,606.3 M

Waste: 4.1 m (0.256%)

rebar market lengths used in the optimization process.

The dominant length in the rebar fabrication of plinth beam is
Cutting pattern

12,200 mm for all the three bar types. The 10-M bar has no effect of
9 × 1;150
9 × 1;150 2 × 763
6 × 1;150 2 × 763
4 × 1;150 8 × 950
10 M

length variations because of less variety and smaller lengths for

major quantity used as stirrups. A significant variation in waste
with length is observed for the 20-M bar because of the variety of
length combinations used as extra negative and positive reinforce-
1,847

ment at supports and midspans. Therefore, further optimization is

possible in case of 20-M type bar with the use of special lengths
and available cutoff lengths. For example, 20-M rebar waste of
Quantity

5.5 m reduces to 1.36 m (5:5
1:525
2:615) when 1.525- and

1
2
1
1
127

132

2.615-m-long trim lengths are reused in further construction works.

This makes actual waste equal to 1:36∕592:3 ¼ 0:229%. Fig. 12
shows the percent waste further reduced by combination of source
Total
B01
B02
B03
B04
B05
B06
B07
B08
B09
B10
B11
B12
B13
Bar

bars with 12,200-mm market length and available cutoff lengths to

952 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 8 7.363
reinforcement bars could be reduced substantially by better
7
Trim Loss (%)
7.111 10M Rebar planning and optimization techniques.
6
15M Rebar
Generally, SA models are successful in complex combinatorial
5
optimization problems through controlled randomization. For this
4 20M rebar
reason, a simulated annealing heuristic approach was proposed to
3
1.962 solve the cutting stock problem in the construction industry. The
2 1.398
1
1.398 waste-optimization model that was developed was validated with
0.256 0.932 0.256
0
0.256 a case study. The model was successful in reducing the cost and
amount of cutting losses. The method has the advantage of being
6100 12200 18300
able to account for influence of rebar market lengths on the design
Market Lengths of Rebar (mm) changes. This method gives more cost-effective results when
Fig. 11. Comparison of estimated trim loss for available market
higher-diameter rebars are used, as cost increases with an increase
lengths
in diameter for even smaller cutoff length. The model is capable of
handling any kind of reinforced concrete frame structure.
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 12. Waste ratio with combination lengths of source bar
0.113, 0.341, and 0.229% for 10-, 15-, and 20-M bars, respectively.
This shows that waste of material as high as 1.6% can be saved
when proper bars in market length are selected for combination
and longer lengths.
On the basis of the results of modeling and optimization pro-
cess, the following alternatives are recommended during the design
stage:
• Accurate refinement of the BIM model by all building stake-
holders in the design stage,
• Continuous structural members with longer lengths,
• Similar diameter of rebar for main reinforcement,
• Model geometry to avoid special structural members (e.g., doubly
reinforced beams),
• BIM model according to the construction stages at site, and
• Exact placement of reinforcement in the BIM model to reflect
actual fabrication.
Conclusions
The objective of this research was to extend BIM into construction
waste management to minimize rebar waste. None of the available
and published methods and research focused on minimizing trim
loss at the early in the design stage of a project. Their focus rather
was at the user end during the construction phase. The main
contribution of this paper is to introduce a novel methodology
to minimize rebar waste at the project design stage by analyzing
BIM models with optimization technique quickly. As one of the
main contributors to construction waste, steel rebar waste was
chosen to be the focus of this research. The cutting losses of
References
Bangash, M. Y. H. (2003). Structural detailing in concrete, 2nd Ed.,
Thomas Telford Publishing, London.
Begum, R. A., Siwar, C., Pereira, J. J., and Jaafar, A. H. (2006).
“A benefit–cost analysis on the economic feasibility of construction
waste minimisation: The case of Malaysia.” Resour. Conserv. Recycl.,
48(1), 86–98.
buildingSMARTAlliance. (2011). “A council of the national institute of
building sciences.” 〈www.buildingsmartalliance.org〉 (Jan. 2011).
Canada Border Service Agency. (2011). 〈www.cbsa-asfc.gc.ca/sima-lmsi/
i-e/ad1225/ad1225p-eng.html〉 (Mar. 12, 2011).
Canadian Construction Association (CCA). (1992). “National round table
on the environment and the economy.” A Rep. on Waste Management
for the Construction Industry, Ottawa.
Canadian Standards Association (CSA)/National Standard of Canada.
(2000a). “Billet-steel bars for concrete reinforcement.” CAN/CSA
G30.18-M92, Toronto.
Canadian Standards Association (CSA)/National Standard of Canada.
(2000b). “Concrete materials and methods of concrete construction.”
CSA A23.1, Toronto.
Chen, C.-L. S., Hart, S. M., and Tham, W. M. (2005). “A simulated
annealing heuristic one-dimensional cutting stock problem.” Eur. J.
Oper. Res., 93(3), 522–535.
Eastman, C. (1999). “Building product models: Computer environments.”
Supporting design and construction, CRC Press, Boca Raton, FL.
Fu, C., Aouad, G., Lee, A., Mashall, A., and Wu, S. (2006). “IFC model
viwer to support nD model application.” Autom. Constr., 15(2),
178–185.
Geodert, D. J., and Meadati, P. (2008). “Integrating construction process
documentation into building information modeling.” J. Constr. Eng.
Manage., 134(7), 509–516.
Gilmore, P. C., and Gomory, R. E. (1961). “A linear programming approach
to the cutting stock problem.” Oper. Res., 9(6), 849–859.
Goulimis, C. (1990). “Optimal solutions for the cutting stock problem.”
Eur. J. Oper. Res., 44(2), 197–208.
Gradisar, M., Kljaji, M., Resinovi, G., and Jesenko, J. (1999). “A sequential
heuristic procedure for one-dimensional cutting.” Eur. J. Oper. Res.,
114(3), 557–568.
Haessler, R. W. (1975). “Controlling cutting patterns changes in one-
dimensional trim problems.” Oper. Res., 23(3), 483–493.
Holland, J. H. (1970). Adaptation in natural and artificial systems, Uni-
versity of Michigan Press, Ann Arbor, MI, 1–200.
Isikdag, U., Underwood, J., and Aouad, G. (2008). “An investigation into
the applicability of building information models in geospatial environ-
ment in support of site selection and fire response management
process.” Adv. Eng. Inf., 22(4), 504–519.
Jahromi, M. H. M. A., Tavakkoli-Moghaddam, R., Givaki, E., and
Rezapour-Ziba, A. (2011). “A simulated annealing approach for a
standard one-dimensional cutting stock problem.” Int. J. Acad. Res.,
3(1), 353–357.
Kim, S. K. (2002). “A system development for automatic detail design and
estimation of rebar work.” The 1st Year Research Rep., Gyeonggi

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012 / 953

J. Constr. Eng. Manage., 2012, 138(8): 943-954

 Regional Small & Medium Business Administration, Gyeonggi, South
Korea, 84–90.
Kim, S.-K., Hong, W.-K., and Joo, J.-K. (2004). “Algorithms for reducing
the waste rate of reinforcement bars.” J. Asian Archit. Build. Eng., 3(1),
17–23.
Kirkpatrick, S., Gelatt, C. D., and Vecchi, M. P. (1983). “Optimization by
simulated annealing.” Science, 220(4598), 671–680.
Li, H., Chen, Z., Yong, L., and Kong, S. (2005). “Application of integrated
GPS and GIS technology for reducing construction waste and
improving construction efficiency.” Autom. Constr., 14(3), 323–331.
Metro Vancouver Construction Waste Management Disposal. (2008).
“Model specifications.” 〈http://www.metrovancouver.org〉 (May 2010).
Metropolis, N., Rosenbluth, A., Rosenbluth, M., Teller, A., and Teller, E.
(1953). “Equation of state calculations by fast computing machines.”
J. Chem. Phys., 21(6), 1087–1092.
Downloaded from ascelibrary.org by Universitas Diponegoro on 03/19/18. Copyright ASCE. For personal use only; all rights reserved.
Pierce, J. F., and Lasky, J. S. (1973). “Improved combinatorial program-
ming algorithms for a class of All-Zero-One integer programming prob-
lems.” Manage. Sci., 19(5), 528–543.
Poon, C. S., Yu, A. T. W., and Jaillon, L. (2004). “Reducing building waste
954 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2012
J. Constr. Eng. Manage., 2012, 138(8): 943-954
at construction sites in Hong Kong.” Constr. Manage. Econ., 22(5),
461–470.
Popov, V., Juocevicius, V., and Migilinskas, D. (2010). “The use of virtual
building design and construction model for developing an effective
project concept in 5D environment.” Autom. Constr., 19(3), 357–367.
Salem, O., Shahin, A., and Khalifa, Y. (2007). “Minimizing cutting wastes
of reinforcement steel bars using genetic algorithms and integer
programming models.” J. Constr. Eng. Manage., 133(12), 982–992.
Schlueter, A., and Thesseling, F. (2009). “Building information model
based energy/exergy performance assessment in early design stages.”
Autom. Constr., 18(2), 153–163.
Steel Construction Institute (SCI). (2003). “The logical product model
(LPM/5).” CIS/2, Berkshire, UK, 〈http://dcom.arch.gatech.edu/
cs6754/STEP/Files/SCI-P268onNov00.pdf〉 (Jul. 2010).
Syal, M. G., Mago, S., and Moody, D. (2007). “Impact of LEED-NC credits
on contractors.” J. Archit. Eng., 13(4), 174–179.
U.S. Green Building Council (USGBC). (2011). “USGBC research
program.” 〈www.usgbc.org〉 (Apr. 14, 2011).