ISSN: 2277-3754

ISO 9001:2008 Certified

International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014

Optimization of Plane and Space Trusses

Using Genetic Algorithms
Osman Shallan, Atef Eraky, Tharwat Sakr, Osman Hamdy
Department of Structural Engineering, faculty of engineering, Zagazig University, Zagazig, Egypt
applied which imitate natural evolution to produce new
Abstract— weight optimization of trusses is so important due to candidate solutions. At the end of the process, the newly
economic and sustainability considerations. Geometry, topology created generation replaces the previous generation and
and sizing optimization is extensively found in literature.
evolution is repeated until obtaining appropriate solution to
Applications found in literature uses the traditional deign
variables containing node coordinates, elements connectivity and the problem while ensuring certain design criteria are satisfied
member cross sections. This paper presents an approach based on or reaching the pre-determined maximum number of
the genetic algorithm for optimum design of plane and space generations. The optimization of truss structures can be
trusses subjected to specified set of constraints. The proposed classified into three categories depending on which
approach defined innovative design variables in terms of node component of the structure is used as a design variable: 1)
coordinates and displacements. Such limited design variables lead Sizing, 2) Shape and 3) Topology optimization as shown in
to the reduction of genotype length resulting in less execution Table 1. In sizing optimization cross-sectional area of the
time. Topology and cross sections are estimated after using
members are the design variables and the coordinates of the
strength criteria. The proposed approach was applied on
benchmark problems repeated in literature, the proposed nodes and the connectivity of various members are fixed.
approach resulted in more optimized results with less However, in truss design problems, cross sections are
mathematical effort. considered discrete variables such that member
cross-sectional areas are specific predefined values. In Shape
Index Terms—Optimization; plane truss; Space truss; optimization the design variables are the nodal coordinates,
Genetic algorithms and in topological optimization the number of nodes and the
connectivity between nodes are the design variables while
I. INTRODUCTION nodal coordinates are assumed known. However, the most
Unequivocally the material cost is one of the foremost efficient design will be obtained by considering all three
driving factors in the construction of a buildings; it can be categories simultaneously. Generally, multilevel optimization
minimized by reducing the weight or volume of the structural methods are used in which topological optimization first
system. In addition, such reduction in the used material serves performed keeping the shape and cross sectional sizes fixed.
as a tool for sustainable design as step for green buildings. All When an optimal topology is found, shape and/or sizing
of the methods used for decreasing the weight intend to reach optimization is performed on the topology found in the
an optimum design having a set of design variables under previous step. But this technique may not lead to the most
specific design constrains. It is essential to comprehend the optimal solution as all the three problems are not mutually
characteristics of the problem while going for an appropriate independent. As a result, traditional methods of optimization
optimization method for structural design. The focal attribute have not been suitable for such problem and the use of other
of structural design optimization is that the solution sought is techniques such as GAs is gaining popularity in the field of
the global optimal solution [1] and the design variables are structural optimization. Although it becomes difficult to
discrete and must be chosen from a pre- determined set which optimize in complex structures where variable interactions
is suitable for engineering design problems. Genetic increase. Classical optimization methods can produce
Algorithms is a part of evolutionary computational technique, sub-optimal results because of these interactions [28] In this
it is a global search method which has been preferred by paper, new approach is introduced for simultaneous shape,
various researchers over other classical optimization topology and size optimization of plane trusses and sizing
techniques for specific applications. [2] Furthermore, it can optimization for space trusses. Nodal coordinates and
be used in wide ranges of optimization problems. In contrast deflections are used as design variables instead of instead of
to traditional optimization methods which begin from single using topology and cross sectional area. Topology and cross
point solution, GA starts with population of solutions within a sections are determined after to cope with the used design
search space. Moreover, this technique works with a coding variables satisfy strength conditions. This method avoids
of a parameter set as opposed to traditional optimization conventional disadvantages of traditional methods of conflict
methods which work with the parameters themselves Each use of optimization categories and the assumption of constant
individual combination in population has a fitness value A/L and keeps it variable.
determined by a fitness function. Afterwards, according to
used crossover and mutation values the selection process is

66
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
Table 1: Examples of truss optimization and x, y, z are the net displacements of the member and
defined as:
Sizing Shape Topology (connectivity)
(2.1)
(2.2)
(2.3)
(3.1)
(3.2)
(3.3)
Where xi, yi, zi, ui, vi, wi are the nodal coordinates and
displacements of member joints, respectively. The strain of
each member is then compared to the allowable strain of used
material (all) as:

II. THE PROPOSED APPROACH

(4)
As described before, the core idea of our proposed
approach is to carry out topology and shape optimization
using nodes deflections and coordinates as a design variable. Excluding the members not complying with the condition
Design variables of the truss in this approach are as follows: at Eq. 4, the topology matrix is developed. This avoid us the
 Coordinates of each node (xi, yi for plane truss and xi, complications resulting from adding topology as design
yi, zi for space truss, where i=1 to N, N is the number variable which result in huge design solution chromosome
of truss nodes). The range of nodal coordinates is containing contradicting connectivity. At this stage, the shape
and connectivity of the truss are defined and the next step is to
defined by the truss proposed geometric limits.
estimate the member cross sections. Stiffness analysis of the
 Deflections of each node (ui, vi for plane truss and ui,
developed truss is then carried out assuming constant area to
vi, wi for space truss, where i=1 to N, N is the number length of members to get the member forces Fi for each
of truss nodes). The range of nodal deflections is member i. Now the cross section of each member can be
defined by the code limits of allowable deflection. derived using the strength and deflection criteria as:
For each chromosome in the generation at which the above
mentioned design variables are included, the following (5)
procedure is applied. At first, from the nodal coordinates of
the solution, member original lengths (Li) can be calculated
assuming that all nodes are connected and that is represented Where Fi is the force obtained in member i and all, E are
in topology matrix. Topology matrix describes the the allowable stresses and modulus of elasticity of truss
distribution of members in truss between nodes defined by: member material respectively. While the truss member forces
obtained from the analysis are not corresponding to global
deflections estimated in the design variables, the cross section
1 D11 D12 D13 D14 D15 D16
obtained by Eq. 5 can be used as just suggested cross section.
2 D21 D22 D23 D24 D25 D26 Such suggested sections are optimal or near optimal final
cross sections especially when the assumed deflections in the
GA is near that resulting from analysis so that helps GA to
m Dm1 Dm2 Dm3 Dm4 Dm5 Dm6 reach optimal solution more easily and quickly. The estimated
Topology matrix area is then used to select the member section from discrete
Where each row represents member and first column available cross sections. Lastly we will calculate the forces,
represent member number from 1 to m where m is the number stresses and deflections corresponding to these selected
of truss members and columns No. 2, 3, 4 represent degree of suggested cross sections to check if stresses or deflection
freedom of start node of member and columns No. 5, 6, 7 exceed the allowable limits. The master flow chart of the
represent degree of freedom of end node of same member proposed procedure is shown in Fig. 2.
Using the other group of design variables which represent
nodal deflections, the deformed length of each truss member III. PROPOSED VERSUS TRADITIONAL
can be calculated. The change in length (L) can be derived APPROACHES
using the relation The traditional approach of combining shape, topology and
(1) sizing design variables in the Genetic algorithm genotype has
several drawbacks as reported in literature [27]. Our proposal
Where x, y, z are the direction angles of the truss member constitutes a trial to possess powerful merits to overcome such

67
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
drawbacks by altering and reducing the design variables and limits of deflection which are lower than the limits of using
obtaining the remaining design variables using engineering all available of topology and cross sections. For some cases
design roles. For the traditional approach encountered in if we can easily predict the direction of deflection for each
literature, individual chromosomes combination in population point so the range of deflection variable can be reduced by
may produces optimum topology and shape but due to random 50%.
selection of cross sections this combination produces
Input data
section(s) smaller than minimum needed cross section(s) and
that cause exceeding constrain(s) limit(s) such stress or
deflection so due to this selected section(s) which lead to unfit Create initial population
structure due to the penalty resulting from design violations
usually expresses as [26].
For Each chromosome
(6)
Where F(x) is fitness value, f(x) weight of truss and is the
count of the number of constraints violated by a given Assemble the basic structure and translate the binary
solution including check of stability and constructability. chromosome in to FEM structure
Also sometimes individual chromosomes combination
produces a good topology and shape but due to random
selection this combination produce section(s) larger than Update topology matrix
minimum needed cross section(s), so the outcome truss is
overly weighted despite if we decrease this cross section(s)
Analyze the truss and get cross sections
slightly until reaching the minimum safe cross section(s) the
fitness will be improved. In addition, the chromosome
length needed for representing sizing variable depend on the
number of discrete available cross sections. The Check on constructability and stability
chromosome length reserved for member cross section is
estimated as: No
Passed

(7) Yes

Calculate the suggested sections

Where are the upper and lower bounds of the
variable respectively and is the desired precision
So if the available cross sections are few, the solution will Calculate Element Stresses and Node Deflection
not be accurate enough and if the available sections are
more the chromosome length will be long and it means more Penalty any violations
complexity and more time consuming. So, it is clear that
using sizing as variable doesn’t produce the optimum fitness
for most of individual chromosomes combination especially Evaluate the Fitness
in case of using infinite search and that is unpractical
method [26]. In addition, Being no relation between
Reached
topology as variable and other variables i.e. selecting Termination
topology is not depend on sizing or shape of truss that Criteria
makes it very complicated to make topology sizing and
shape optimization in same time and the number of possible Not Reached
solutions reaches extreme levels, which means very big
populations and long calculation time and also stuck Cross over & Mutation
problems may occurs [27]. On the other hands, the proposed
approach proved to give better solution than using sizing as Selection & New Generation
variable to avoid the abovementioned drawbacks and make
topology optimization by simple condition without need
any chromosomes and also avoiding the disadvantages of
Translate the binary chromosome and draw
traditional method. It also lead to the reduction of the
the truss and results
chromosome length because as we mentioned before the
chromosome length depend on the limits of variable and the
Fig.2: Master Flow Chart of the Optimization Algorithm

68
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
The node deflection variable is also associated with nodes
not members and as we know the number of truss node usually
less than the number of its members so that also reduces the
chromosome length needed to represent that variable. This
procedure is willing to get the maximum fitness for each
proposed shape and Topology because we get here the
minimum safe sections depending on engineering design rules
which can reduce the probability of getting unfit trusses
exceed constrains limits.

IV. RESULTS Fig.3: Structure of the10-bar truss example.

To verify our proposed approach and investigate its
stability and efficiency regarding its results and
computational effort, two benchmark problems are
considered. The 10 member plane truss problem and the 25
member space truss problems are selected to represent plane
and space trusses. The results of our proposed approach are
compared with results of optimization of these examples
found in previous work. Summary definition of problems and
comparison of results are included in this section.

A. 10-Bar plane Truss

The first problem is shape, sizing and topology
optimization for 10 bar plane truss. This ten-bar truss is often
used as a benchmark problem in structural optimization. This
structure is frequently found in literature related to plane truss
optimization. The truss has two vertical supports with a
distance 'a' of 9.144 meters (360 inches) and two loads 'F' of
445.374 kN (100 kips) at 9.144 and 18.288 meters from the
lower support as shown in Figure 3. Weight is minimized by Fig. 4: Convergence history of 10-bar plane truss structure.
GA with parameters as follows: population size 600,
maximum generation 200, 0.9 crossover and 0.05 mutation
probability as justified. Aluminum is used, with Modulus of
elasticity E = 68.95 GPa (104 ksi) , density ρ = 2 ,768 kg/m3
(0.1 lb/in3) and element stresses are limited to 172.37 MPa
(25 ksi) in both tension and compression while buckling is
ignored. The displacements are limited to 50.8 mm (2 in) both
horizontally and vertically as per code requirements. All
available cross sections are used here, as mentioned before
the large number of available cross section doesn't affect time
consumed or GA efficiency and not need much populations /
Fig.5: Optimized structure of the 10-bar truss.
generations number as member sections are not used as design
variables. Shown in Figure 4, is the convergence history of the Table 2 lists the results encountered in literature for the 10
10 bars truss example. At the figure, the fitness value is member benchmark problem compared with the results of the
plotted against the generation number to clarify the how the proposed approach. As illustrated from the results, our
GA converges to the optimum solution. As investigated from propose approach resulted in the most optimized value of
the plot, the fitness value improvement is very limited during fitness (weight) which is less than almost all results found in
the first 30 generation, and then greatly improved till the 100th literature. This may be attributed to the consideration of
generation after which stability is observed while Deb and limited optimization categories in literature in terms of shape
Gulati [12] reached stability around the 140th generation using and sizing or topology and sizing while our approach
110 populations and optimizing size and topology only. considers the three categories simultaneously. Only Deb and
Gulati [12].resulted in better solution, but as indicated, their
optimum solution contains two members over each other one
member between nodes P6 & P4 and other member between

69
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
nodes P6 & P2 which is considered unreal and undesirable
Coordinate. Of Variable Node P3
overlap [27].
Figure 5 shows the shape of optimized truss while table 3. X Y
And 4 show the results of nodal coordinates and member Cm Inch Cm Inch
sections. The maximum deflection investigated in the
optimum solution is 50.763 for node P2 in Y direction which 11.73 461.81 6.40 251.96
reaches 99.92% of the maximum permissible value which
means that the selected cross sections are almost optimum.
Table 6: Loading conditions for 25-bar space truss.
Table 3: Coordinate. Of Variable Point P3
B. 25-bar Space Truss Node Fx (KN) (lb) Fy (KN) (lb) Fz (KN) (lbs.)
1 4.45374 1000 -44.5374 -10000 -44.5374 -10000

2 0 0 -44.5374 -10000 -44.5374 -10000

3 2.22687 500 0 0 0 0

6 2.672244 600 0 0 0 0

joints and the member groups for section selection and

applied loads are shown in Tables 5, 6 and 7.
Aluminum is used with modulus of elasticity E =68.95 GPa
(104 ksi) and density, ρ = 2,768 kg/m3 ( 0.1 lb/in3) and
element stresses are limited to 275.8 MPa (40 ksi) in both
tension and compression while buckling is ignored.
The displacements are limited to 8.9 mm (0.35 in) in all
Fig.6: 25-bar space truss structure. directions as per code requirements.
Table 5: Coordinates of the joints of the 25-bar space truss.
Node X (m) (Inch) Y (m) (Inch) Z (m) (Inch)
1 -0.9525 -37.5 0 0 5.08 200
2 0.9525 37.5 0 0 5.08 200
3 -0.9525 -37.5 0.9525 37.5 2.54 100
4 0.9525 37.5 0.9525 37.5 2.54 100
5 0.9525 37.5 -0.9525 -37.5 2.54 100
6 -0.9525 -37.5 -0.9525 -37.5 2.54 100
7 -2.54 -100 2.54 100 0 0
8 2.54 100 2.54 100 0 0
9 2.54 100 -2.54 -100 0 0
Fig. 7: Convergence history of 25-bar Space truss structure
10 -2.54 -100 -2.54 -100 0 0
The set of areas available for this truss is
Table 7: Group membership for 25-bar space truss. S={0.l, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1,
Group Members 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 2.1, 2.2, 2.3, 2.4, 2.5,
Number
1 1-2 2.6, 2.8, 3.0, 3.2, 3.4, 3.6} (in2).
2 1-4, 2-3, 1-5, 2-6 Weight is minimized by GA with parameters as follows:
3 2-5, 2-4, 1-3, 1-6 population size 400, maximum generation 100 , 0.9 crossover
4 3-6, 4-5 and 0.05 mutation probability as justified.
5 3-4, 5-6 Shown in Figure 7, is the convergence history of the 25 bars
6 3-10, 6-7, 4-9, 5-8 truss example. At the figure, the fitness value is plotted
7 3-8, 4-7, 6-9, 5-10 against the generation number to clarify the how the GA
8 3-7, 4-8, 5-9, 6-10 converges to the optimum solution. As investigated from the
plot, the fitness value is greatly improved from first
The other benchmark problem is sizing optimization for generation till the 60th generation after which stability is
the 25 bar space truss shown in Figure 6. The coordinates of observed that is because of making optimization for sizing
only.

70
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
Table 8: Optimization results for 25-bar space truss.
Talasl
Rajeev Erbat Kaveh
Rajeev and To˘gan ioglu
Design and Erbatur ur Lee et and Tayfu
Krishnamo Coello and [20] Li et al. Camp
variable Krishnam Zhu et al. et al. al. Shoja n This
or-thy et al. Cao [18] Dalo˘gl BGA [21] [23]
(area oor-thy [15] [16] [16] [22] ee Dede study
[13] [17] u wEIS HPSO BB-BC
in2) [13] GAOS1 GAOS HSH [24] [25]
Ps=40 [19] Ps =
Ps=20 2 ACO
300
A1 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
A2 0.1 1.8 1.9 0.1 1.2 0.7 0.5 0.3 0.1 0.3 0.3 0.3 0.3 0.3 0.3
A3 2.3 2.3 2.6 3.4 3.2 3.2 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.6
A4 0.2 0.2 0.1 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
A5 0.1 0.1 0.1 0.6 0.1 1.4 1.9 2.0 1.9 2.1 2.1 2.1 2.1 2.1 1.7
A6 0.8 0.8 0.8 1.1 0.9 1.1 0.9 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.8
A7 1.8 1.8 2.1 0.9 0.4 0.5 0.5 0.5 0.7 0.5 0.5 0.5 0.5 0.5 0.4
A8 3.0 3.0 2.6 3.0 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.4 3.6
Weight
546.76 546.01 562.93 515.27 493.8 493.94 485.05 483.35 485.9 484.85 484.85 484.85 484.85 484.85 476.33
(lbs.)
Member space truss problem compared with the results of the
proposed approach. As illustrated from the results, our
propose approach resulted in the most optimized value of
fitness (weight) which is less than all results found in
literature. The maximum deflection investigated in the
optimum solution is 8.89 mm for node 1 in Y direction which
reaches 99.89% of the maximum permissible value which
means that the selected cross sections are almost optimum.
from huge numbers of a pre‐determined set without any
V. CONCLUSIONS increasing the chromosome length.
When dealing with simultaneous size, shape and topology
optimization, the number of possible solutions reaches great REFERENCES
levels, which need long chromosome, big populations and [1] Nanakorn P, Meesomklin K. An adaptive penalty functions in
long calculation time. So to overcome these problems we genetic algorithms for structural design optimization.
should work on reducing the chromosome length without Computer and Structures 2001; 79:2527–39.
affecting the GA efficiency. An approach is proposed based [2] Goldberg DE. Genetic algorithm in search optimization and
on using nodal deflections as design variable instead of the machine learning. New York: Addison-Wesley; 1989.
member sections in addition to the nodal coordinates. This [3] Deb K and Gulati S, Design of truss-structures for minimum
will reduce the length of genotype as nodes are always less weight using genetic algorithms, Finite elements in analysis
than members in truss and as the range of nodal displacement and design, 37 (2001) 447-465
is less than the range of available steel sections for truss [4] Hajela, P. Lee, E. (1995). Genetic algorithms in truss
members. In addition, according to loads and configurations topological optimization. International Journal of Solids and
the direction of deflection can be expected which reduces the Structures. Volume 32, issue 22. Pages3341-3357.
deflection variables to 50% which can improve the [5] Li, Lijuan. Huang, Zhibin. Liu, Feng. (2006). An Improved
calculations. By using simple condition it also allows Particle Swarm Optimizer for Truss Structure Optimization.
removing and keeping on members i.e. making topology 2006 International Conference on Computational Intelligence
optimization without using chromosome for it and also and Security. Volume 1.Pages 924-928.
reduce Complexity and stuck problems of making Topology [6] Kripakaran, Prakash. Gupta, Abhinav. Baugh Jr. John W.
optimization by traditional method because this condition (2007). A novel optimization approach for minimum cost
make topology optimization depend on other variables design of trusses. Computers & Structures. Volume 85, issue
(coordinates of nodes and deflections of nodes). The 23-24. Pages 1782-1794.
proposed procedure was applied to two of the classical truss [7] Galante, Miguel. (1996). genetic algorithms as an approach to
problems and the results were compared to the results of optimize real-world trusses. International journal for
previous work found in literature. The presented results not Numerical Methods in engineering. Volume 39, issue 3.
only produce better optimum weight than previous results Pages 361-382.
but also reduced the calculation time and effort by using a [8] Su Ruiyi, Gui Liangjin, Fan Zijie (2009). truss topology
few numbers of chromosomes and ability to choose sections optimization using genetic algorithm with individual

71
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
identification technique Proceedings of the World Congress [27] Max Hultman. Weight optimization of steel trusses by a
on Engineering 2009 Vol II genetic algorithm – Size, shape and topology optimization
according to Eurocode. (2010)
[9] Rajan SD. Sizing, shape and topology design optimization of
trusses using genetic algorithm. J Struct Eng 1995; [28] Brian J. Auer. Size and Shape Optimization of Frame and
121(10):1480–7. Truss Structures through Evolutionary Methods (2005).
[10] Tang W, Tong L, Gu Y. Improved genetic algorithm for
design optimization of truss structures with sizing, shape and
topology variables. Internat J Numer Methods Engrg 2005;
62:1737–62.
[11] H. Rahami, A. Kaveh,_, Y. Gholipour. Sizing, geometry and
topology optimization of trusses via force method and genetic
algorithm. Engineering Structures 30 (2008) 2360–2369
[12] Kalyanmoy Deb.Surendra Gulati.(2001) Design of
truss-structures for minimum weight using genetic algorithms
Finite elements in analysis and design Volume 37 (2001)
447-465
[13] S. Rajeev, C.S. Krishnamoorthy, Discrete optimization of
structures using genetic algorithms, J. Struct. Eng. 11–5
(1992) 1233–1250.
[14] James M. Gere and William Weaver. Analysis of
Framed Structures. D. Van Nostrand Company, Inc.,1965.
[15] D.M. Zhu, An improved Templeman’s algorithm for optimum
design of trusses with discrete member size, Eng. Opt. 9
(1986) 303–312.
[16] F. Erbatur, O. Hasancebi, I. Tutuncu, H. Kılıc，, Optimal
design of planar and space structures with genetic algorithm,
Comput. Struct. 75 (2000) 209–224.
[17] C.A.C. Coello, M. Rudnick, A.D. Christiansen, Using genetic
algorithm for optimal design of trusses, IEEE (1994) 88–94.
[18] G. Cao, Optimized design of framed structures using a genetic
algorithm, Ph.D.Thesis, The University of Memphis, 1996.
[19] V. To˘gan, A.T. Dalo˘glu, An improved genetic algorithm
with initial population strategy and self-adaptive member
groupings, Comput. Struct. 86 (2008)1204–1218.
[20] T. Talaslioglu, A new genetic algorithm methodology for
design optimization of truss structures: bi population-based
genetic algorithm with enhanced interval search, Model.
Simul.Eng. (2009) 28, Article ID 615162.
[21] L.J. Li, Z.B. Huang, F. Liu, A heuristic particle swarm
optimization method fortruss structures with discrete
variables, Comput. Struct. 87 (2009) 435–443.
[22] K.S. Lee, Z.W. Geem, S.H. Lee, K.W. Bae, The harmony
search heuristic algorithm for discrete structural optimization,
Eng. Opt. 37–7 (2005) 663–684.
[23] C.V. Camp, Design of space trusses using big bang-big
crunch optimization, J.Struct. Eng. (2007) 999–1008.
[24] A. Kaveh, S. Shojaee, Optimal design of skeletal structures
using ant colony optimization, Int. J. Numer. Methods Eng.
70 (2007) 563–581.
[25] TayfunDedea, Serkan Bekiro˘glu, Yusuf Ayvaz , Weight
minimization of trusses with genetic algorithm , Applied Soft
Computing 11 (2011) 2565–25
[26] Sivanandam S. N. Deepa S. N. (2008). Introduction to
Genetic Algorithms. Berlin:Springer.

72
 ISSN: 2277-3754
ISO 9001:2008 Certified
International Journal of Engineering and Innovative Technology (IJEIT)
Volume 3, Issue 7, January 2014
APPENDIX
Table 2: Results of previous works with same conditions
Optimization category
Weight
Search Method Size Shape Topology (lbs.)
Galante (1996) [7]. Genetic algorithm √ √ 5119.3
Kripakaran, Gupta and Baugh Jr. (2007) [6]. Hybrid search method. √ 5073.03
Li, Huang and Liu (2006) [5]. Particle swarm √ 5060.9
Rajan (1995) [9]. Genetic algorithm √ √ 4962.1
Su Ruiyi, Gui Liangjin, Fan Zijie (2009) [8]. Genetic algorithm √ √ 4962.07
Hajela and Lee (1995) [4]. Genetic algorithm √ √ 4942.7
Wenyan (2005) [10]. Genetic algorithm √ √ 4921.25
Deb and Gulati (2001) [3]. Genetic algorithm √ √ 4899.15
H. Rahami, A. Kaveh (2008) [11]. Force method √ √ 4855.2
Deb and Gulati (2001) [12]. Genetic algorithm √ √ 4731.65
This study Genetic algorithm √ √ √ 4762.1

Table 4: Optimization results for 10-bar plan truss

Area Stress
Element No. Dimensions mm % Stress of allowable
in2 m2 KSI Mpa

6.04 0.00 23.42 161.

M1 200x200x5 93.7%
5 4 4 5
24.1 0.01
M2 400x400x10 -8.398 -57.9 33.6%
8 6
19.1 0.01
M3 260x260x12.5 -5.782 -39.9 23.1%
8 2
0.00
M4 180x180x12 12.5 -8.236 -56.8 32.9%
8
21.0 0.01 6.812 46.9
M5 350x350x10 27.2%
8 4 2 7
28.8 0.01 7.124 49.1
M6 400x400x12 28.5%
7 9 5 2

73