Está en la página 1de 5

Moth-Flame optimization algorithm for solving Real

Challenging Constrained Engineering optimization


problems
Indrajit N.Trivedi

Authors Name/s per 2nd Affiliation (Author)

line 1 (of Affiliation): dept. name of organization


line 2-name of organization, acronyms acceptable
line 3-City, Country
line 4-e-mail address if desired

line 1 (of Affiliation): dept. name of organization


line 2-name of organization, acronyms acceptable
line 3-City, Country
line 4-e-mail address if desired

Abstract In this paper, a novel nature-inspired optimization


algorithm based on the navigation strategy of Moths in universe
called the Moth-Flame optimization (MFO) Algorithm, is applied for
constrained optimization and engineering design problems. A
comparative analysis of MFO algorithm expresses the optimum
functional value in term of accuracy and standard deviation over rest
of well-known constraint optimization algorithms. Five constrained
benchmark function of engineering problems have been calculated
and gained solutions were compared with other recognized
algorithms. The gained solution expresses that MFO algorithm
provides better results in various design problems compared to other
optimization algorithms.

Keywords Moth-Flame optimization; constrained benchmark


function; Nature-inspired; Constrained handling; Algorithm

I. INTRODUCTION
A novel natureinspired Moth-Flame optimization algorithm
[1] based on the transverse orientation of Moths in space.
Transverse orientation for navigation uses a constant angle by
Moths with respect to Moon to fly in straight direction in
night. In MFO algorithm that Moths fly around flames in a
Logarithmic spiral way and finally converges towards the
flame. Spiral way expresses the exploration area and it
guarantees to exploit the optimum solution.
In past years many algorithms based on gradient search for
solving linear and non-linear equation but in gradient search
method value of objective function and constraint unstable and
multiple peaks if problem having more than one local
optimum.
Recently several meta-heuristic techniques used for
obtaining optimum solution of engineering design problems
like as Car side impact design, Marine ship propeller design,
Speed Reducer design, Rolling Element Bearing design,
Belleville Spring design.
Constrained optimization engineering design problem are
compared to the most popular algorithms are: Genetic
Algorithms (GA) [14], artificial bee colony (ABC) [8],
Teaching-learning-based optimization (TLBO) [9, 10, and 11],

modified differential evolution (MDE) [12, 13] and mine blast


algorithm (MBA) [6].
The genetic algorithm (GA), most popular stochastic
optimization algorithm has ability to avoid local optima and
overall average fitness of population are improved over the
course of generations. These two qualities is key reason of
success of GA in solving engineering constrained optimization
problem.
II. MOTH-FLAME OPTIMIZER
Moth-Flame optimizer is first introduced by Seyedali
Mirjalili In 2015 [1]. MFO a population based algorithm; we
represent the set of moths in a matrix:
m1,1 ,
m
2,1,
M

mn ,1

m1,2
m2,2

mn ,2

, , m1, d
, , m2, d

, , mn , d

Where n is the number of moths and d is the number of


variables (dimension).
For all the moths, we also assume that there is an array for
storing the corresponding fitness values as follows:
OM 1
OM 2

OM .

OMn

Where n is the number of moths.


Note that the fitness value is the return value of the fitness
(objective) function for each moth. The position vector (first
row in the matrix M for instance) of each moth is passed to
the fitness function and the output of the fitness function
is assigned to the corresponding moth as its fitness
function (OM1 in the matrix OM for instance).

Other key components in the proposed algorithm are flames.


We consider a matrix similar to the moth matrix:
F 1, 1
F 2, 1

F .

.
Fn , 1

F 1, 2
F 2, 2
.
.
Fn , 2

.
.
.
.
.

F 1, d

F 2, d

Fn , d

.
.
.
.
.

Where n is the number of moths and d is the number of


variables (dimension).
We know that the dimension of M and F arrays are equal.
For the flames, we also assume that there is an array for
storing the corresponding fitness values:
OF 1
OF 2

OF .

OFn

Where n is the number of moths.


It should be noted here that moths and flames are both
solutions. The difference between them is the way we treat
and update them each iteration. The moths are actual search
agents that move around the search space, whereas flames are
the best position of moths that obtains so far. Therefore, each
moth searches around a flame and updates it in case of finding
a better solution. With this mechanism, a moth never loses its
best solution.
The MFO algorithm is three-tuple that approximates the
global optimal of the optimization problems and defined as
follows:

MFO

I, P, T

(5)

I is a function that generates a random population of moths


and corresponding fitness values. The methodical model of
this function is as follows:
I : {M , OM }
(6)
The P function, which is the main function, moves the moths
around the search space. This function received the matrix of
M and returns its updated one eventually.
P : M M
The T function returns true if the termination criterion is
satisfied and false if the termination
Criterion is not satisfied:
T : M {true, false}
In order to mathematical model this behavior; we update the
position of each moth with respect to a flame using the
following equation:
Mi S ( Mi, Fj )
th

S M i , Fj Di * ebt cos 2 t Fj
Where: Di indicates the distance of the moth for the j

flame, b is a constant for defining the shape of the logarithmic


spiral, and t is a random number in [-1, 1].

Di Fj Mi
th

Fj indicates the j th flame,


th
and where Di indicates the distance of the i moth for the
j th flame.
Where: Mi indicate the i moth,

The number of flames is adaptively decreased over the


course of iterations. We use the following formula:
N 1 (12)

flame no round N l *

Where l is the current number of iteration, N is the


maximum number of flames, and T indicates the maximum
number of iterations.
We utilize Quicksort algorithm, the sort is of O ( n log n )
2

and O ( n ) in the best and worst case, respectively.


Considering the P function, therefore, the overall
computational complexity is defined as follows:

O MFO O t O Quick sort O position update

O( MFO) O(t (n n * d )) O(tn tnd )


2

(13)
Where n is the number of moths, t is the maximum number
of iterations, and d is the number of variables.
III. ENGINEERING BENCHMARK CONSTRAINTED PROBLEM
A. Car side impact design problem
The aim of Car side impact design is to minimize weight
using eleven influence parametric quantities including
thicknesses of B-Pillar inner, B-Pillar inner reinforcement,
floor side inner, cross members, door beam, door beltline
reinforcement and roof rail(x1-x7), materials of B-Pillar inner
and floor side inner(x8-x9) and barrier height and hitting
position(x10-x11).The car side impact design problem is
proposed by Gu Et al. [2,3].
Moth-flame optimizer convergence curve in Fig.1.for car
side impact design and statistical results are compared with
various well known algorithms in TABLE I.

th

Where Mi indicate the i moth, F j indicates the j flame,


and S is the spiral function.
Considering these points, we define a logarithmic spiral for
the MFO algorithm as follows:

th

Fig. 1. Convergence curve of car side impact design

TABLE I.

COMPARISON OF STATISTICAL RESULTS FOR CAR


SIDE IMPACT DESIGN USING SIX ALGORITHMS

Method
Best
value
Mean
value
SD

PSO
22.844

DE
22.842

GA
22.856

FA
22.842

CS
22.842

MFO
22.186

22.894

23.228

23.515

22.893

22.858

22.875

0.1501

0.3445

0.6655

0.1666

0.0761

3.7E15

Moth-flame optimizer convergence curve in Fig.2.for speed


reducer design problem and statistical results are compared
with artificial bee colony (ABC) [8], Teaching-learning-based
optimization (TLBO) [9, 10, and 11], modified differential
evolution (MDE) [12, 13] and mine blast algorithm (MBA) [6]
in TABLE III.

B. Marine ship propeller design problem


A Propeller gives the required thrust for moving marine and
aero vehicles. The high resistance of water due to high
density, the efficiency of the propeller should be high.
A Propeller is designed as follows for achieving highest
efficiency:

Va
K ( x)
T
2 nD KQ ( x)

KT ( x )
Where

T
n2 D 2

(1)
Fig. 2. Convergence curve of speed reducer problem

(2)

- axial velocity, D diameter length =2 m, n

rotation speed of propeller, K T -thrust coefficient,

KQ -

torque coefficient, -fluid density, T -thrust.


The design parameters of a propeller are defined based on
the representation method of the shape of the blades.
Constraint and other details are not given in this paper, so
for detailed information readers are referred to Carltons book
[4, 5].
Statistical results are compared with ant lion optimizer
(ALO) and moth-flame optimizer (MFO) algorithms in
TABLE II.
TABLE II.
COMPARISON OF STATISTICAL RESULTS FOR
MARINE SHIP PROPELLER DESIGN USING TWO ALGORITHMS
Five blade value
Name

ALO

MFO

Efficiency

0.68647

0.68878

Adverse Efficiency

0.82919

0.82909

TABLE III.
COMPARISON OF STATISTICAL RESULTS FOR
SPEED REDUCER DESIGN USING TWO ALGORITHMS
Method

Mean value

Best value

SD

ABC

2997.0580

2997.05800

TLBO

2996.3481

2996.34817

MDE

2996.3672

2996.35668

8.2E-3

MBA

2996.7690

2994.48245

1.56

MFO

2994.47106

2994.471066

D. Belleville spring design problem


The main purpose is to design with minimum weight and
satisfying a number of constraints. The control variables are
external diameter ( De ) ,thickness (t ) , internal diameter

( Di ) and height ( h) of the spring. In designing various


parameters were considered like slope, deflection, stress,
height to deflection, length and both diameters..
The statistical results for Belleville spring design problem
were compared in terms of mean, best and SD values.

C. Speed reducer design problem


The main aim of this constraint benchmark function or
problem is to minimize the weight of speed reducer. In this
constrained optimization problem various parameter are
involved such as bending stress of the gear teeth, transverse
deflections of shafts, and stresses in shafts and surface stress
[6]. This problem involved eleven constraints, produces high
complexity of problem reported in [7] solutions are also given
but that are infeasible.

Moth-flame optimizer convergence curve in Fig.3.for speed


reducer design problem and statistical results are compared
with ABC, TLBO, and MBA in TABLE IV.
Belleville spring is used in various applications where high
spring loads are required. They are particularly useful where
vibration, differential thermal expansion, relaxation, and bolt
creep are problematic.
The results for Belleville spring design problem are slightly
better than other algorithms.

TABLE V.
COMPARISON OF STATISTICAL RESULTS FOR
ROLLING ELEMENT BEARING DESIGN USING FOUR ALGORITHMS
Method

Fig. 3. Convergence curve of belleville spring design problem

Mean value

Best value

SD

ABC

1.995475

1.979675

0.07

TLBO

1.979687

1.979675

0.45

MBA

1.984698

1.979674

7.78E-03

MFO

2.02954437

1.97967463

0.130595

E. Rolling element bearing design problem


The main purpose of rolling element bearing design is to
maximize the dynamic load carrying capacity. The control
parameters are pitch diameter ( Dm ) , ball diameter ( Db ) ,
inner and outer raceway curvature coefficient ( f i , f 0 ) ,
number of ball ( Z ) ,

K D max , KD min , , e and .

All design variables are continuous except number of balls


that is discrete variable means number of balls is rounded in a
nearest integer value. Rolling element consist of
manufacturing consideration, kinematic and according to
application requirement. Rolling element exists in between
two surfaces that are moveable and provides a facility of less
friction.

SD

81843.3

N.A

N.A

ABC

81496.000

81859.7416

0.69

TLBO

81438.987

81859.74

0.66

MBA

85321.403

85535.9611

211.52

MFO

85379.937

85220.6813

1.4551E-11

Moth-flame optimizer convergence curve in Fig.4.for rolling


element bearing design problem and statistical results are
compared with Genetic algorithm (GA) [14], ABC, TLBO,
and MBA in TABLE V.

Acknowledgment
The authors would like to thank Professor Seyedali Mirjalili
and staff of LE College Morbi for their valuable support.

Conclusion
Constrained optimization an ability to find out optimum
solution with constraints handling which includes both
equality and inequality constraints. While obtaining optimum
solution constraint limits should not be violated. Real
engineering problems are mainly constrained problem. A
constraint handling method should be integrated with
optimizer for optimizing constrained problem. In this paper
MFO provides best solution in compared to other well-known
algorithms with different kinds of problem like highly
complex constrained, expensive, continuous and discrete
control parameters. The result of various constrained problems
proves that it is also an effective method in solving
challenging problems with unknown search space.

References
1.

2.

3.

4.
5.
6.
Fig. 4. Convergence curve of rolling element bearing

Best

GA

TABLE IV.
COMPARISON OF STATISTICAL RESULTS FOR
BELLEVILLE SPRING DESIGN USING FOUR ALGORITHMS
Method

Mean

Seyedali
Mirjalili,"Moth-flame
optimization
algorithm: A novel nature-inspired heuristic
paradigm," Knowledge-Based System, vol. 89, pages
228-249, 2015.
Gu L, Yang RJ, Cho CH, Makowski M, Faruque M,
Li Y (2001) Optimization and robustness for
crashworthiness. Int J Vehicle Design 26(4):348360.
Gandomi AH, Yang XS, Alavi AH, Mixed variable
structural optimization using firefly algorithm.
Computers & Structures.
Carlton J (2012) Marine propellers and propulsion.
Butterworth-Heinemann, Oxford.
Mirjalili S (2015) The ant lion optimizer. Adv Eng
Softw 83:8098.
Sadollah, A. Bahreininejad, H. Eskandar, and M.
Hamdi, "Mine blast algorithm: A new population
based algorithm for solving constrained engineering

7.

8.

9.

10.

11.

12.

13.

14.

optimization problems," Applied Soft Computing,


vol. 13, pp. 2592-2612, 2013.
J.K. Kuang, S.S. Rao, L. Chen, Taguchi-aided search
method for design optimization of engineering
systems, Engineering Optimization 30 (1998) 123.
B. Akay, D. Karaboga, Artificial bee colony
algorithm for large-scale problems and engineering
design optimization, Journal of Intelligent
Manufacturing (2010).
R.V. Rao, V.J. Savsani, D.P. Vakharia, Teachinglearning-based optimization: a novel method for
constrained
mechanical
design
optimization
problems, Computer-Aided Design 43 (2011) 303
315.
R.V. Rao, V. Patel, An elitist teaching-learning-based
optimization algorithm for solving complex
constrained optimization problems, International
Journal of Industrial Engineering Computations 3
(2012) 535560.
R.V. Rao, V. Patel, Multi-objective optimization of
heat exchangers using a modified teaching-learning
based-optimization algorithm, Applied Mathematical
Modelling (2012).
E. Mezura-Montes, J. Velazquez-Reyes, C.A.C.
Coello, Modified differential evolution for
constrained optimization, in: Evol. Comput., CEC
2006, IEEE Congress, 2006, pp. 2532.
E. Montes-Montes, C.A.C. Coello, J. VelazquezReyes, Increasing successful offspring and diversity
in differential evolution for engineering design, in:
Proceedings of the Seventh International Conference
on Adaptive Computing in Design and Manufacture,
2006, pp. 131139.
S. Gupta, R. Tiwari, B.N. Shivashankar, Multiobjective design optimization of rolling bearings
using genetic algorithm, Mechanism and Machine
Theory 42 (2007) 14181443.

También podría gustarte