Bat Algorithm to Series- Parallel Power System

Design
Guerraiche K1., Rahli M1., and Dekhici L2.
1Laboratory, optimization of electrical networks, University
of Sciences and the Technology (USTO-MB)
BP 1505 El MNaouer, Oran, Algeria,
2 LAMOSI laboratory, Mathematics and Computer Science
Faculty
University of Sciences and the Technology (USTO-MB)
31000, Oran, Algeria
khguerraiche@yahoo.fr, rahlim@yahoo.fr
latifa.dekhici@univ-usto.dz

Abstract In this paper, we intend to solve a well-known

power Redundancy Optimization Problem (ROP) also
called Power system design optimization using bio-inspired
metaheuristics Bat Algorithm (BA). We combine them to
Universal Moment Generating Function (UMGF).The
problem consists on selecting the appropriate components
from a system given the minimum cost so to fulfill a system
reliability constraint. Experimental results show that the
algorithms can seek to the optimal design and give good
results.
Keywords Bat algorithm, Series-Parallel Power,
reliability, Power system design, Universal moment
generating function (UMGF).

I.

INTRODUCTION

Optimization is a field of science in which the best values

of parameters of the problem are explored under
specified conditions i.e. basically it targets to find those
particular parameter values which enable an objective
function to generate the minimum or maximum value
under constrained [3] [8] [7]. Many modern optimization
methods and algorithms are nature-inspired algorithms,
that is, they have been developed by mimicking the
successful characteristics of natural systems. Among
these nature-inspired algorithms, most are bio inspired.
Their fundamental characteristics are based on the
evolutionary characteristics of biological systems. A
popular subset of bio-inspired algorithms can be called
swarm intelligence-based algorithms because they have
been developed by drawing inspiration from the so-called
swarm intelligence (SI) [28]
The bat algorithm (BA) is a very interesting approach
recently proposed by Yang [27]. It is based on the nature
behavior of micro bats when looking for food. These bats
use echolocation to guide their search. This feature is
idealized and implemented by the BA to optimization
problems. This idealized idea of bats is that they generate
sound waves with some given frequencies and pulse

Zeblah A.
University of Sidi Bel Abbes,
B.P. 22000, Sidi Jillali, Sidi Bel Abbes, Algeria.
zeblaha@yahoo.fr

rates. When they tend to get close to its prey, the pulse
rate increases, while the loudness decreases. BA
algorithm mimics this idea, i.e., the design parameters are
the bat position and the prey is the objective [5].
II.

POWER SYSTEM PROBLEM

Let us consider a power system containing N

electrical power components connected in series-parallel.
Each component I in turn contain a number of different
components connected in parallel. All components of any
given subsystem belong to different versions.
Components are characterized by their Reliability (Aij),
cost (Cij) and performance (ij). The structure of
subsystem i can be defined by the numbers of parallel
components kij for 1jVi, where Vi is the number of
versions available for component of type i.
The multi-state electrical power system redundancy
optimization problem can be formulated as follows: find
the minimal cost system configuration k1, k2,kN, such
that the corresponding reliability exceeds or equal the
specified reliabilityA0.

C =

Minimize

Subject To
III.

Vi

i =1

j =1

ij

c ij

(1)

A(d , t , k1 , k2 ,..., k N ) A0
ABBREVIATIONS AND ACRONYMS

i, j, l: respectively indices of series, versions and

demand period interval.
N: number of series i
Cij: Cost of electrical component j of type i.
Vi: number of Available electrical components
technologies of type i.
kij : number of occurrence of component j in series i.

i: Performance of power component i.

A. Parallel devices
n

Aij: Reliability of power component j of type i.

(u1 ( z ), u 2 ( z )) = ( Pi z ai ,

A0 : Minimum avaibility required.

i =1

Kmax: maximum number that can be taken from each

component j.
Pi: Performance probability of i-th device.

= Pi Q j z

M: number of demand period interval.

B. Series devices

( u 1 ( z ), u 2 ( z )) = Pi z a ,
i =1

: Operator for parallel device.

PQ
i

min a i , b j

Q
j =1

b
z j

(6)

AVAIBILITY EVALUATION METHOD

R (W ) = P{ f W } = 1 P{ W }
(2)
The procedure used to estimate this index is based on
a modern mathematical technique: the UMGF (or utransform) technique in [24,25,26].This method was first
applied to real power system reliability assessment and
optimization in [15], and represent an extension of
ordinary moment generating function [22]. The UMGF,
in our case, of a discrete variable E is defined as a
polynomial
J

i =1 j =1

The problem defined above is one of combinatorial

optimization problem, it is necessary to enumerate a huge
number of possible system states. Thus, it is required to
use an effective and fast procedure for structure
reliability estimation. As shown above, the main problem
is to evaluate the index R for arbitrary series and parallel
system. The probability that the total capacity of the
electrical power system is not less than a specific load
demand level W must be calculated as:

u ( z ) = Pj z

i =1 j =1

W: Demand levels.

IV.

bj

ai + b j

: Operator for series device.

j =1

(5)

Qi: Performance probability of j-th device.

Q z

gj

Applying composition operators and

consecutively, one can obtain the UMGF of the entire
series-parallel system. To do this we must first determine
the individual UMGF of each device.
V.

BAT METHOD

A. A brief review
The bat algorithm (BA) was first presented in [27] and it
has been applied to benchmark functions. The standard
bat algorithm has been applied in lot of area of
optimization such as economic power dispatch problems
[20], data compression technique [4], bloom filter
optimization [19] and phishing website detection [6]. Kmeans clustering was studied in [13] using bat algorithm
.In [12] [11], authors combined it with fuzzy logic in the
context for e-learning [10], to classify micro array data
[18]. Authors in [17] used bat algorithm to minimize the
makespan and mean flow time to solve multi-processors
flexible flow shop scheduling problems. In the area of
Image Processing, it was used to full body human pose
estimation [1], Image matching [9]. In Other
Applications, we can find its application in: optimal
capacitor placement for loss reduction in distribution
systems [21] and exergy modelling [14] [23].

j =1

(3)
Consider single devices with total failures and each
device i has nominal performance i and reliability Ai.
The UMGF of such an device has only two terms can be
defined as:

u ( z ) = (1 A ) z + A z
0

Ei

= (1 Ai ) + Ai z i

(4)
To evaluate the MSS availability of a series-parallel
system, two basic composition operators are introduced.
These operators determine the polynomial u(z) for a
group of devices.

B. Echolocation of bats
Bats are fascinating animals and most of them have the
advanced capability of echolocation.
The bat algorithm (BA) is inspired from an
idealization of echolocation that can be summarized as
follows:
A virtual bat flies randomly at position(solution)xi
with a velocity vi with a varying frequency or wavelength
Fi and loudness Ai. As it searches and finds its prey, it
changes frequency, pulse emission rate r and loudness.
Search is step up by a local random walk. Selection of
the best continues until certain stop criteria are met.
The algorithm, for simplicity, uses the following
idealized principles:

All bats use echolocation to sense distance, and

they also know the difference between prey and
background barriers in some magical way;

 Bats fly randomly with velocity vi at position xi

with a fixed frequency fmin, varying wavelength
and loudness A0 to search for prey. They can
automatically adjust the wavelength (or
frequency) of their emitted pulses and adjust the
rate of pulse emission r depending on the
proximity of their target;

D. Pseudo code
Using all assumptions the pseudo-code of Bat
algorithm can be as follow:
Algorithm 1. Pseudo code of bat algorithm
0< <1; >0 //for example = =0.9
Initialize fmin=0; fmax=100; //to be ajusted

Although the loudness can vary in many ways, it is

assumed that the loudness varies from a large positive A0
to a minimum constant value Amin [3] [27]. Another
assumption is that the frequency f in range [fmin, fmax]
corresponds to wavelength range [min, max] like
frequency range of [20,500] (in KHz) corresponds to
range of wavelength from [0.7, 17](in mm) [3] [2]. It is
also assumed that f [0, fmax] because higher frequencies
have short wavelengths and travel shorter distance like in
case of bats, so rate of pulse can be simple be in range of
[0,1] where 0 means no pulses and 1 means maximum
rate of pulse emission [27].

Initialize bats population xi and their velocity vi

Define impulsion frequencies fi: fi=fmin+[fmax-fmin]rand( )
Initialize impulsions:ri0 (0,1)
and intensities: Ai (1,2)
Evaluate f(xi) and find global best(gbest) and fgbest
Fitness(i)=f(xi) , best(i)=xi
For iter=1 to max_gen do
adjust frequencies (equation 7)
update velocities (equation 8)

C. Bat Algorithm Parameters

1) Frequency Representation. The frequency f is an
integer or a float number depending on the selected
minimum and maximum frequency.

generate new bats xi (equation 9)

for i=1 to nb do //for all bats
if rand>ri

fi=fmin+[fmax-fmin] rand( )
(7)

generate a new position arround

the best (equation 10)
end if

rand[0,1]

generate new position by flying

randomly (equation 11)

2) Velocity Representation. The velocity v of each bat is

represented as a positive integer number. Velocity
suggests how many of the bats attributes should be
changed at a certain instant. The bats will
communicate with each other through the global best
gbest.
vi=vi+(xi-gbest)fi

(8)

if rand< Ai and f(xi)<fitness(i))

accept new solutions
Increase ri and reduce Ai (equation 12 and
13)
end if

3) Position. the position is updated according to

equation
xi=besti+vi
(9)

sort bats to find the best

done, done , end

VI.

It can be arround the global best with loundness :

xi=gbest + random(-1,1) .Amoy
(10)

A. Test DATA
TABLE I.

Or randomly :
xi=besti + random(-1,1) .Amoy
(11)

Cost

Capacity

C [mln $]

[%]

0.980

0.590

120

0.977

0.535

100

0.982

0.470

85

0.978

0.420

85

0.983

0.400

48

0.920

0.180

31

0.984

0.220

26

HT

0.995

0.205

100

Transf
ormer

0.996

0.189

92

0.997

0.091

53

Versions

Power
1
Units

5) Pulse Rate Representation. Pulse rate ri value will

play a role in whether a local search procedure
around the global best solution should be conducted
or skipped.
ri= ri0(1-exp(-.iter)), ri0 Initial pulse rates, >0
(13)

CHARACTERISTICS OF RELIABILITY SYSTEM

COMPONENTS ON THE MARKET [16]
Reliability

Subsystems

4) Loudness Representation. A is the average sound

loudness of all the bats. It has a range, between the
maximum and minimum loudness. It can decrease.
Ai= Ai , ]0,1[
(12)

COMPUTATIONAL RESULTS

0.997

0.056

28

VII. CONCLUSION

0.998

0.042

21

0.971

7.525

100

HT

0.973

4.720

60

In this study, the algorithm bat bio inspired, uses the

basic principles of the art of bat inspired. Bat was used
for solving constrained optimization problems.

lines

0.971

3.590

40

0.976

2.420

20

0.977

0.180

115

0.978

0.160

100

0.978

0.150

91

MT

0.983

0.121

72

Transf
ormer

0.981

0.102

72

0.971

0.096

72

0.983

0.071

55

0.982

0.049

25

0.977

0.044

25

0.984

0.986

128

MT

0.983

0.825

100

Lines

0.987

0.490

60

0.981

0. 475

Abdel-Rahman, E. M., Ahmad, A. R., Akhtar, S., A metaheurisic

bat-inspired algorithm for full body human pose estimation, in:
Ninth Conference on Computer and Robot Vision, 2012, pp. 369
375.

[2]

Altringham J. D. , T. McOwat, and L. Hammond, Bats: biology

and behaviour, Oxford University Press Oxford, 1996.

[3]

Arora S., and S. Singh, A Conceptual Comparison of Firefly

Algorithm, Bat Algorithm and Cuckoo Search. ICCCCM. 2013.

51

[4]

. PARAMETERS OF THE CUMULATIVE DEMAND CURVE

[16]

Banu F., A., Chandrasekar, C., An optimized appraoch of

Modified bat algorithm to record deduplication, Int. Journal of
Computer Applications, 2012, Vol. 62, No. 1, pp. 1015.

[5]

Bora T. C., L. d.S. Coelho, and L.Lebensztajn, Bat-Inspired

Optimization Approach for the Brushless DC Wheel Motor
Problem, IEEE Transactions on magnetics, February, 2012, vol.
48, NO. 2.

[6]

Damodaram, R., Valarmathi, M. L.,. Phishing website detection

and optimization using modified bat algorithm, Int. J. Engineering
Research and Applications, 2012, Vol. 2, No. 1, pp. 870876.

[7]

Deb K., A. Pratap, S. Agarwal, and T. Meyarivan, A fast and

elitist multiobjective genetic algorithm: Nsga-ii, Evolutionary
Computation, IEEE Transactions on, 2002, vol. 6, no. 2, pp. 182
197,.

Wm [%]

100

80

50

20

Tm [h]

4203

788

1228

2536

B. Optimization result and discussion

One of the ways to enhance system reliability and to
adapt system capacity to meet the demand is to optimize
the power structure. For each study period a new redesign
is obtained by the proposed bat algorithm for different
values of A0 (0.95, 0.97 and 0.99). Table 3 illustrates the
computed cost and availability index corresponding to
their new design.
TABLE III.

REFERENCES
[1]

TABLE II.

The proposed method allows a convenient way to

solve a wide variety of cases of optimization problems of
reliability of power multistate systems, unlimited number
of versions of components in parallel. The algorithm
developed to solve this problem is based on the use of the
universal moment generating function UMGF and the bat
method.

.OPTIMAL SOLUTION OBTAINED BY BAT ALGORITHM

Reliability
Constraint
A0

0.950

0.970

0.99

Bat algorithm
A

0,964

0,971

0.997

Optimal
Power
Design

11,87

6[6]-2[7]
5[4]-5
2-3
6[8]
3[4]

14,5

16.96

5-5[7]
6[2]-4
1
7[4]-9
2[4]-5[3]
7[6]
4[3]-4[5]
2[2]-3
6-2[7]
4[4]

[8]

Dorigo M. and G. Di Caro, New ideas in optimization, 1999.

[9]

Du, Z. Y., Liu B.,. Image matching using a bat algorithm with
mutation, Applied Mechanics and Materials, 2012, Vol. 203, No.
1, pp. 8893.

[10] Khan, K., and Sahai, A.,. A comparison of BA, GA, PSO, BP and
LM for training feed forward neural networks in e-learning
context, Int. J. Intelligent Systems and Applications [IJISA],
2012a, Vol. 4,No.7, pp. 2329.
[11] Khan, K., and Sahai, A.,. A fuzzy c-means bi-sonar-based
Metaheuristic optimization algorithm, INt. J. of Interactive
Multimedia and Artificial Intelligence, 2012b, Vol. 1, no. 7, pp.
2632.
[12] Khan, K., Nikov, A., Sahai A., 2011, A fuzzy bat clustering
method for ergonomic screening of office workplaces, S3T 2011,
Advances in Intelligent and Soft Computing, 2011, Volume
101/2011, pp. 5966.
[13] Komarasamy, G., and Wahi, A., An optimized K-means clustering
tech- niqueusing bat algorithm, European J. Scientific Research,
2012. Vol. 84, No. 2, pp. 263-273.
[14] Lemma, T. A., Bin Mohd Hashim, F., Use of fuzzy systems and
Bat algorithm for exergy modelling in a gas turbine generator,

IEEE Colloquium on Humanities, Science and Engineering

CHUSER2011, 5-6 Dec. 2011, pp. 305310.
[15] Levitin G., A. Lisniaski, H. Ben-haim, D. Elmakis, Redundancy
optimization for series-parallel multi-state system, IEEE
Transaction on reliability, 1998,47:2, pp-165-172.
[16] Levitin G: Multistate Series-Parallel System ExpansionScheduling subject to Availability Constraints: IEEE Transactions
on reliability, March, 2000, VOL. 49, NO. 1.
[17] Marichelvam, M. K., and Prabaharam, T., A bat algorithm for
Realistic hybrid flowshop schedulihng problems to minimize
makespan and mean flow time, ICTACT Journal on Soft
Computing, Vol. 3, No. 1,2012, pp. 428433.

[21] Reddy, V. U., Manoj, A., Optimal capacitor placement for

lossReduction in distribution systems using bat algorithm, IOSR
Journal of Engineering, 2012. Vol. 2, No. 10, pp. 2327.
[22] Ross S. M., Introduction to probability model, Academic Press,
York, 1993.
[23] Tamiru, A. L., Hashim, F. M.,. Application of bat algorithm and
Fuzzy systems to model exergy changes in a gas turbine, in:
Artificial
Intelligence,
Evolutionary
Computing
and
Metaheuristics [Eds. X. S.Yang], Studies in Computational
Intelligence, 2013, Vol. 427, Springer, Heidelberg, pp. 685719.
[24] Ushakov I.A., A Universal generating function, Sov. Journal,
computer system science, 1986, 24:6, pp-37-49,.

[18] Mishra, S., Shaw, K., Mishra, D, A new meta-heuristic bat

Inspired classification approach for microarray data, Procedia
Technology, 2012, Vol. 4, pp. 802806.

[25] Ushakov I.A., Optimal standby problems and a universal

generating Function, Sov. Journal, computer system science,
1987 Vol N25, pp- 61-73.

[19] Natarajan, A., Subramanian, S., Premalatha, K., A comparative

study of cuckoo search and bat algorithm for Bloom filter
optimisation in spam filtering, Int. J. Bio-Inspired Computation,
2012, Vol. 4, No.pp. 8999.

[26] Ushakov I.A.: Universal generating function, Sov. J. Compt. Syst.

Sci, 1986, 24:5, pp-118-129,.

[20] Ramesh, B., Mohan, V. C. J., Reddy, V. C. V., Application of bat

algorithm for combined economic load and emission dispatch, Int.
J of Electricl Engineering and Telecommunications, 2013, Vol. 2,
No 1, pp.1-9.

[27] Yang X. S., A new bat-inspired algorithm, Nature Inspired

Cooperative Strategies for Optimization, vol. 284, Springer,
Studies Computational Intelligence, 2010, pp. 6574.
[28] Yang Xin-she, and He Xing-shi, Swarm intelligence and smart
optimization algorithms. Basic sciences journal of textile, Sept.
,2013, Vol. 26 ,No. 3.