A Study of Firefly Algorithm and its Application in Non-Linear Dynamic Systems

@ IJTSRD | Available Online @ www.ijtsrd.com
ISSN No: 2456
International
Research
A Study of
Application in Non
Gautam Mahapatra
Department of Computer Science
and Engineering, Birla Institute of
Technology, Jharkand, India
ABSTRACT
Firefly Algorithm (FA) is a newly proposed
computation technique with inherent parallelism,
capable for local as well as global search, meta
heuristic and robust in computing process. In this
paper, Firefly Algorithm for Dynamic System
(FADS) is a proposed system to find instantaneous
behavior of the dynamic system within a single
framework based on the idealized behavior of the
flashing characteristics of fireflies. Dynamic sy
where flows of mass and / or energy is cause of
dynamicity is generally represented as a set of
differential equations and Fourth Order Runge
(RK4) method is one of used tool for numerical
measurement of instantaneous behaviours of dynamic
system. In FADS, experimental results are
demonstrating the existence of more accurate and
effective RK4 technique for the study of dynamic
system.
Keywords: FADS, Firefly Flashing, First Order
Differential Equations, RK4
Introduction
S. X. Yang of Cambridge University in the year 2008
has proposed new nature inspired, swarm intelligence
based optimization method on the idealized behavior
of the flashing characteristics of the common fireflies
like Photinus [1].This algorithm assumes th
firefly is a solution of the optimization problem and
quality of the solution is proportional to the intensity
of glowing, and consequently brighter firefly attracts
its partners which may explore the search space for
@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 2 | Jan-Feb 2018
ISSN No: 2456 - 6470 | www.ijtsrd.com | Volume
International Journal of Trend in Scientific
Research and Development (IJTSRD)
International Open Access Journal
dy of Firefly Algorithm and its
Application in Non-Linear Dynamic Systems
Srijita Mahapatra
Department of Electronics &
Communication Engineering,
UEM, Kolkata, India
Soumya Banerjee
Firefly Algorithm (FA) is a newly proposed
computation technique with inherent parallelism,
capable for local as well as global search, meta-
process. In this
paper, Firefly Algorithm for Dynamic System
proposed system to find instantaneous
behavior of the dynamic system within a single
framework based on the idealized behavior of the
flashing characteristics of fireflies. Dynamic system
where flows of mass and / or energy is cause of
dynamicity is generally represented as a set of
differential equations and Fourth Order Runge-Kutta
(RK4) method is one of used tool for numerical
measurement of instantaneous behaviours of dynamic
m. In FADS, experimental results are
demonstrating the existence of more accurate and
effective RK4 technique for the study of dynamic
FADS, Firefly Flashing, First Order
S. X. Yang of Cambridge University in the year 2008
has proposed new nature inspired, swarm intelligence
based optimization method on the idealized behavior
of the flashing characteristics of the common fireflies
like Photinus [1].This algorithm assumes that each
firefly is a solution of the optimization problem and
quality of the solution is proportional to the intensity
of glowing, and consequently brighter firefly attracts
its partners which may explore the search space for
the better solution. Yang has
and effectiveness for different multi
nonlinear and multimodal optimization problems [2].
In the year 2010 he has tested & designed several
benchmark functions where this algorithm is
outperforming other existing meta
algorithms [3]. Lukasik et al., 2009 has shown how
FA performs better for continuous constraints
optimization problems [4]. After the introduction this
efficient computation technique it is found in
literature that different applications[6] have
to solve various common NP-
clustering[8], bi-level optimization[7], non
optimization problem [9], load forecasting problem
[10] and multi-objective job-shop scheduling problem
[12].Different improvements based chaos
there is hybridization with support
(SVM) [10] have been also emerging as more
efficient optimization technique. Yang has shown in
Levy Flight distribution is outperforming than
Gaussian distribution in randomized characteristics o
fireflies [13].Dynamic systems are generally
represented as a system of differential equation as
parameters are varying with time and interrelated
properties [14-20] and to measure different
instantaneous property values the numerical solution
of such differential equations are more practical and
different numerical analysis are performed for the
same [16,17]. In this present work FA
studied and used to know the instantaneous behaviors
of such dynamic systems modeled as system first
order differential equations [18]. We are looking for
Feb 2018 Page: 542
6470 | www.ijtsrd.com | Volume - 2 | Issue – 2
Scientific
(IJTSRD)
International Open Access Journal
Linear Dynamic Systems
Soumya Banerjee
shown higher efficiency
and effectiveness for different multi-dimensional,
nonlinear and multimodal optimization problems [2].
In the year 2010 he has tested & designed several
benchmark functions where this algorithm is
outperforming other existing meta-heuristic
algorithms [3]. Lukasik et al., 2009 has shown how
FA performs better for continuous constraints
optimization problems [4]. After the introduction this
efficient computation technique it is found in
literature that different applications[6] have developed
-Hard problems like data
level optimization[7], non-linear
optimization problem [9], load forecasting problem
shop scheduling problem
[12].Different improvements based chaos [11] and
there is hybridization with support-vector machine
(SVM) [10] have been also emerging as more
efficient optimization technique. Yang has shown in
Levy Flight distribution is outperforming than
Gaussian distribution in randomized characteristics of
fireflies [13].Dynamic systems are generally
represented as a system of differential equation as
parameters are varying with time and interrelated
20] and to measure different
instantaneous property values the numerical solution
fferential equations are more practical and
different numerical analysis are performed for the
same [16,17]. In this present work FA has been
studied and used to know the instantaneous behaviors
of such dynamic systems modeled as system first
ntial equations [18]. We are looking for

International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 2 | Jan-Feb 2018 Page: 543
numerical solutions, so we are focusing on numerical
analysis based solutions and successfully proposed
more accurate formula for the RK4which has been
derived using Taylor Series[19] and embedded
successfully in our proposed FADS.
Firefly Algorithm
Firefly Colony
Firefly Algorithm is most general optimization
algorithm which can handle multi-dimensional, linear,
non-linear, multimodal, non-convex, non-
differentiable and continuous constraints optimization
problem. Yang 2010 has shown that the Particle
Swarm Optimization (PSO) is a special case it [3, 8].
FA has designed after inspiring on the unique short
and rhythmic flashing pattern generated to attract
mating partners and / or potential prey by the most of
the fireflies’ species. The radiation of this flashing
light obeys the mathematical relation: 𝐼 ∝ 1
𝑟 where
𝐼 is the intensity of flashing light at distance𝑟, also
due to absorption this intensity again decreases and
combined effect limits the fireflies to communicate
within few hundred meter of distance. Yang
formulated this flashing behavior as an objective
function and developed a new bio-inspired, swarm
intelligence based meta-heuristic optimization
algorithm.
Modeling the Attraction of Firefly
This flashing behavior of fireflies can be idealized and
following three idealized rules for the firefly colony
system were proposed by Yang 2010:
Rule I: All fireflies are unisex, so that one firefly is
attracted to other fireflies regardless of their sex
Rule II: Attractiveness is proportional to their
brightness, thus for any two flashing fireflies, the less
bright one will move towards the brighter one. The
attractiveness is proportional to the brightness and
they both decrease as their distance increases. If no
one is brighter than a particular firefly, it moves
randomly.
Rule III: The brightness or light intensity of a firefly
is affected or determined by the landscape of the
objective function to be optimized.
Using these three ideal rules and natural absorption of
the media it is possible to formulate the attractiveness
of the firefly as follows:
𝛽(𝑟) = ≈ 𝛽 𝑒 (Gaussian form)
(1)
where 𝛾 is the absorption co-efficient (or attenuation
factor) of the media and 𝛽 is the initial attractiveness.
From the inverse square law of light intensity it is
possible to write for practical media:
𝐼(𝑟) = ≈ 𝐼 𝑒 (Gaussian form)
(2)
Here one is added to the denominator to avoid the
divisibility by zero and 𝐼 is the intensity at light
source.
Modeling the Movement of Firefly System
The Cartesian Distance between two fireflies 𝐹
and 𝐹 , are presently lying in 𝑛-dimensional search
space at 𝑋 and 𝑋 respectively is given in following
equation (3) and used in the new position calculation
formula:
𝑟 = 𝑋 − 𝑋 = ∑ (𝑥 , − 𝑥 , ) (3)
Fireflies are moving due to attractiveness generated
for rhythmic pattern of flashing can be written as:
𝑋 = 𝑋 + 𝛽 𝑒 , 𝑋 − 𝑋 + 𝛼(𝑟𝑎𝑛𝑑 − ) (4)
Here firefly 𝐹 is the brighter or more attractive than
firefly𝐹 , so firefly 𝐹 is moving towards 𝐹 and the
position of 𝐹 i.e. 𝑋 is updating using equation (4)
and it is also updating the intensity of 𝐹 using
equation (2) and hence the objective functions of the
optimization problem. Attractiveness is updated using
formula (1). For maximizing problems, intensity is
directly proportional to the objective function value
(𝐼(𝑋) ∝ 𝑓(𝑋)) and for minimization it will be
inversely proportional (𝐼(𝑋) ∝ 1/𝑓(𝑋)). In this
position changing equation second part is due to
attractiveness while the third part is for
randomization, like mutation in GA [23] to avoid
stuck-at-local optima problem, 𝛼 is the randomization
parameter and 𝑟𝑎𝑛𝑑 is a random number generator
uniformly distributed in[0,1].

Algorithm for Simulation of Firefly Colony
Based on above mentioned modeling of attractiveness
and then the movement of firefly due to this
attractiveness with certain randomization due
environmental interactions here is the pseudo code of
the firefly system to solve the optimization problem.
Algorithm Firefly Algorithm (FA)
(*𝑚𝑎𝑥𝑓(𝑋) is the objective function,𝑋 = (𝑥 , 𝑥 , … , 𝑥 ), position or solution vector in the constraint 𝑛 −
𝑑𝑖𝑚𝑒𝑛𝑠𝑖𝑜𝑛𝑎𝑙 search space*)
Step 1: [Initialize]
Create 𝑁 number of Fireflies 𝐹 = (𝐹 , 𝐹 , … , 𝐹 )
Place them at random positions 𝑆 = (𝑋 , 𝑋 , … , 𝑋 )in the feasible solution region.
Evaluate intensities 𝐼 = (𝐼 , 𝐼 , … , 𝐼 ) for all 𝑁 Fireflies
Initialize light absorption co-efficient (𝛾)
Set maximum number of generation (𝐺𝑒𝑛 )
Set current generation number (𝑡 = 1)
Step 2: [Observe the swarming effect of the fireflies’ colony for optimization]
while(𝑡 < 𝐺𝑒𝑛 )do
begin
for𝑖 = 1to𝑁do
begin
for𝑗 = 1to𝑖do
begin
if(𝐼 < 𝐼 )then
begin
2.1 Move 𝐹 towards 𝐹 in 𝑛 − 𝑑𝑖𝑚𝑒𝑛𝑠𝑖𝑜𝑛𝑎𝑙 space using (3)
2.2 Update light intensity using and Evaluate new solutions (2)
end
Update attractiveness using (1)
end
end
Rank the fireflies using Sort operation
Find the current best firefly
Move all fireflies randomly in the search space
Update generation counter 𝑡 = 𝑡 + 1
end
Step 3: [Finished]
Return best Solution Vector and Optimized Value
Here the attractiveness depends on both flashing intensity and objective function value and distance among the
fireflies which monotonically decreases with distance. But, for FA visibility is adjustable and more versatile in
attractiveness variations, and this leads to higher mobility so the search space is explored more rapidly and
execution time decreases.
Dynamic System
Dynamical system is a system in which any
measuring parameter is varying with time. The world
is full-of dynamic systems; all populations, human
and otherwise, are dynamic systems, epidemics are
dynamic systems, economies at all scales are dynamic
systems, weather is complex dynamic system.
Anywhere there are flows of mass and / or energy
observed is a dynamic system. For the mathematical
study of such varying system, instantaneous rate of
changes parameters are modeled as system of time

dependent Differential Equations. To measure any
such variable property or characteristic we need to
solve all such differential equations after imposing
boundary conditions. In all cases solving such
equations are not so simple and / or in some cases it is
impossible and generally numerical approximate
solutions are preferred over complex symbolic
mathematical expression [16-20].Euler, Euler-
Cauchy, Taylor Series, Runge-Kutta Second Order &
Fourth Order etc. methods most common to use for
numerical solution. Runge-Kutta Fourth Order
Method is most popular method because it is
advanced form of general Taylor Series Method [18-
20] and mostly used for the dynamic systems which
can be modeled as first order differential equations.
Example of Dynamic System
Fluid system in nature is very common dynamic system, in the year 1872, Joseph Boussinesq[21,22] has
proposed very simplified model for water waves known as solitary wave or soliton still now it considered as
reference and valid for weakly non-linear and fairly long waves. Following system of first order differential
equations [equations no. (5) & (6)] known as the Boussinesq equations arerepresenting this model.
𝑫
= 𝒆 𝟑 × 𝑩 (5)
𝑩
= 𝑫 × 𝑩 (6)
𝑫 = (𝐷 , 𝐷 , 𝐷 ) - is horizontal displacement and𝑩 = (𝐵 , 𝐵 , 𝐵 ) is vertical displacement, is a non-
dimensional constant and 𝒆 𝟑=
0 −1 0
1 0 0
0 0 0
.
= − 𝐵 (7)
= 𝐵 (8)
= 0 (9)
= (𝐷 𝐵 − 𝐷 𝐵 ) (10)
= (𝐷 𝐵 − 𝐷 𝐵 ) (11)
= (𝐷 𝐵 − 𝐷 𝐵 ) (12)
In above differential equations[equation no. (7) – (12)] we are showing component of different velocities.For
time dependent instantaneous behavior these equations must be solved and use of RK4 has been shown by
Desale et al in the year 2013 [21].
Runge-Kutta Methodfor Dynamic Systems
To study the dynamic behavior of any system the gradient or variation relations of the system are expressed as a
set of differential equation of equal or different orders. Most of the simple form is the first order form called as
Ordinary Differential Equations (ODE). The solutions of these ODEs describe the functional behavior of the
system and for this purpose different calculus oriented methods like integrating factor etc are used. The RK4
method is very efficient, as it avoids all kind of differentiations required in Taylor Series method. In the
derivation of the RK4 method we get eleven independent algebraic relations for thirteen different coefficients
and these are forming a system of equations with rank less than number unknowns. To find the value of
unknowns Runge-Kutta had assigned empirically fixed values for two such variables and then solve the system
of equation using numeric algorithm [19], finally we got the most popular recursive equation for evaluation.Let
for a dynamic system there are 𝑁 numbers of properties to measure and these are represented by the time
dependent vector𝑌(𝑡) = {𝑦 (𝑡)} and 𝑌 = 𝑌(𝑡 ) = {𝑦 (𝑡 )} be initially measures or know parameters.
Then the instantaneous dynamic behavior represented by 1st
order differential equation form as follows:

( )
= 𝐹 𝑡, 𝑌(𝑡) = {𝑓 (𝑡, 𝑌(𝑡))} with initial condition (𝑡 , 𝑌 ) (13)
And the solution will be of the form𝑌 = 𝐺(𝑡) = {𝑔 (𝑡)} . Different complex forms of 𝐹(𝑡, 𝑌(𝑡)) are possible
and to get 𝐺(𝑡) different techniques are used and for numeric solution different approximate numeric
algorithms are used. Most popular numeric method is RK4. This method is derived from Taylor Series method
after approximating it up to 4th order term in the series and Runge-Kutta had assumed the recursive solution of
the following form [19, 20]:
𝑦 = 𝑦 + 𝑤 𝑘 + 𝑤 𝑘 + 𝑤 𝑘 + 𝑤 𝑘 , where (14)
𝑘 = ℎ𝑓 (𝑡 , 𝑦 (t), 𝑦 (t), … , 𝑦 (t) ) (15)
𝑘 = ℎ𝑓 (𝑡 + 𝑎 ℎ, 𝑦 + 𝑏 𝑘 , 𝑦 + 𝑏 𝑘 , … , 𝑦 + 𝑏 𝑘 ) (16)
𝑘 = ℎ𝑓 (𝑡 + 𝑎 ℎ, 𝑦 + 𝑏 𝑘 + 𝑏 𝑘 , 𝑦 + 𝑏 𝑘 + 𝑏 𝑘 , … , 𝑦 + 𝑏 𝑘 + 𝑏 𝑘 ) (17)
𝑘 = ℎ𝑓 𝑡 + 𝑎 ℎ, 𝑦 + 𝑏 𝑘 + 𝑏 𝑘 + 𝑏 𝑘 , 𝑦 + 𝑏 𝑘 + 𝑏 𝑘 + 𝑏 𝑘 , … , 𝑦 + 𝑏 𝑘 + 𝑏 𝑘 + 𝑏 𝑘 , ∀𝑖 =
1, … , 𝑁 (18)
After coefficient comparison with the Taylor’s series method following eleven equations for thirteen
unknowns{𝑎 } , {𝑏 } and{𝑤 } of above equations (14) – (18) have generated, for which any conventional
method is not suitable to solve.
𝑏 = 𝑎 , 𝑏 + 𝑏 = 𝑎 , 𝑏 + 𝑏 + 𝑏 = 𝑎 (19)-(21)
𝑤 + 𝑤 + 𝑤 + 𝑤 = 1, 𝑤 𝑎 + 𝑤 𝑎 + 𝑤 𝑎 = (22)-(23)
𝑤 𝑎 + 𝑤 𝑎 + 𝑤 𝑎 = , 𝑤 𝑎 + 𝑤 𝑎 + 𝑤 𝑎 = (23)-(25)
𝑤 𝑎 𝑏 + 𝑤 (𝑎 𝑏 + 𝑎 𝑏 ) = , 𝑤 𝑎 𝑎 𝑏 + 𝑤 𝑎 (𝑎 𝑏 + 𝑎 𝑏 ) = (26)-(27)
𝑤 𝑎 𝑏 + 𝑤 (𝑎 𝑏 + 𝑎 𝑏 ) = , 𝑤 𝑎 𝑏 𝑏 = (28)-(29)
These eleven equations [equations no (19)-(29)]are forming a system of equations for which we have thirteen
unknowns and eleven relations, so the rank is less than number of unknowns [18]. Runga-Kutta equalized the
rank after assuming𝑎 = , 𝑏 = 0, then solved these equations and finally derived the popular recursive
relation from equation (14) as:
𝑦 = 𝑦 + (𝑘 + 2𝑘 + 2𝑘 + 𝑘 ) (30)
with the solutions𝑎 = , 𝑎 = 1, 𝑏 = 𝑏 = , 𝑏 = 𝑏 = 0, 𝑏 = 1, 𝑤 = 𝑤 = , 𝑤 = 𝑤 =
Firefly Algorithm for Dynamic System (FADS)
For the design of FADS for general dynamic system we use these thirteen parameters as 13-dimensional
position vector: 𝑋 = (𝑎 , 𝑎 , 𝑎 , 𝑏 , 𝑏 , … , 𝑏 , 𝑤 , 𝑤 , 𝑤 , 𝑤 ) ∈ 𝑅 for the different fireflies of FA. Above
equations (19) to (29) are nothing but equality constraints to satisfyfor best possible position vector (𝑋) to get
the more accurate recursive formula represented in equation (14). For fitness or objective function calculation
thesquare difference of left hand sidealgebraic expressions of these thirteen parameters with right hand side

constant for these eleven equations for certain position vector or firefly is considered as square error 𝛿 and sum
of these square errors as objective function and we have to minimize it towards zero value:
𝑀𝑖𝑛𝑖𝑚𝑖𝑧𝑒𝐼(𝑋) = ∑ 𝛿 (31)
For Algorithm Firefly Algorithm (FA) mentioned above we have incorporated position vector of the fireflies
along with square error measurement in equation (31) as intensity of these fireflies. After getting more accurate
13-parameters values recursive formula (14) with other formulas for the calculations of 𝑘 ′𝑠 mentioned
equations (15)-(18) are used for numerical values of characteristics parameters on time for a dynamic system.
Experimental Results
Proposed FADS has been implemented using C++ programming language and tested with FA related
parameters setting: Number of Firefly (N) = 20, Number of Generation (𝐺𝑒𝑛 ) = 1000, 𝛽 = 0.2 , and
𝛾 = 1.0. This has been executed for 30 independent runs with said parameter values and we get different
values for these thirteen variables which were found by Runge-Kutta and these are shown in Table 1. From this
we can conclude that here is a new form of the recurrence relation which is outperforming the existing system.
Table 1.
Parameters RK Solution Average FADS Solution
𝑎 0 0.197491
𝑎 0.665014
𝑎 1 0.813665
𝑏 0.201191
𝑏 -0.737769
𝑏 1.409750
𝑏 0 1.005670
𝑏 0 -0.864207
𝑏 1 0.667915
𝑤 0.109450
𝑤 0.277182
𝑤 0.362196
𝑤 0.248217
To study the performance of this dynamic system analyzer we have tested an expontial type system defined by
the following first order differential equation:
( )
=
( )
= 𝐹 𝑡, 𝑌(𝑡) = 𝑡𝑦(𝑡) with initial condition (𝑡 , 𝑌 ).
Fig. 1. 3D Comparative Study

The effectiveness of the system is presented by a comparative study of existing solution and standard C++
language library with the proposed FADS and this has been shown graphically in above Fig. 1.
Fig 2: FA Convergence
The convergence of FA to achieve the best possible parameters values is also shown in Fig. 2.
Conclusion
This work has proposed how the robust and efficient
FA optimization technique can be used to get numeric
solutions of complex dynamic system represented by
a set of differential equations form. This also suggests
more accurate alternative of RK4 and also suggesting
simpler way to develop RK-N where 𝑁 > 4 for more
accurate behavioral study of complex dynamic
system.
Acknowledgements
The authors wish to acknowledge the support of the
Post Graduate Teaching &Research Council of
Asutosh College.
References
1. Yang, X. S., 2008. Nature-inspired meta-heuristic
algorithms. Luniver press, UK.
2. Yang, X. S., 2009.Firefly algorithms for
multimodal optimization. In Stochastic
algorithms: foundations and applications,
Springer Berlin Heidelberg, p. 169-178.
3. Yang, X. S., 2010. Firefly algorithm, stochastic
test functions and design optimization.
International Journal of Bio-Inspired
Computation, 2(2), 78-84.
4. Yang, X.S., 2010. Firefly algorithm. Engineering
optimization, pp.221-230.
5. Łukasik, S., and Żak, S., 2009. Firefly algorithm
for continuous constrained optimization tasks.
In Computational Collective Intelligence.
Semantic Web, Social Networks and Multiagent
Systems. Springer Berlin Heidelberg, p. 97-106.
6. Yang, X.S. and He, X., 2013. Firefly algorithm:
recent advances and applications. International
Journal of Swarm Intelligence, 1(1), pp.36-50.
7. Mahapatra, G. and Banerjee, S., 2014. Bilevel
Optimization using Firefly Algorithm."5th
International Conference on Electronics
Engineering & Computer Science (IEMCON).
Elsevier Publications, pp. 313-322.
8. Senthilnath, J., Omkar, S.N. and Mani, V., 2011.
Clustering using firefly algorithm: performance
study. Swarm and Evolutionary
Computation, 1(3), pp.164-171.
9. Pal, S.K., Rai, C.S. and Singh, A.P., 2012.
Comparative study of firefly algorithm and
particle swarm optimization for noisy non-linear
optimization problems. International Journal of
intelligent systems and applications, 4(10), p.50.
10. Kavousi-Fard, A., Samet, H. and Marzbani, F.,
2014. A new hybrid modified firefly algorithm
and support vector regression model for accurate
short term load forecasting. Expert systems with
applications, 41(13), pp.6047-6056.
11. Gandomi, A.H., Yang, X.S., Talatahari, S. and
Alavi, A.H., 2013. Firefly algorithm with
chaos. Communications in Nonlinear Science and
Numerical Simulation, 18(1), pp.89-98.
12. Karthikeyan, S., Asokan, P., Nickolas, S. and
Page, T., 2015. A hybrid discrete firefly algorithm
for solving multi-objective flexible job shop
scheduling problems. International Journal of
Bio-Inspired Computation, 7(6), pp.386-401.
13. Yang, X.S., 2010. Firefly algorithm, Levy flights
and global optimization. Research and

development in intelligent systems XXVI, pp.209-
218.
14. Stuart, A.M., 1994. Numerical analysis of
dynamical systems. Acta numerica, 3, pp.467-572.
15. Joseph, D.D., 1966. Nonlinear stability of the
Boussinesq equations by the method of
energy. Archive for Rational Mechanics and
Analysis, 22(3), pp.163-184.
16. Guckenheimer, J., 2002. Numerical analysis of
dynamical systems. Handbook of dynamical
systems, 2, pp.345-390.
17. Antoulas, A.C., 2005. Approximation of large-
scale dynamical systems. Society for Industrial
and Applied Mathematics.
18. S. Gilbert,2007, Linear Algebra and its
Applications. Pacific Grove: Brooks Cole
19. Mathews, J.H., 2014. Numerical methods for
mathematics, science and engineering. Prentice-
Hall.
20. W. H. Press, S.A. Teukolsky, W.T. Vetterling, and
B.P. Flannery, 2007, Numerical Recipes The art
of scientific computing, 3rd Edition, Cambridge
University Press.
21. Desale, B.S. and Dasre, N.R., 2013. Numerical
solution of the system of six coupled nonlinear
ODEs by Runge-Kutta fourth order
method. Applied Mathematical Sciences, 7(6),
pp.287-305.
22. Shi, F., Kirby, J.T., Harris, J.C., Geiman, J.D. and
Grilli, S.T., 2012. A high-order adaptive time-
stepping TVD solver for Boussinesq modeling of
breaking waves and coastal inundation. Ocean
Modelling, 43, pp.36-51.
23. Goldberg, D.E., 2006. Genetic algorithms.
Pearson Education India

A Study of Firefly Algorithm and its Application in Non-Linear Dynamic Systems

More Related Content

What's hot (17)

Similar to A Study of Firefly Algorithm and its Application in Non-Linear Dynamic Systems (20)

More from ijtsrd (20)

Recently uploaded (20)

A Study of Firefly Algorithm and its Application in Non-Linear Dynamic Systems