Substitution Box Design Based on Chaotic Maps

and Cuckoo Search Algorithm

1st Tanveer Akhtar 2nd Nizamud Din 3rd Jamal Uddin
Department of Computer Science Department of Computer Science Department of Computer Science
Iqra National University University of Chitral Qurtaba University of Science & IT
Peshawar, Pakistan Chitral, Pakistan Peshawar, Pakistan
[email protected] [email protected] [email protected]

Abstract—The widespread use of the Internet for transmission This sensitivity to the initial conditions contributes to produce
of data demands the well-defined data protection algorithms. the secure S-box. Moreover, real-time applications demand
Symmetric block cipher protects the digital information from fast encryption algorithms. Encryption algorithms based on
the un-authorized recipient. The cryptographic strength of sym-
metric cipher mainly depends on substitution box (S-box). So chaotic systems perform fast encryption and require less
the main challenge is to create an S-box with higher value computational complexity [3]. S-box is the only component
of nonlinearity. Meta-heuristic optimization algorithm has been that resists against differential attacks [4]. Differential
applied to create a secure S-box. In this paper, a design scheme approximation probability of an S-box defines the level of
based on chaotic maps and Cuckoo Search Algorithm (CSA) is security shield against differential cryptanalysis. Chaotic
proposed to enhance the non-linearity of S-box. Chaotic maps are
widely used to create secure S-box due to their aperiodic nature. map generates S-box with low differential probability which
The proposed idea produced an S-box with non-linearity 109.25, promises strength against differential attacks. The S-box
which strengthens the cipher security against linear attacks. based on chaotic map achieves speed and security in the
The results are compared with the recently created S-boxes and encryption algorithm [5]. Chaotic maps attracted the attention
the S-boxes obtained using meta-heuristic algorithms such as of many researchers in the creation of S-box. With the
Ant Colony Optimization (ACO), Genetic Algorithm (GA) and
Firefly algorithm (FA). The obtained S-box is evaluated against encouraging results from one-dimensional chaotic maps [6],
cryptographic criteria such as bijectivity, nonlinearity, strict 2D chaotic map [7] and 3D chaotic map produced desirable
avalanche criteria, bit independence criteria, linear probability results in the image encryption algorithms [8]. Usually a
and differential probability. single S-box is used in the image encryption technique. In
Index Terms—S-box, Cuckoo Search Algorithm, Chaotic [9], the authors applied three different S-boxes for red, green
Maps,Cryptosystem, Nonlinearity
and blue components of an image. Unlike chaos-based S-box,
the author [10] used algebraic technique to produce nonlinear
I. I NTRODUCTION
component in the encryption algorithm.
Cryptography is one of the techniques that play fundamental
role in the security of digital information. It can be broadly The confusion factor in the design of secure encryption
categorized into asymmetric cipher and symmetric cipher [1]. algorithm only depends on S-box. However, this dependency
Asymmetric block cipher applies different keys on a block of on S-box is susceptible to cipher attacks [11]. Therefore, the
data for encryption and decryption. Public key encrypts the author synthesized the combined effect of S-box and Arnold
plain text into cipher text and private key reverts this process. transformation to strengthen the cryptographic properties
Symmetric block cipher uses same key for encryption and against cipher attacks [12]. Chaotic maps create continuous
decryption. The substitution box (S-box) plays very signifi- output. However, the computers operate on discrete inputs.
cant role in symmetric cryptographic systems such as Data So the output value of chaotic maps needs to be converted
Encryption Standard (DES), International Data Encryption into discrete format. The author presented a scheme based
Algorithm (IDEA) and Advanced Encryption Standard (AES). on discrete chaotic maps to obtain secure S-box without the
Shannon [2] showed that a strong crypto system must have two need of any approximation [13]. Chaos-based S-boxes are
fundamental properties of confusion and diffusion. In block favorite to the researchers due to their ease of implementation
cipher, S-box provides the property of confusion that hides in software and hardware [14].
the relationship between secret key and cipher text. This is
the only non-linear component that determines the strength of Advance Encryption Standard (AES) uses fixed S-box
entire block encryption algorithm. Therefore, the main concern in that the S-box remains same despite changing the key.
in the construction of strong crypto-system is the design of a Thus static S-box opens ways for cipher text pair attacks;
secure s-box. and the key-dependent S-box escalates the complexity of the
Chaotic systems are favorable for use in cryptography as algorithm [15]. Although key-dependent S-boxes enhance
they behave differently on the basis of their initial parameters. the strength of the algorithm, the main problem is the

978-1-5386-8317-0/19/$31.00 ©2019 IEEE

efficient development of dynamic S-boxes. This problem has II. C UCKOO S EARCH A LGORITHM
been attempted in [16] [17]. The dynamic S-box increases Cuckoo Search Algorithm, originally invented by Yang and
the efficiency and security of the encryption algorithm as Deb [27], is a population based algorithm. Cuckoo bird does
compared to the static S-box [18]. not build his own nest for hatching eggs. She cunningly lays
her eggs in host bird nest. The other specie unknowingly
Chaos-based encryption system lacks in systematic design takes responsibility to hatch these eggs and raises the children.
to enhance crypto-graphic properties of S-box. Systematic Cuckoo bird inherently has the ability to disguise the color and
design is a prime need. Having this challenge, optimization voice. Due to this, the host bird can-not tell the cuckoo birds
techniques are used to get near-optimal S-box in terms from his own birds. So the host bird raises the cuckoo birds
of non-linearity. These techniques are accompanied with along with his own. Hence, the possibility of distinguishing
chaotic maps to produce the combined effect in the creation the foreign chicks reduces, and this increases the chances of
of S-box [19] [20] [21]. Combinatorial optimization finds the reproduction of the cuckoo birds. However, when the host
the optimal discrete solution which satisfies certain criteria. bird discovers the alien eggs, she takes one of the two actions.
Combinatorial problems have finite many solutions to a Either the host bird leaves the current nest and finds another
particular problem. Evolutionary methods find the best nest or she drops the alien eggs. This parasitic behavior of the
optimal solution. The application of these methods has been cuckoo birds makes this algorithm interesting and attractive.
investigated in single key cipher and double keys cipher The prominent features of this algorithm are enumerated as:-
[22]. S-box based on biological DNA is proposed [23]. In 1) Each cuckoo bird chooses a nest at random, and lays
all the previous attempts, optimization algorithms improve one egg at a time.
only one cryptographic property such as nonlinearity. This 2) Only the nest having fittest eggs (solutions) enters into
problem has been addressed in [24] where the authors applied the next population.
artificial bee colony algorithm for the creation of S-box. The 3) The host nests are fixed in number. The probability of
attractive feature of this design is considering nonlinearity and alien eggs discovery lies between 0 and 1. The native
differential uniformity as objective function simultaneously. bird has option either to drop the alien eggs from the
Population based meta-hueristic algorithms are widely used nest, or leave the nest and construct a new nest.
in the development of cryptosystem. These algorithms are Cuckoo Search Algorithm (CSA) has been proved very
very efficient in finding the s-box having good cryptographic useful in finding the optimal solution in many combi-natorial
properties. Wang [19] proposed a method based on chaos and problems such as Travelling Sales Person [28] and Scheduling
genetic algorithm for generating stronger 8x8 S-box. This problems [29]. By employing CSA, Ashok K. Bhateja [30]
method comprises two phases. In the first phase, initial S-box broke vigenere cipher. The main characteristic of this cipher
is generated by iterating the logistic map. In the second phase, is that it has a large key space. So to find the secret key using
the S-box thus generated is used as input to genetic algorithm conventional methods such as brute force seems infeasible.
in order to obtain secure 8x8 S-box. In [25] , author applied The author of CSA proved the superiority of CSA over its
Ant Colony Optimization for obtaining stronger S-box. contemporary population-based algorithms such as PSO and
Chaotic logistic map has been applied to produce the initial GA in terms of finding optimal solution. Therefore, this is our
population. So the Chaotic map complements the Ant Colony main source of inspiration to apply CSA to construct secure
Optimization in the creation of stronger 8x8 S-box. In [26], S-box.
S-box is designed with the help of firefly algorithm and
discrete chaotic maps. III. P ROPOSED S CHEME
The problem of finding an S-box with higher nonlinearity
An 8x8 S-box contains 256 distinct elements. The total is transformed into a combinatorial problem. Population based
permutation (search space) comes to 256! Although chaotic meta-heuristic techniques are the best methods for the solution
systems are very helpful in the design of secure S-box, of such problems. The proposed solution is composed of
these methods lack in directed search through such a huge Cuckoo Search Algorithm and Chaotic Maps. Cuckoo Search
search space. Therefore, a combination of chaotic theory and manipulates the initial population of solutions (S-boxes) to
Cuckoo Search Algorithm (CSA) has been applied in the find the better solution. Chaotic maps help create the initial
development of secure S-box. population of S-box. Chaotic maps generate nonlinear S-box
which provides a good starting point for the CSA Algorithm
to enhance the nonlinearity of the S-boxes.
The organization of the paper is as follows: Introduction
section briefly highlights the importance of an S-box, which A. Fitness Function
is followed by Related Work in section II. Section III briefly Fitness function is a function the value of which is
describes CSA; the proposed idea is explained in section IV; increased or decreased by the meta-heuristic searching
the comparison of results is summarized in section V followed algorithm. This function evaluates the fitness of a solution.
by the conclusion in section VI. In our proposed solution, the nonlinearity of S-box plays the
role of fitness function. The objective function is measured 256 distinct elements. In this way, the initial S-box is created.
using Walsh Transformation as defined in Eq. (1):- The S-box thus generated has good value of nonlinearity. It
provides a good starting point for the proposed scheme.
1
Nf = 2n−1 − maxw∈GF (2n ) |Sf (w)| (1) C. Generaing New Solution (S-Box)
2
X As 8x8 S-box contains 256 distinct integer values in a
Sf (w) = (−1)f (x)⊕x.w (2) specific order, new S-box can be generated by changing the
x∈GF (2n ) order of any two or more elements. The Inversion Operator
is applied over a selected segment of elements. A random
Where Sf (w) is the Walsh spectrum of f (x) and x.w is the
number i between [0,255] is selected, which represents
dot product of x and w.
the ith element in the sequence. Another random number
B. Initial S-box j between [1,15] is selected. The number j is added into
number i such as (i + j ) which represents the last element
Chaotic maps are used to provide initial populations. These
of the segment. If (i + j ) ≥ 256, (i − j ) is taken as the last
are inherently non-linear. This property of non-linearity is very
element of the segment. These segment values are swapped
important for substitution box. So the population of initial
to get the new solution (S-Box).
S-boxes is produced using the theory of chaotic maps. This
provides a good starting point for Cuckoo Search Algorithm
Generating S-Box based on Cuckoo Search Algorithm
(CSA) so that it can quickly find the S-box with greater value
Input:
of non-linearity. The initial S-box is generated by iterating
Fitness Function as defined in Eq. 1
chaotic logistic map and tent map defined in Eqs. (3)& (4).
Probability coefficient Pa
xi+1 = µxi (1 − xi ) (3) Output: Xi max
 Begin
(xi )/(b) 0 < xi ≤ 0;
x= (4) 1 Generate-initial-solution using chaotic maps
(1 − xi )/(1 − b) b < xi < 1.
for i=1 to n do
Where µ and b are the control parameters. The output of these Generate-initial-population
chaotic maps is very sensitive to these parameters. The values end for
of these parameters are taken as µ=3.99999 and b=0.499999 2 Repeat
aiming to induce random like behavior to a greater extent. a ← Get-cuckoo-randomly(x)
Generate-new-solution(a) for m times
Generating initial S-box based on chaotic maps b ← Get-best-cuckoo(x)
Input: Generate-new-solution(b) for m times
Define S as a sequence 3 If (rand . pa) then
Output: c ← Get-worst-cuckoo(x)
Sequence S Generate-new-solution(c) for m times
Begin: Remove c from population x
End if
While (S.length ≤ 256)T hen 4 Select n best cuckoos
f ori = 0to10 5 Until stop condition true
xi ← Logistic Map defined in Eq. 3 6 Output the best cuckoo Xi
end for End
for i=0 to 10
yi ← Tent Map defined in Eq. 4 The proposed algorithm is elaborated as follows:
end for 1 Generate the initial S-box using chaotic systems defined
X = yj × 256 in Algorithm 1. From the initial S-box, a population of
If X not exists in S S-boxes n is generated as described in section IV.
S←X 2 Select a random S-box from the existing population.
end while Create new S-boxes m from the selected S-box. Now,
Output sequence S select the S-box with highest value of nonlinearity and
create new S-boxes m. Total population becomes n+2m.
End 3 Generate a random value between 0 and1. If the value is
Algorithm 1 generates the initial S-box. Two chaotic maps less than the probability Pa=0.25, generate new S-boxes
are applied to create a sequence of 256 distinct elements. The from the S-box with least value of nonlinearity.
output of the logistic map defined in Eq. 3 is fed into the tent 4 Sort the S-boxes based on nonlinearity. Remove S-boxes
map defined in Eq.4. Initially, an empty sequence S is defined, with less value of nonlinearity and keep the population
which is populated using logistic map and tent map until it has of S-boxes n with greater value of nonlinearity.
 5 Iterate over steps from 2 to 4 until the stop condition is presented between the proposed S-box and the S-boxes
becomes true. obtained using population based optimization techniques. It
6 Output the S-box with the best value of nonlinearity. is evident that the proposed scheme generates the S-box with
Where Pa =0.25 and n=20. It produces n best solutions. higher value of nonlinearity 109.25. The comparative analysis
The solution with the higher fitness value is selected as the in Table III shows that the proposed S-box is better than the
optimal solution. The S-box obtained by the proposed scheme recent S-boxes in terms of minimum, maximum and average
is shown in Table I. value of nonlinearity.

IV. R ESULTS AND D ISCUSSIONS C. Strict Avalanche Criteria

In this section, the strength of the obtained S-box is evalu- The idea of strict avalanche criteria (SAC) is initially
ation against standard performance criteria. presented by Webster and Tavares [33]. This property is
important when some complex mapping f of n bits on to one
A. Bijectivity
bit is needed. Strict Avalanche Criteria ensures that changing
For an n × n S-box, the Boolean Function fi is bijective if a single input bit spreads the change over half of the output
it satisfies the following equation bits. The SAC dependency matrix of our S-box is shown
wt(a1 f1 ⊕ a2 f2 ⊕ ...... ⊕ an fn ) = 2n−1 (5) in Table IV. The average value of SAC in the dependency
matrix is 0.507568 which is very close to the ideal value
where ai ∈ {0, 1} and wt(.) is the hamming weight. This of 0.50. The mean SAC of the proposed S-box is compared
property ensures that the s-box has one-to-one and onto with that of S-boxes found in recent literature as presented in
mapping from input vectors to output vectors. It means that Table III. The mean SAC of proposed S-box is 0.5076 with
an 8x8 S-box satisfies bijectivity if it has unique values in a negligible deviation of 0.0169 from the ideal SAC value of
the range [0, 255]. If the S-box contains the distinct values 0.50.
between [0, 255], it means that the input values have a unique
corresponding output values. The obtained S-box presented in
Table-1 has the distinct element in the specified range. So, it D. Bit Independence Criteria
satisfies bijectivity. Webster and Tavares invented Bit Independence Criteria
B. Nonlinearity (BIC). It is one of the important properties of S-box. This
property states that the degree of independence between any
Nonlinearity of a Boolean function f (x) : F2n → F2 is de-
two bits of cipher text must be large when a single bit in the
fined as the minimum distance from the set of affine functions.
plane text is complemented [33]. The proposed S-box has the
It ensures that output vectors are not linearly mapped from
minimum value of BIC 92 and the maximum value of 108
input vectors. It is measured using Walsh transformation. Non-
as depicted in Table V. The average value of BIC is 102.929
linearity provides a kind of security shield against statistical
which contribute to the strength of the nonlinear component.
security attack. To compute nonlinearity of an 8x8 S-box, all
The comparison of mean BIC of our S-box with that of
the 256 elements are arranged in a column. This column of
recent S-boxes is depicted in the Table III. The average value
256 integer values is converted into binary numbers. Now this
of BIC of our S-box is 102.929, which is comparable with the
vertical column consists of 8 binary columns. Each binary
recently developed S-boxes.
column represents a binary truth table. Each binary truth table
has a nonlinear value. E. Linear Approximation Probability
The linear approximation of the S-box gives the probability
bias of an event. Minor change in input bits pattern generates
imbalance in the output bits. The linear approximation value
of the S-box is 0.1406. Table III presents the comparative
analysis. The value of linear probability shows that the S-box
offers resistance against linear attacks.

F. Differential Approximation Probability

The S-box resists against differential attacks. This resis-
tance is quantified using differential approximation probability
value. The differential uniformity is achieved in the differential
input mapping to differential output mapping. Differential
Fig. 1. Comparison of nonlinearity with optimization schemes uniformity is measured using differential approximation prob-
ability. The result comparison is presented in Table III. The
The nonlinearity values of the S-box are 108, 110, 110, 108, maximum differential probability of the proposed S-box has
110, 110, 108 and 110 making an average nonlinearity value the value of 12/256= 0.0468. The new S-box shows resistance
of 109.25. In Table II & Figure 1, the nonlinearity comparison against differential attacks.
 TABLE I
S- BOX OBTAINED THROUGH PROPOSED SCHEME

8 72 114 156 153 41 193 82 78 15 112 141 203 6 159 161

74 3 50 215 44 157 12 34 99 190 133 208 183 224 27 212
249 220 4 178 13 145 251 195 139 54 68 7 69 148 240 172
206 163 125 20 175 51 83 124 117 239 180 47 32 210 149 109
73 5 33 197 228 25 221 105 173 103 94 88 92 48 42 90
23 1 244 204 253 179 242 181 230 66 246 86 184 61 122 129
76 55 182 108 174 96 77 87 213 107 120 248 130 60 144 143
113 137 234 10 162 11 245 151 222 91 43 0 241 36 199 40
225 136 84 127 207 79 187 59 85 132 167 104 176 229 14 147
138 160 102 168 219 29 116 150 223 39 152 217 16 97 81 200
211 100 135 95 164 191 189 2 121 106 62 205 142 254 177 235
89 155 93 46 146 226 26 209 30 80 227 115 196 154 185 158
202 71 70 194 98 63 201 218 110 67 214 169 64 243 17 216
128 58 192 19 236 31 232 22 247 49 255 131 18 35 165 24
134 186 21 65 231 28 101 37 170 171 188 38 45 166 140 233
118 56 237 198 238 123 111 252 119 250 75 52 57 126 9 53

TABLE II
C OMPARISON OF NONLINEARITY WITH OPTIMIZATION SCHEMES

S-Boxes 0 1 2 3 4 5 6 7 Min Max Mean

Proposed 110 108 110 110 108 110 110 108 108 110 109.25
GA [19] 108 108 108 108 108 108 108 108 108 108 108
ACO [20] 108 106 106 106 106 110 106 108 106 110 107
Firefly [26] 108 108 108 108 108 108 106 106 106 108 107.5

TABLE III
P ERFORMANCE COMPARISON OF PROPOSED S- BOX WITH RECENT SCHEMES

S-Boxes Min NL Max NL Avg NL Avg SAC BIC-NL DP LP

Proposed 108 110 109.25 0.5075 102.929 12 0.1406
GA [19] 108 108 108 0.5068 103.36 10 0.1406
ACO [20] 106 110 107 0.5015 104.21 10 0.1484
Firefly [26] 106 108 107.5 0.4943 104.35 10 0.1250
[3] 96 106 103.2 0.5151 103.07 44 0.1562
[8] 104 108 105.8 0.4976 104.50 10 0.1172
[13] 106 108 106.75 0.5034 103.78 10 0.1328
[21] 104 108 106.5 0.4991 104.57 10 0.1172
[31] 103 109 105.12 0.5061 103.67 10 0.1563
[32] 102 108 104.7 0.5056 104.28 12 0.1172

TABLE IV
D EPENDENCY MATRIX FOR STRICT AVALANCHE CRITERIA

0.531250 0.484375 0.5 0.546875 0.546875 0.484375 0.546875 0.546875

0.531250 0.515625 0.531250 0.484375 0.562500 0.578125 0.437500 0.515625
0.515625 0.531250 0.562500 0.515625 0.484375 0.515625 0.546875 0.468750
0.515625 0.546875 0.515625 0.531250 0.468750 0.468750 0.484375 0.468750
0.484375 0.453125 0.484375 0.5 0.515625 0.5 0.531250 0.468750
0.5 0.515625 0.531250 0.546875 0.515625 0.515625 0.421875 0.5
0.546875 0.562500 0.484375 0.468750 0.546875 0.531250 0.468750 0.515625
0.515625 0.515625 0.437500 0.515625 0.453125 0.484375 0.484375 0.484375
 TABLE V
BIC CORRESPONDING TO N ONLINEARITY

0 102 100 102 102 104 98 106

102 0 106 102 106 102 102 102
100 106 0 104 102 106 108 102
102 102 104 0 104 104 104 104
102 106 102 104 0 92 100 102
104 102 106 104 92 0 106 104
98 102 108 104 100 106 0 106
106 102 102 104 102 104 106 0

V. C ONCLUSION [10] Attaullah, S. S. Jamal, and T. Shah, “A Novel Algebraic Technique

for the Construction of Strong Substitution Box,” Wireless Personal
Substitution- box (S-box) induces the confusion in the sym- Communications, vol. 99, no. 1, pp. 213–226, 2018. [Online]. Available:
metric cipher. In this Paper, the design of S-box is transformed https://doi.org/10.1007/s11277-017-5054-x
[11] Y. Zhang and D. Xiao, “Cryptanalysis of S-box-only chaotic image
into combinatorial problem. Cuckoo Search Algorithm (CSA) ciphers against chosen plaintext attack,” Nonlinear Dynamics, vol. 72,
along with chaotic maps has been applied to get optimal S-box. no. 4, pp. 751–756, 2013.
The S-box is compared with the S-boxes obtained using GA, [12] S. Farwa, N. Muhammad, T. Shah, and S. Ahmad, “A Novel Image
Encryption Based on Algebraic S-box and Arnold Transform,” 3D
ACO and FA. The proposed S-box shows good results. The Research, vol. 8, no. 3, 2017.
proposed scheme enhances the nonlinearity, which is 109.25. [13] D. Lambić, “A novel method of S-box design based on discrete chaotic
The remaining properties such as BIC, SAC, LP and DP are map,” Nonlinear Dynamics, vol. 87, no. 4, pp. 2407–2413, 2017.
not considered in the proposed scheme as we applied only [14] F. A. Khan, J. Ahmed, J. S. Khan, J. Ahmad, M. A. Khan, and S. O.
Hwang, “A new technique for designing 8 8 substitution box for image
single objective version of CSA. It can further be enhanced encryption applications,” in 2017 9th Computer Science and Electronic
to maximize more than one cryptographic properties. The Engineering Conference, CEEC 2017 - Proceedings, 2017, pp. 7–12.
proposed technique synthesizes the combined effect of chaotic [15] G. Manjula and H. S. Mohan, “Constructing key dependent dynamic
S-Box for AES block cipher system,” in Proceedings of the 2016 2nd
theory and CSA in the creation of cryptographic secure 8x8 S- International Conference on Applied and Theoretical Computing and
box. The resulting S-box has encouraging cryptographic prop- Communication Technology, iCATccT 2016, no. 10, 2017, pp. 613–617.
erties which contributes to the security of block cipher. The [16] T. Ao, J. Rao, K. Dai, and X. Zou, “Construction of high quality key-
dependent S-boxes,” IAENG International Journal of Computer Science,
comparative analysis highlights the strength of the proposed vol. 44, no. 3, pp. 337–344, 2017.
S-box against linear and differential attacks. [17] B. K. Maram and J. M. Gnanasekar, “Evaluation of Key Dependent
S-Box Based Data Security Algorithm using Hamming Distance and
R EFERENCES Balanced Output,” Journal, vol. 5, no. 1, pp. 67–75, 2016.
[18] S. Katiyar, N. Jeyanthi, and S.-B. O. X. P. Roperties, “Pure Dynamic
R EFERENCES S-box Construction,” International Journal of Computers S. Katiyar, N.
Jeyanthi, vol. 1, no. 1, pp. 42–46, 2016.
[1] C. Paar and J. Pelzl, Understanding Cryptography, 2010. [Online]. [19] Y. Wang, K. W. Wong, C. Li, and Y. Li, “A novel method
Available: http://link.springer.com/10.1007/978-3-642-04101-3 to design S-box based on chaotic map and genetic algorithm,”
[2] C. Shannon, “Communication theory of secrecy system,” Bell System Physics Letters, Section A: General, Atomic and Solid State
Technical Journal, vol. 28, no. 4, pp. 656–715, 1949. Physics, vol. 376, no. 6-7, pp. 827–833, 2012. [Online]. Available:
[3] M. Khan and Z. Asghar, “A novel construction of substitution box for http://dx.doi.org/10.1016/j.physleta.2012.01.009
image encryption applications with Gingerbreadman chaotic map and [20] M. Ahmad, D. Bhatia, and Y. Hassan, “A Novel Ant Colony
S8permutation,” Neural Computing and Applications, vol. 29, no. 4, pp. Optimization Based Scheme for Substitution Box Design,” Procedia
993–999, 2018. Computer Science, vol. 57, pp. 572–580, 2015. [Online]. Available:
[4] H. M. Heys, “A tutorial on linear and differential cryptanalysis,” http://dx.doi.org/10.1016/j.procs.2015.07.394
Cryptologia, vol. 26, no. 3, pp. 189–221, 2002. [21] T. Farah, R. Rhouma, and S. Belghith, “A novel method for designing S-
[5] M. A. Khan, A. Ali, V. Jeoti, and S. Manzoor, “A Chaos- box based on chaotic map and TeachingLearning-Based Optimization,”
Based Substitution Box (S-Box) Design with Improved Differential Nonlinear Dynamics, vol. 88, no. 2, pp. 1059–1074, 2017.
Approximation Probability (DP),” Iranian Journal of Science and [22] T. Shirabe, “Combinatorial Optimization in Cryptography,” vol. 32, pp.
Technology - Transactions of Electrical Engineering, vol. 42, no. 2, 1324–1330, 2005.
pp. 219–238, 2018. [Online]. Available: https://doi.org/10.1007/s40998- [23] A. H. S. Al Wattar, R. Mahmod, Z. A. Zukarnain, and
018-0061-9 N. I. Udzir, “Generating a new S-Box inspired by biological
[6] F. A. Khan, J. Ahmed, J. S. Khan, J. Ahmad, and M. A. Khan, “A novel DNA,” International Journal of Computer Science and Application,
substitution box for encryption based on Lorenz equations,” in 2017 vol. 4, no. 1, pp. 32–42, 2015. [Online]. Available: http://dpi-
International Conference on Circuits, System and Simulation, ICCSS journals.com/index.php/IJCSA/article/view/1260/1118
2017, 2017, pp. 32–36. [24] Y. Tian and Z. Lu, “S-box: Six-dimensional compound hyperchaotic map
[7] J. Peng, A. A. El-Latif, A. Belazi, and Z. Kotulski, “Efficient chaotic and artificial bee colony algorithm,” Journal of Systems Engineering and
nonlinear component for secure cryptosystems,” in International Confer- Electronics, vol. 27, no. 1, pp. 232–241, 2016.
ence on Ubiquitous and Future Networks, ICUFN, 2017, pp. 989–993. [25] M. Ahmad, D. Bhatia, and Y. Hassan, “A Novel Ant Colony
[8] G. Liu, W. Yang, W. Liu, and Y. Dai, “Designing S-boxes based on 3-D Optimization Based Scheme for Substitution Box Design,” Procedia
four-wing autonomous chaotic system,” Nonlinear Dynamics, vol. 82, Computer Science, vol. 57, pp. 572–580, 2015. [Online]. Available:
no. 4, pp. 1867–1877, 2015. http://dx.doi.org/10.1016/j.procs.2015.07.394
[9] H. Liu, A. Kadir, and P. Gong, “A fast color image encryption scheme [26] H. Ahmed, M. Zolkipli, and M. Ahmad, “A novel efficient substitution-
using one-time S-Boxes based on complex chaotic system and random box design based on firefly algorithm and discrete chaotic map,”
noise,” Optics Communications, vol. 338, pp. 340–347, 2015. [Online]. Neural Computing and Applications, Springer, vol. 3, 2018. [Online].
Available: http://dx.doi.org/10.1016/j.optcom.2014.10.021 Available: https://doi.org/10.1007/s00521-018-3557-3
[27] X.-S. Yang and S. Deb, “Cuckoo Search via Levy Flights,” World
Congress on Nature {&} Biologically Inspired Computing, pp.
210–214, 2009. [Online]. Available: http://arxiv.org/abs/1003.1594
[28] A. Ouaarab, B. Ahiod, and X. S. Yang, “Discrete cuckoo search
algorithm for the travelling salesman problem,” Neural Computing and
Applications, vol. 24, no. 7-8, pp. 1659–1669, 2014.
[29] A. Ouaarab, B. Ahiod, X.-S. Yang, and M. Abbad,
“Discrete Cuckoo Search algorithm for job shop scheduling
problem,” in IEEE International Symposium on Intelligent
Control (ISIC), 2014, pp. 1872–1876. [Online]. Available:
http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=6967636
[30] A. K. Bhateja, A. Bhateja, S. Chaudhury, and P. K. Saxena,
“Cryptanalysis of Vigenere Cipher using Cuckoo Search,”
Applied Soft Computing Journal, 2014. [Online]. Available:
http://dx.doi.org/10.1016/j.asoc.2014.10.004
[31] F. Özkaynak and S. Yavuz, “Designing chaotic S-boxes based on time-
delay chaotic system,” Nonlinear Dynamics, vol. 74, no. 3, pp. 551–557,
2013.
[32] I. Hussain, T. Shah, and M. A. Gondal, “A novel approach for designing
substitution-boxes based on nonlinear chaotic algorithm,” Nonlinear
Dynamics, vol. 70, no. 3, pp. 1791–1794, 2012.
[33] W. Tavares, “On the design of S-boxes. In: Williams, H.C. (Ed.),
CRYPTO’85. LNCS 218, 523534.” 1986.