Evolutionary Genetic Algorithm for Encryption

Wafa' Slaibi Alsharafat

Prince Hussein Bin Abdullah Faculty of Information Technology, Information Systems Department
Al al-Bayt University
Mafraq, Jordan
[email protected]

Abstract— Encryption is a fundamental mechanism for

protecting information since the significant of security is
increasing nowadays. Different evolutionary methods have been
adapted to satisfy high level of security. Genetic algorithm is an
exploration and evolutionary algorithm which based on natural
selection which optimizing problem solution and to be away from
producing one ciphertext for the same plaintext. Therefore, the
attempts of attack, as ciphertext attack, will be reduced if not
eliminated. In this paper, GA as evolutionary algorithm will be
employed in a symmetric encryption and decryption where the
user's message which represents plaintext and user secret data, Fig. 2. ASymmetric Encryption/decryption
key, will be transferred into ciphertext by applying crossover and
mutation beside substitution function. Symmetric Encryption is recognized as a single-key
encryption where one-key is used in encryption and decryption.
In opposite, asymmetric encryption algorithms use multi keys
Keywords— GA; Encryption; Decryption during encryption and decryption. In addition, encryption
algorithms can implement substitution or transposition [2].
Substitution substitutes each plaintext symbol with another
I. INTRODUCTION
symbol to form ciphertext. Transposition technique described
Recently, Security is acritical mission which concerns as repositioning plaintext to produce ciphertext.
about keeping data and transactions away from any attempts of
attacks especially when number of possible attack and methods An evolutionary method as Genetic Algorithm (GA) [5] is
of attacks increased rapidly. Several security solutions are implemented to raise the level robustness and variation
arisen by adopting different mechanisms as Encryption, between plaintext and ciphertext. To satisfy the randomness
Intrusion Detection Systems (IDS), firewalls [1] to satisfy main and robustness, a set of operation will be executed which
security requirements; Availability, Confidentiality and represented by selecting, crossover and mutation to keep the
Integrity [2]. Encryption is concerns about converting and diversity and mismatch between plaintext and ciphertext.
encoding information or messages into a format that readable
and accessible by authorized users [2]. In an encryption, the II. RELATED WORK
text referred as a plaintext which will be input for encryption Different researches have been conducted to put into effect
algorithm combined with secret information called a key to different security mechanisms to keep high level of security.
generate ciphertext that can only be read if decrypted [3]. For that purpose, new encryption algorithm have been
Encryption algorithms divided into; symmetric encryption developed and enhanced by using artificial intelligence.
algorithm and asymmetric encryption algorithm as provided in
Fig. 1, 2 respectively. Researchers in [6] have proposed algorithm by using
several keys for the encryption and decryption process and this
mission will be achieved by using fuzzy logic. According to
user needs, he will specify number of key to be used in
encryption algorithm where short key size specifies low
security level while long key size specifies high security level.
Tabu-genetic algorithm for evolving S-boxes in symmetric
algorithms like DES was presented in [7]. Researchers
implemented a niche technique to keep diversity between
populations and evade premature convergence. Niche
Fig. 1. Symmetric Encryption/decryption technique is used to find several optimal solutions, but also to
reduce execution time.

978-1-4799-3975-6/14/$31.00 ©2014 IEEE

Authorized licensed use limited to: Banasthali Vidyapith. Downloaded on February 12,2025 at 13:12:32 UTC from IEEE Xplore. Restrictions apply.
 2014 IEEE International Conference on Computational Intelligence and Computing Research

Su Su Maung and others in [8] have proposed a fast In next section, provide an implementation on GA in
encryption by adapting chaotic maps. Logistic map will be encryption process.
used to produce Two dimensional standard map; 8*8 S-box.
Then, a sequence of random bytes is generated by using 2D IV. GENETIC ALGORITHM INTO IMPLEMENTATION
chaotic cat map to index the entries of the S-box. Finally,
ciphertext will be produced from implementing S-box and GA used to enforce expanding in population size and
XOR on plaintext. diversity through crossover and mutation. The main steps for
encryption which starts from plaintext to ciphertext are
Artificial Neural Network (ANN) in [9] has been used to described as follow:
generate ciphertext by simulating natural noise. For
implementing ANN, several decisions must be taken which A. Plaintext Presentation:
concern about number of ANN layers, and weight for each
input from plaintext. In addition, Researchers in [4] have In this step, suitable representation method will influence
applied artificial neural network. They focused on receiving the effectiveness of GA in encryption.
plaintext as input processed through multilayer neural network. There are different ways for encoding and represent
While key is represented as a weight and neural function as plaintext which consists from user message and secret key.
internal components for multilayer neural network Here, binary coding will be implemented because ASCII code
will be implemented on plaintext. Furthermore, crossover and
mutation work efficiently on binary code and we must take into
So, there are different approaches have been used in account that the selected method depends on the problem itself.
encryption and generation of secure key [10, 13]. Some of
these approaches are listed as follows: Message text will be formatted using two dimension a
matrix =n*n, here matrix=3*3. To increase complexity, shift
• Artificial Neural Network can be performed on plaintext by shift bits to left or right
several bits.
• Bayesian Network
Another process will be applied which concerns about
• Artificial Immune System adding secret information to message called key. Key must be
• K- nearest neighbor algorithm presented using matrix =n*n as size of plaintext to enable
additive operation between text and key, for example user
• Support Vector Machine message=NETWORK which mapped into matrix=3*3 which
later converted into ASCII matrix and shifted 2 bits, as shown
• Decision Tree
in Fig.3 and Fig.4 respectively. Same operation implemented
• Fuzzy Logic on secret key e.g. SECURITY.
• Machine Learning

III. GENETIC ALGORITHM

Genetic Algorithm (GA) is an adaptive exploratory search
algorithm premised on the evolutionary ideas of natural
selection and genetic. There are several causes for
implementing GA which starts with expanding the size of the
population and guarantee randomness by adapting crossover
and mutation operations. Fig. 3. Plaintext matrix and ASCII code.
To encrypt sender message, GA contained the following
steps [11]:
1. Initiation which concerns about specifying sender
message and secret key.
2. State stopping criterion,
3. Using genetic operators
a. Crossover Fig. 4. Message afetr shifting

b. Mutation Finally, substitution and ASCII are implemented

respectively to prepare final version of plaintext into bits
4. Creating ciphertext. format then divide text into two equivalent pairs, or more, as
shown in Fig.5.

Authorized licensed use limited to: Banasthali Vidyapith. Downloaded on February 12,2025 at 13:12:32 UTC from IEEE Xplore. Restrictions apply.
 2014 IEEE International Conference on Computational Intelligence and Computing Research

Fig. 5. Final version of plaintext. Fig. 7. Mutation representation

B. Applying crossover and mutation After mutation, final ciphertext is produced and converted
into readable format by using hexadecimal or octal or decimal
system.
Crossover is one of main characteristics that distinguished
GA from other revolution techniques. Crossover is a process of In decryption, the steps for encryption process are reversed
and start with ciphertext instead of plaintext.
exchange of information between two texts to reproduce new
text, individual, inherent their parent's behavior. For comparison, the expected complexity for the proposed
All crossover methods depend on encoding method and method by using GA in encryption will be O (n ). 2

problem type. In crossover, two parents are taken and a new

text, child, is generated by taking some attributes from both
parents. In GAs a crossover can be one of following types [5, V. CONCLUSION
12] Evolutionary methods as GA have shown during encryption
• Single Point Crossover: works by selecting efficiency and effectiveness through. Different improvement
random position k (k Щ {1, 2,.., l-1}, l= is text length) can be implemented on GA operation as crossover and
to produce new text by intersect parents text at position mutation to reduce possibilities to retrieve plaintext from
k. ciphertext. Furthermore, a set of enhancement will be
introduced by modifying crossover and mutation by applying
• Two Point Crossover: works by selecting two fuzzy logic to guarantee a high level of diversity between
crossover points; k1, K2 are selected text will be plaintext and ciphertext.
intersected at position k1 and K2.
• Multi point Crossover: instead of selecting one
intersection point or 2 point we select more than 2
intersection point. REFERENCES
In this paper, multiple point crossovers have been
performed to produce the result in Fig. 6. [1] W.Alsharafat, “ Classifier System in Cloud Environment to Detect
Denial of Service Attack,” nternational Journal of Computer
Applications, Vol. 85, No. 14, 2014.
[2] W.Stallings, “Cryptography and Network Security,” 4th Edition, Pearson
publisher, 2006.
[3] O.Goldreich, “Foundations of Cryptography Basic Applications,”
Cambridge university press, vol. 2, 2004
[4] S. Lonkar, N.Charniya, “Neural Network based Cryptography, ”
International Journal of Application or Innovation in Engineering &
Management (IJAIEM), 2014.
[5] H. Bhasin and S. Bhatia, “Application of Genetic Algorithms in
Fig. 6. Crossover implementation Machine learning, ” IJCSIT, Vol. 2 (5), 2011.
[6] R. Madanayake , R. Peiris, N. Ranaweera, and Th. Jayathilake,
“Advanced Encryption Algorithm Using Fuzzy Logic,” 2012
Mutation is applied to keep diversity between current International Conference on Information and Computer Networks
(ICICN 2012). IPCSIT vol. 27 (2012) © (2012) IACSIT Press,
generation and next generations. In mutation process plaintext Singapore.
bits are changed randomly to introduce certain diversity in the [7] Y.Xiangdong, “S Box Construction and Result Analysis Based on
ciphertext. One of strength point for mutation is produce new Optimal Tabu-genetic Algorithm,” 2nd International Conference on
text different from existing for more exploration to discover Education Technology and Computer (ICETC), Shanghai , vol. 3, pp.
unknown situations in search space. 539-542, 2010.
[8] S.Su Maung, and M. Myint Sein, “A Fast Encryption Scheme Based on
Here, mutation takes a place after performing crossover. Chaotic Maps,” GMSARN International Conference on Sustainable
One of mutation techniques concerns about flip-flop text by Development: Issues and Prospects for the GMS, 2008.
converting randomly 0's to 1's or 1's to 0's as provided in Fig.7. [9] J.Blackledge, S.Bezobrazav, P.Tobin, and F. Zamora, “ Cryptography
Uning Evolutionary Computing,” ISSC 2013, LYIT Letterkenny, June
20–21.

Authorized licensed use limited to: Banasthali Vidyapith. Downloaded on February 12,2025 at 13:12:32 UTC from IEEE Xplore. Restrictions apply.
 2014 IEEE International Conference on Computational Intelligence and Computing Research

[10] A.Dadhich, S. Kumar Yadav, “ Evolutionary Algorithms, Fuzzy Logic

and Artificial Immune Systems applied to Cryptography and
Cryptanalysis: State-of-the-art review,” International Journal of
Advanced Research in Computer Engineering & Technology
(IJARCET), Volume 3 Issue 6, June 2014.

[11] M. Mitchell, “An Introduction to Genetic Algorithms,” The MIT

Press, Cambridge, USA, 1999.

[12] D.Pisinger, “Linear Time Algorithms for Knapsack Problems with

Bounded Weights, ” Journal of Algorithms, Volume 33, Number 1,
October 1999, pp. 1–14.

[13] E. C ,Laskari, G.C., Meletiou,Y.C., Stamation, and M.N.,Vrahatis,

“Evolutionary Computation based Cryptanalysis: A first study.Nonlinear
Analysis, ” vol. 63, no.(5- 7), pp. 823-830, 2005.

Authorized licensed use limited to: Banasthali Vidyapith. Downloaded on February 12,2025 at 13:12:32 UTC from IEEE Xplore. Restrictions apply.