Professional Documents
Culture Documents
The Golden'' Matrices and A New Kind of Cryptography
The Golden'' Matrices and A New Kind of Cryptography
www.elsevier.com/locate/chaos
Abstract
We consider a new class of square matrices called the ‘‘golden’’ matrices. They are a generalization of the classical
Fibonacci Q-matrix for continuous domain. The ‘‘golden’’ matrices can be used for creation of a new kind of cryptog-
raphy called the ‘‘golden’’ cryptography. The method is very fast and simple for technical realization and can be used
for cryptographic protection of digital signals (telecommunication and measurement systems).
2006 Elsevier Ltd. All rights reserved.
1. Introduction
In the last decades the theory of Fibonacci numbers [1–3] was reinforced and invigorated by new scientific results
and applications [4–31]. Let us consider only two of them.
The following property of the nth power of the Q-matrix was proved in [2]:
F nþ1 F n
Qn ¼ ; ð2Þ
F n F n1
where n = 0, ±1, ±2, ±3, . . ., Fn1, Fn, Fn+1 are Fibonacci numbers given by the following recurrence relation:
F nþ1 ¼ F n þ F n1 ð3Þ
1
Professor, Academician of the Ukrainian Academy of Engineering Sciences.
E-mail address: goldenmuseum@rogers.com
URL: www.goldenmuseum.com
0960-0779/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chaos.2006.03.069
A.P. Stakhov / Chaos, Solitons and Fractals 32 (2007) 1138–1146 1139
Alexey Stakhov, Ivan Tkachenko and Boris Rozin developed recently a theory of the hyperbolic Fibonacci and Lucas
functions [8,14,15].
Let us consider so-called symmetrical hyperbolic Fibonacci functions introduced in [15].
Symmetrical hyperbolic Fibonacci sine:
sx sx
sFsðxÞ ¼ pffiffiffi : ð6Þ
5
Symmetrical hyperbolic Fibonacci cosine:
sx þ sx
cFsðxÞ ¼ pffiffiffi ; ð7Þ
5
pffiffi
where s ¼ 1þ2 5 (the Golden Proportion).
Note that the symmetrical hyperbolic Fibonacci functions are connected to the Fibonacci numbers (3) and (4) by the
following correlations:
sFsðnÞ; for n ¼ 2k;
Fn ¼ ð8Þ
cFsðnÞ; for n ¼ 2k þ 1:
It was proved in [15] that the following identities connect the symmetrical hyperbolic Fibonacci functions:
½sFsðxÞ2 cFsðx þ 1ÞcFsðx 1Þ ¼ 1; ð9Þ
½cFsðxÞ2 sFsðx þ 1ÞsFsðx 1Þ ¼ 1: ð10Þ
Note that the identities (9) and (10) are a generalization of the ‘‘Cassini formula’’ (5) for continuous domain.
The main purpose of the present paper is to develop a theory of the ‘‘golden’’ matrices that are a generalization of
the matrix (2) for continuous domain. It is also considered to be a new kind of cryptography, which we call the
‘‘golden’’ cryptography, because it is based on the ‘‘golden’’ matrices.
Table 1
The explicit forms of the matrices Qn
n 0 1 2 3 4 5 6 7
n
Q 1 0 1 1 2 1 3 2 5 3 8 5 13 8 21 13
0 1 1 0 1 1 2 1 3 2 5 3 8 5 13 8
1 0 1 1 1 1 1 2 2 3 3 5 5 8 8 13
Qn
0 1 1 0 1 2 2 3 3 5 5 8 8 13 13 21
A representation of the matrices Qn (n = 0, ±1, ±2, ±3, . . .) in explicit form based on the recurrence relations (12) and
(13) are given in Table 1.
Note that Table 1 gives the ‘‘direct’’ matrices Qn and their ‘‘inverse’’ matrices Qn in explicit form. Comparing the
‘‘direct’’ and ‘‘inverse’’ Fibonacci matrices Qn and Qn given in Table 1 it is easy to see that there is a very simple
method to get the ‘‘inverse’’ matrix Qn from its ‘‘direct’’ matrix Qn. In fact, if the power of the ‘‘direct’’ matrix Qn
given with (2) is odd number (n = 2k) then for getting of its ‘‘inverse’’ matrix Qn it is necessary to rearrange in matrix
(2) its diagonal elements Fn+1 and Fn1 and to take its diagonal elements Fn with an opposite sign. It means that for the
case n = 2k the ‘‘inverse’’ matrix Qn has the following form:
F 2k1 F 2k
Q2k ¼ : ð15Þ
F 2k F 2kþ1
For the case n = 2k + 1 for obtaining the ‘‘inverse’’ matrix Qn from the ‘‘direct’’ matrix Qn it is necessary to rear-
range in (2) the diagonal elements Fn+1 and Fn1 and to take them with an opposite sign, so that one finds:
!
F 2k F 2kþ1
Qð2kþ1Þ ¼ : ð16Þ
F 2kþ1 F 2kþ2
Let us represent now the matrix (2) in the form of two matrices that are given for the odd (n = 2k) and even
(n = 2k + 1) values of n:
!
2k
F 2kþ1 F 2k
Q ¼ ; ð17Þ
F 2k F 2k1
!
2kþ1
F 2kþ2 F 2kþ1
Q ¼ : ð18Þ
F 2kþ1 F 2k
If we use the correlation (8), we can write the matrices (17) and (18) in the terms of the symmetrical hyperbolic Fibo-
nacci functions (6) and (7):
cFsð2k þ 1Þ sFsð2kÞ
Q2k ¼ ; ð19Þ
sFsð2kÞ cFsð2k 1Þ
sFsð2k þ 2Þ cFsð2k þ 1Þ
Q2kþ1 ¼ ; ð20Þ
cFsð2k þ 1Þ sFsð2kÞ
It is clear that the matrices (21), (22) are a generalization of the Q-matrix (2) for continuous domain. They have a
number of the unique mathematical properties. For example, for x ¼ 14 the matrix (14) takes the following form:
!
1 pffiffiffiffi cFsð32Þ sFsð12Þ
Q2 ¼ Q ¼ : ð23Þ
sFsð12Þ cFsð 12Þ
It is impossible to imagine a meaning for the ‘‘square root of the Q-matrix’’, but such ‘‘Fibonacci fantasy’’ follows from
(23) amazing on it may be.
Let us represent the inverse matrices (15) and (16) in the terms of the symmetrical hyperbolic Fibonacci functions (6)
and (7):
!
2k
cFsð2k 1Þ sFsð2kÞ
Q ¼ ; ð24Þ
sFsð2kÞ cFsð2k þ 1Þ
!
2kþ1
sFsð2kÞ cFsð2k þ 1Þ
Q ¼ ; ð25Þ
cFsð2k þ 1Þ sFsð2k þ 2Þ
Let us calculate now the determinants of the matrices (21) and (22):
Det Q2x ¼ cFsð2x þ 1Þ cFsð2x 1Þ ½sFsð2xÞ2 ; ð37Þ
2xþ1 2
Det Q ¼ sFsð2x þ 2Þ sFsð2xÞ ½cFsð2x þ 1Þ : ð38Þ
Let us compare the expression (37) to the identity (8). Because the identity (8) is true for any values of the variable x, in
particular, for the value 2x, it follows from this consideration the following identity:
Det Q2x ¼ 1: ð39Þ
By analogy, we can write:
Det Q2xþ1 ¼ 1: ð40Þ
If we now return back to the ‘‘Cassini formula’’ (5), we can conclude that the unusual identities (39) and (40) are a
generalization of the ‘‘Cassini formula’’ (5) for continuous domain.
What we have introduced above, namely the ‘‘golden’’ direct and inverse matrices (21), (22), (26) and (27), allow us
to develop the following application to cryptography. Let the initial message be a ‘‘digital signal’’, which is any
sequence of real numbers:
a1 ; a2 ; a3 ; a4 ; a5 ; a6 ; a7 ; a8 ; . . . ð41Þ
Separate real numbers of the sequence (41) are called readings.
There are many examples of the ‘‘digital signals’’ (41): digital telephony, digital TV, digital measurement systems and
so on.
The problem of protecting the ‘‘digital signal’’ (41) from the hackers is solved usually with application of crypto-
graphic methods. Consider a new cryptographic method based on the ‘‘golden’’ matrices.
To this end let us choose the first four readings a1, a2, a3, a4 of (41) and form from them a square 2 · 2-matrix M:
a1 a2
M¼ : ð42Þ
a3 a4
Note that the initial matrix M can be considered as plaintext [32].
Note that there are 4! = 4 · 3 · 2 · 1 = 24 variants (permutations) to form the matrix (42) from the four readings a1,
a2, a3, a4. Let us designate the ith permutation by Pi (i = 1, 2, . . . , 24). The first step of cryptographic protection of the
four readings a1, a2, a3, a4 is a choice of the permutation Pi.
Then we choose the direct ‘‘golden’’ matrices (21) or (22) as enciphering matrices and their inverse matrices (26) and
(27) as deciphering matrices.
Let us consider now the following encryption/decryption algorithms based on matrix multiplication (see Table 2).
Here M is the plaintext (42) that is formed according to the permutation Pi; E1(x), E2(x) are ciphertexts; Q2x, Q2x+1
are the enciphering matrices (21) and (21); Q2x and Q2x+1 are the deciphering matrices (26) and (27). We can use the
variable x as a cryptographic key or simply a key. This means that in dependence on the value of the key x there is an
infinite number of transformation of the plaintext M into the ciphertext E(x).
In general the key K consists of two parts: permutation Pi and the variable x, that is,
K ¼ fP ; xg:
Table 2
Encryption/decryption algorithm
Encryption Description
M · Q2x = E1(x) E1(x) · Q2x = M
M · Q2x+1 = E2(x) E2(x) · Q2x+1 = M
A.P. Stakhov / Chaos, Solitons and Fractals 32 (2007) 1138–1146 1143
Let us prove that the cryptographic method given with Table 2 ensures one-valued transformation of the plaintext
into the ciphertext E and then the ciphertext E into the plaintext M. Let us consider this transformation for the case
when we choose the ‘‘golden’’ matrix (21) as the enciphering matrix. For the given value of the cryptographic key
x = x1 the ‘‘golden’’ encryption can be represented as follows:
a1 a2 cFsð2x1 þ 1Þ sFsð2x1 Þ e11 e12
M Q2x ¼ ¼ ¼ Eðx1 Þ; ð43Þ
a3 a4 sFsð2x1 Þ cFsð2x1 1Þ e21 e22
where
where
For calculation of the matrix elements given by (49)–(52) we can use the expressions (44)–(47). Then we have:
Using the fundamental identity (9) we can write the expression (53) as follows:
d 11 ¼ a1 : ð54Þ
d 12 ¼ a2 ; ð55Þ
d 21 ¼ a3 ; ð56Þ
d 22 ¼ a4 : ð57Þ
Using (54) and (55) we can write the matrix D (48) as follows:
d 11 d 12 a1 a2
D¼ ¼ ¼ M: ð58Þ
d 21 d 22 a3 a4
This means that a cryptographic method given by Table 2 ensures one-valid transformation of the initial plaintext M at
the entrance of the coder into the same plaintext M at the exit of the decoder.
Let us calculate now the determinant of the ciphertexts, that is, the matrices E1(x), E2(x):
Det E1 ðxÞ ¼ Det M Det Q2x ; ð59Þ
Det E2 ðxÞ ¼ Det M Det Q2xþ1 : ð60Þ
1144 A.P. Stakhov / Chaos, Solitons and Fractals 32 (2007) 1138–1146
If we use the identities (39) and (40), we can write the expressions (59) and (60) in the following form:
Det E1 ðxÞ ¼ Det M; ð61Þ
Det E2 ðxÞ ¼ Det M: ð62Þ
This means that the determinants of the matrices E1(x) and E2(x) is determined identically by the determinant of the
initial matrix M.
According to (43) the encryption consists in calculation of four elements e11, e11, e11, e11 of the matrix (43). According
to (44)–(47) a calculation of every element include two multiplications and one addition. This means that a full encryp-
tion time Te is equal:
T e ¼ 8Dtm þ 4Dtad ; ð63Þ
where Dtm is a time of one multiplication and Dtad is a time of one addition.
By analogy, if we consider the expressions (48)–(52) we can write the expression for a full decryption time:
T d ¼ 8Dtm þ 4Dtad : ð64Þ
Analysis of the expressions (63) and (64) show that the ‘‘golden’’ cryptography is fast cryptography. This means that the
‘‘golden’’ cryptography can be used for cryptographic protection of digital signals in real scale of time.
We can improve the cryptographic protection of the method if we use multiple encryption and decryption. This idea
consists in the following. The first step of encryption is a use of the key
K 1 ¼ fP i ; x1 g
that consists of any permutation Pi and any value x1, taken in random manner. As a result of the encryption we got the
matrix E(Pi, x1) given by (43). The second step of encryption is to use the matrix E(x1) as the initial matrix for encryp-
tion. To this end we can use the second cryptographic key
K 2 ¼ fP j ; x2 g;
where Pj is the next permutation and x2 is the next value of x. After performing the ‘‘golden’’ encryption with the key x2
we can get a new matrix E that is a function of the two permutation Pi and Pj and two cryptographic keys x1 and x2,
that is, E = E(Pj, x1; Pj, x2). In general case we can repeat this procedure n times, that is, the cryptographic key K is a
totality of n random permutations Pi, Pj, . . . , Pk and n cryptographic keys x1, x2, . . . , xn, that is,
K ¼ fP i ; x1 ; P j ; x2 ; . . . ; P k ; xn g: ð65Þ
As a result of multiple encryption we can get the matrix
E ¼ EðKÞ:
For the decryption we have to apply the inverse cryptographic key K1 that is an inverse form of the initial crypto-
graphic key, that is,
K 1 ¼ fP k ; xn ; P r ; xn1 ; . . . ; P j ; x2 ; P i ; x1 g:
It is clear the ‘‘golden’’ cryptographic method relates to symmetrical cryptography [32]. It is considered now that the
main deficiency of the symmetrical cryptography is a problem of key transmission. To this end in the recent decades so-
called asymmetrical cryptography was developed [32]. In the asymmetric cryptosystems we use two keys: open or public
key and secret key. The encryption of the message before transmission is realized by the public key and the decryption of
ciphertext is realized by the secret key. However, the asymmetric cryptography has two deficiencies:
(1) They use very complicated encryption and decryption algorithms. This means that sometimes they cannot be
used for protection of digital signals in real scale of time.
A.P. Stakhov / Chaos, Solitons and Fractals 32 (2007) 1138–1146 1145
(2) Because encryption and decryption algorithms are very complicated this fact puts forward a problem to guaran-
tee that encryption and decryption would be realized without errors (a problem of reliable computations).
The ‘‘golden’’ cryptographic method allows us to solve these two problems very elegantly for the following
conditions:
(1) We can use the existing asymmetrical cryptosystems for the transmission of the cryptographic key (65).
(2) We can use the ‘‘golden’’ cryptographic system given by Table 2 for the fast transmission of the digital signal (41).
(3) We can use the unique mathematical property of the ‘‘golden’’ cryptography given with (61) and (62) for checking
of the encryption and decryption algorithms.
This means that using the ‘‘golden’’ cryptography method (Table 2) we can design fast, simple for technical realiza-
tion and reliable cryptosystems.
Note that for every session of transmission we can change the cryptographic key (65). This means that the analysis of
the previous transmissions cannot be used for uncovering of the current cryptographic key (65). We can change the
cryptographic key (65) using a generator of random numbers and so on.
4. Conclusion
As is shown in [4–31] the application field of the Fibonacci numbers theory is developing with such intensity that we
can call that it is the beginning of an era of ‘‘global Fibonaccization’’ of modern science. ‘‘Harmony Mathematics’’
[10,12,13,16,20], which is modern development of the Fibonacci numbers theory, touches the fundamentals of mathe-
matics and can lead to a new number theory [8,16] and a new measurement theory [4,5,7]. The hyperbolic Fibonacci and
Lucas functions [8,14,15] can lead to creation of Lobachevsky–Fibonacci and Minkovsky–Fibonacci geometry [15] that
is of great importance for theoretical physics. However, the works of Mauldin and Willams, El Naschie, Vladimirov
[22–31] and other physicists show that it is impossible to imagine a development of modern physical research without
deep knowledge of the Golden Section.
The main result of the present article is a development of one more application of the Golden Proportion, that is, a
creation of a new kind of matrices called the ‘‘golden’’ matrices. They are an original synthesis of two important dis-
coveries of modern Fibonacci numbers theory, the Fibonacci Q-matrix (2) and hyperbolic Fibonacci functions (6) and
(7). A new kind of cryptography, the ‘‘golden’’ cryptography, follows from this approach. A common use of the existing
asymmetric cryptosystems (for the transmission of the secret key) and the ‘‘golden’’ cryptosystems (for transmission of
ciphertexts) can lead to the fast, simple for technical realization and reliable cryptosystems for protection of digital sig-
nals used in telecommunication and measurement systems.
References
[1] Vorobyov NN. Fibonacci numbers. Moscow: Nauka; 1978 [in Russian].
[2] Hoggat VE. Fibonacci and Lucas numbers. Palo Alto, CA: Houghton-Mifflin; 1969.
[3] Vajda S. Fibonacci and Lucas numbers, and the golden section. Theory and applications. Ellis Horwood limited; 1989.
[4] Stakhov AP. Introduction into algorithmic measurement theory. Moscow: Soviet Radio; 1977 [in Russian].
[5] Stakhov AP. Algorithmic measurement theory. Moscow: Znanie, Series ‘‘Mathematics and Cybernetics; 1979. No. 6 [in Russian].
[6] Stakhov AP. Codes of the golden proportion. Moscow: Radio and Communications; 1984 [in Russian].
[7] Stakhov AP. The golden section in the measurement theory. Comput Math Appl 1989;17(4–6):613–38.
[8] Stakhov AP, Tkachenko IS. Hyperbolic Fibonacci trigonometry. Rep Ukr Acad Sci 1993;208(7):9–14 [in Russian].
[9] Stakhov AP. A generalization of the Fibonacci Q-matrix. Rep Natl Acad Sci Ukr 1999(9):46–9.
[10] Stakhov AP. The golden section and modern harmony mathematics. Applications of Fibonacci numbers, 7. Kluwer Academic
Publishers; 1998. p. 393–99.
[11] Stakhov AP. Brousentsov’s ternary principle, Bergman’s number system and ternary mirror-symmetrical arithmetic. Comput J
(Br. Comput. Soc.) 2002;45(2):221–36.
[12] Stakhov AP. The generalized golden sections and a new approach to the geometric definition of a number. Ukr Math J
2004;56(8):1143–50 [in Russian].
[13] Stakhov AP. Sacred geometry and harmony mathematics. Vinnitsa: ITI; 2003 [in Russian].
[14] Stakhov AP. Hyperbolic Fibonacci and Lucas functions: a new mathematics for the living nature. Vinnitsa: ITI; 2003.
[15] Stakhov A, Rozin B. On a new class of hyperbolic function. Chaos, Solitons & Fractals 2004;23:379–89.
1146 A.P. Stakhov / Chaos, Solitons and Fractals 32 (2007) 1138–1146
[16] Stakhov A. The generalized principle of the golden section and its applications in mathematics, science and engineering. Chaos,
Solitons & Fractals 2005;26:263–89.
[17] Stakhov A, Rozin B. The golden shofar. Chaos, Solitons & Fractals 2005;26:677–84.
[18] Stakhov A, Rozin B. Theory of Binet formulas for Fibonacci and Lucas p-numbers. Chaos, Solitons & Fractals 2006;27:1162–77.
[19] Stakhov A, Rozin B. The ‘‘golden’’ algebraic equations. Chaos, Solitons & Fractals 2006;27:1415–21.
[20] Stakhov AP. Fundamentals of a new kind of mathematics based on the golden section. Chaos, Solitons & Fractals
2005;27:1124–46.
[21] Spinadel VW. From the golden mean to chaos. Editorial Nueva Librerı́a, Buenos Aires, Argentina, ISBN 950-43-9329-1, 1998,
2nd ed., Nobuko, 2004.
[22] Mauldin RD, Willams SC. Random recursive construction. Trans. Am. Math. Soc. 1986;295:325–46.
[23] El Nashie MS. Quantum mechanics and the possibility of a Cantorian space–time. Chaos, Solitons & Fractals 1992;1:485–7.
[24] El Naschie MS. On dimensions of Cantor set related systems. Chaos, Solitons & Fractals 1993;3:675–85.
[25] El Nashie MS. Is quantum space a random cantor set with a golden mean dimension at the core? Chaos, Solitons & Fractals
1994;4(2):177–9.
[26] El Naschie MS. Statistical geometry of a cantor discretum and semiconductors. Comput Math Appli. 1995;29(12):103–10.
[27] El Naschie MS. Fredholm operators and the wave-particle duality in Cantorian space. Chaos, Solitons & Fractals 1998;9(6):
975–8.
[28] El Naschie MS. On a class of general theories for high energy particle physics. Chaos, Solitons & Fractals 2002;14:649–68.
[29] El Naschie MS. Elementary number theory in superstrings, loop quantum mechanics, twistors and E-infinity high energy physics.
Chaos, Solitons & Fractals 2006;27(2):297–330.
[30] Vladimirov YS. Quark icosahedron, charged and Vainberg’s angle. In: Proceedings of the international conference problems of
harmony, symmetry and the golden section in nature, science and art. Vinnitsa; 2003, vol. 15. p. 69–79 [in Russian].
[31] Vladimirov YS. Metaphysics. Moscow: BINOM; 2002 [in Russian].
[32] Seberry J, Pieprzyk J. Cryptography: an introduction to computer security. Prentice Hall; 1989.