Asia Pacific Mathematics Newsletter

What
What are are p-Adic
p-Adic Numbers?
Numbers? What are
What areTheyThey
Used Used
for? for?
U A Rozikov
U A Rozikov

Abstract. In this short paper we give a popular intro- rational numbers, such that |2|∗ < 1. This is done
duction to the theory of p-adic numbers. We give some as follows. Let Q be the field of rational numbers.
properties of p-adic numbers distinguishing them to
Every rational number x  0 can be represented
“good” and “bad”. Some remarks about applications
of p-adic numbers to mathematics, biology and physics in the form x = pr mn , where r, n ∈ Z, m is a positive
are given. integer, ( p, n) = 1, ( p, m) = 1 and p is a fixed prime
number. The p-adic absolute value (norm) of x is
1. p-Adic Numbers given by
 −r
p-adic numbers were introduced in 1904 by the  p , for x  0,


|x|p = 
 0, for x = 0.

German mathematician K Hensel. They are used
intensively in number theory. p-adic analysis was The p-adic norm satisfies the so called strong
developed (mainly for needs of number theory) triangle inequality
in many directions, see, for example, [20, 50].
|x + y|p ≤ max{|x|p , |y|p}, (2)
When we write a number in decimal, we can
only have finitely many digits on the left of the and this is a non-Archimedean norm.
decimal, but we can have infinitely many on the This definition of |x|p has the effect that high
right of the decimal. They might “terminate” (and powers of p become “small”, in particular |2n |2 =
become all zeros after some point) but they might 1/2n . By the fundamental theorem of arithmetic,
not. The p-adic integers can be thought of as for a given non-zero rational number x there is a
writing out integers in base p, but one can have unique finite set of distinct primes p1 , . . . , pr and
infinitely many digits to the left of the decimal a corresponding sequence of non-zero integers
(and none on the right; but the rational p-adic a1 , . . . , an such that x = pa11 . . . par r . It then follows
numbers can have finitely many digits on the that |x|pi = p−a i
i
for all i = 1, . . . , r, and |x|p = 1 for
right of the decimal). For example, the binary any other prime p  {p1 , . . . , pr }.
expansion of 35 is 1·20 +1·21 +0·22 +0·23 +0·24 +1·25, For example, take 63/550 = 2−1 · 32 · 5−2 · 7 · 11−1
often written in the shorthand notation 1000112. we have
One has 1 = 0, 111111111 . . .2 = 0, (1)2. But what is 

 2, if p = 2,

. . . 111111, 02 = (1), 02? Compute (1), 02 + 1: 

1/9, if p = 3,





. . . 111111, 02



25, if p = 5,


�� 63 ��� = 
�� � 
+. . . 000001, 02 

. . . 000000, 02 � 550 �p  


 1/7, if p = 7,


11, if p = 11,

Hence (1), 02 = −1. This equality can be written as







 1, if p ≥ 13.
n−1



�
(1), 02 = lim 2i = lim (2n − 1) = −1. (1)
n→∞ n→∞
i=0 We say that two norms � · �1 and � · �2 on Q are
This limit equivalent to limn→∞ 2 = 0. In real
n equivalent if there exists α > 0 such that
case one has limn→∞ qn = 0 if and only if absolute
� · �α1 = � · �2 .
value |q| is less than 1. Remember that to define
real numbers one considers all limit points of It is a theorem of Ostrowski (see [41]) that
sequences of rational numbers, using the absolute each absolute value on Q is equivalent either
value as metric. to the Euclidean absolute value | · |, the trivial
To give a meaning of the limit (1), one has absolute value, or to one of the p-adic absolute
to give a new absolute value | · |∗ , on the set of values for some prime p. So the only norms on

October 2013, Volume 3 No 4 1

 Asia Pacific Mathematics Newsletter 2

The sum ∞

Q modulo equivalence are the absolute value, n=0 n! exists in every Qp . The follow-
the trivial absolute value and the p-adic absolute ing problem has been open since 1971.
∞
value which means that there are only as many Problem. Can n=0 n! be rational for some
completions (with respect to a norm) of Q. prime p?
The p-adic absolute value defines a metric It is not known if ∞

n=0 n!  0 in every Qp .
|x − y|p on Q. Two numbers x and y are p-adically 7. For any x ∈ Q, we have
closer as long as r is higher, such that pr divides 
|x − y|p . Amazingly, for p = 5 the result is that 135 |x| |x|p = 1.
p:prime
is closer to 10 than 35.
The completion of Q with respect to p-adic This formula have been used to solve several
norm defines the p-adic field which is denoted problems in number theory, many of them us-
by Qp . Any p-adic number x  0 can be uniquely ing Helmut Hasse’s local-global principle, which
represented in the canonical form roughly states that an equation can be solved
over the rational numbers if and only if it can be
x = pγ(x) (x0 + x1 p + x2 p2 + · · · ),
solved over the real numbers and over the p-adic
where γ = γ(x) ∈ Z and xj are integers, 0 ≤ xj ≤ numbers for every prime p.
p − 1, x0 > 0, j = 0, 1, 2, ... (see more detail [31,
50, 54]). In this case |x|p = p−γ(x) . The set of p-adic 3. “Bad” Properties of p-Adic Numbers
numbers contains the field of rational numbers Q
but is different from it. 1. Qp is not ordered.
Using canonical form of p-adic numbers, simi- 2. Qp is not comparable with R, for example
√ √
larly as real numbers, one makes arithmetic oper- 7  Q5 , but i = −1 ∈ Q5 .
ations on p-adic numbers (see for example, [41]). 3. Qp is not algebraically closed.
But | · |p can be extended uniquely to the
2. “Good” Properties of p-Adic Numbers algebraic closure Qap and the completion of (Qap , |·|p )
is called Cp , the field of the p-adic complex num-
The ultra-metric triangle inequality, i.e. (2), under- bers. Cp is no locally compact, but separable and
lies many of the interesting differences between algebraically closed.
real and p-adic analysis. The following properties Now define the functions expp (x) and logp (x).
of p-adic numbers make some directions of the Given a ∈ Qp and r > 0 put
p-adic analysis more simple than real analysis:
1. All triangles are isosceles. B(a, r) = {x ∈ Qp : |x − a|p < r}.
2. Any point of ball D(a, r) = {x ∈ Qp : |x−a|p ≤ r}
The p-adic logarithm is defined by the series
is center. Each ball has an empty boundary. Two
∞
balls are either disjoint, or one is contained in the  (x − 1)n
logp (x) = logp (1 + (x − 1)) = ( − 1)n+1 ,
other. n=1
n
3. | · |p1  | · |p2 if p1  p2 . This means that each
which converges for x ∈ B(1, 1);
prime number p generates its own field of p-adic
The p-adic exponential is defined by
numbers Qp .
∞
4. x2 = −1 has a solution x ∈ Qp if and only if  xn
expp (x) = ,
p = 1 mod 4. n=0
n!
5. A sequence {xn } in Qp is a Cauchy sequence
which converges for x ∈ B(0, p−1/(p−1)).
if and only if |xn+1 − xn |p → 0 as n → ∞.
Let x ∈ B(0, p−1/(p−1), then
This has the useful corollary that a sum con-
verges if and only if the individual terms tend to | expp (x)|p = 1, | expp (x)−1|p = |x|p , | logp (1+x)|p = |x|p ,
zero:
6. (A student’s dream) ∞ logp ( expp (x)) = x, expp ( logp (1 + x)) = 1 + x.

n=1 an < ∞ if and only
if an → 0.
4. Some “good” functions become “bad”. For
Since |n!|p → 0 we have, for example,
example exp (x) is very “good” function on R, but
∞ ∞
  as we seen above expp (x) is defined only on ball
( − 1)n n!(n + 2) = 1, ( − 1)n n!(n2 − 5) = −3.
n=0 n=0 B(0, p−1/(p−1) ).

2 October 2013, Volume 3 No 4

 Asia Pacific Mathematics Newsletter
3

4. Remarks about Applications p-adic mathematics in many directions: theory of

distributions [6, 31], differential and pseudodif-
When the p-adic numbers were introduced they ferential equations [32, 56], theory of probability
considered as an exotic part of pure mathematics [31, 56] spectral theory of operators in a p-adic
without any application (see for example [41, analogue of a Hilbert space [7, 8, 33].
42, 50, 56] for applications of p-adic numbers The representation of p-adic numbers by se-
to mathematics). Since p-adic numbers have the quences of digits gives a possibility to use this
interesting property that they are said to be close number system for coding of information. There-
when their difference is divisible by a high power fore p-adic models can be used for the description
of p the higher the power the closer they are. of many information processes. In particular, they
This property enables p-adic numbers to encode can be used in cognitive sciences, psychology and
congruence information in a way that turns out sociology. Such models based on p-adic dynamical
to have powerful applications in number theory systems [3–5].
including, for example, in the famous proof of The study of p-adic dynamical systems arises
Fermat’s Last Theorem by Andrew Wiles (see [42, in Diophantine geometry in the constructions
Chap. 7]). of canonical heights, used for counting rational
What is the main difference between real and points on algebraic varieties over a number field,
p-adic space-time? It is the Archimedean axiom. as in [21].
According to this axiom any given large segment There most recent monograph on p-adic dy-
on a stright line can be surpassed by successive namics is Anashin and Khrennikov [9]; nearly a
addition of small segments along the same line. half of Silverman’s monograph [52] also concerns
This axiom is valid in the set of real numbers and p-adic dynamics.
is not valid in Qp . However, it is a physical axiom Here are areas where p-adic dynamics proved
which concerns the process of measurement. To to be effective: computer science (straight line
exchange a number field R to Qp is the same as programs), numerical analysis and simulations
to exchange axiomatics in quantum physics (see (pseudorandom numbers), uniform distribution
[31, 56]). of sequences, cryptography (stream ciphers, T-
In 1968 two pure mathematicians, A Monna functions), combinatorics (Latin squares), au-
and F van der Blij, proposed to apply p-adic tomata theory and formal languages, genetics.
numbers to physics. In 1972 E Beltrametti and The monograph [9] contains the corresponding
G Cassinelli investigated a model of p-adic valued survey. For a newer results see recent papers and
quantum mechanics from the positions of quan- references therein: [10, 14, 15, 28, 36, 37, 38, 48, 51].
tum logic. Since 80th p-adic numbers are used in Moreover, there are studies in computer science
applications to quantum physics. p-adic strings and cryptography which along with mathematical
and super strings were the first models of p-adic physics stimulated in 1990th intensive research in
quantum physics (see, for example, [17, 29, 50, p-adic dynamics since it was observed that major
54]). The interest of physicists to p-adic numbers computer instructions (and therefore programs
is explained by the attempts to create new models composed of these instructions) can be considered
of space-time for the description of (fantastically as continuous transformations with respect to the
small) Planck distances. 2-adic metric, see [11, 12].
There are some evidences that the standard In [33, 53] p-adic field have arisen in physics in
model based on real numbers is not adequate the theory of superstrings, promoting questions
to Planck’s domain. On the other hand, some about their dynamics. Also some applications
properties of fields of p-adic numbers seem to of p-adic dynamical systems to some biological,
be closely related to Planck’s domain. In partic- physical systems has been proposed in [3, 4, 5,
ular, the fields of p-adic numbers have no order 22, 23, 33, 35]. Other studies of non-Archimedean
structure. dynamics in the neighborhood of a periodic point
The pioneer investigations on p-adic string and of the counting of periodic points over global
theory induced investigations on p-adic quantum fields using local fields appear in [39, 47]. It is
mechanics and field theory (see the books [31, 54, known that the analytic functions play important
55]). This investigations induce a development of role in complex analysis. In the p-adic analysis

October 2013, Volume 3 No 4 3

 Asia Pacific Mathematics Newsletter 4

the rational functions play a role similar to that of References

analytic functions in complex analysis [49]. There-
[1] S. Albeverio and W. Karwowski, A random walk
fore, there naturally arises a question on study the on p-adics — the generator and its spectrum,
dynamics of these functions in the p-adic anal- Stochastic Processes Appl. 53 (1994) 1–22.
ysis. On the other hand, these p-adic dynamical [2] S. Albeverio and X. Zhao, Measure-valued branch-
ing processes associated with random walks on
systems appear while studying p-adic Gibbs mea- p-adics, Ann. Probab. 28 (2000) 1680–1710.
sures [26, 24, 44–46]. In [18, 19] dynamics on the [3] S. Albeverio, U. A. Rozikov and I. A. Sattarov,
p-adic (2, 1)-rational dynamical systems, J. Math.
Fatou set of a rational function defined over some Anal. Appl. 398(2) (2013) 553–566.
finite extension of Qp have been studied, besides, [4] S. Albeverio, A. Khrennikov and P. E. Kloeden,
an analogue of Sullivan’s no wandering domains Memory retrieval as a p-adic dynamical system,
BioSys. 49 (1999) 105–115.
theorem for p-adic rational functions which have [5] S. Albeverio, A. Khrennikov, B. Tirozzi and S.
no wild recurrent Julia critical points was proved. De Smedt, p-adic dynamical systems, Theor. Math.
Phys. 114 (1998) 276–287.
In [27] the behaviour and ergodicity of a p-adic [6] S. Albeverio, A. Y. Khrennikov and V. M.
dynamical system f (x) = xn in the fields of p-adic Shelkovich, Theory of p-Adic Distributions: Linear
numbers Qp and complex p-adic numbers Cp was and Nonlinear Models (Cambridge University Press,
2010).
investigated. Firstly, the problem of ergodicity of [7] S. Albeverio and A. Yu. Khrennikov, Represen-
perturbed monomial dynamical systems which tation of the Weyl group in spaces of square
integrable functions with respect to p-adic valued
was posed in these papers and which stimulated Gaussian distributions, J. Phys. A 29 (1996) 5515–
intensive research, was solved in [13]. Secondly, 5527.
quite recently a far-going generalisation of the [8] S. Albeverio, R. Cianci and A. Yu. Khrennikov, A
representation of quantum field Hamiltonians in
problem for arbitrary 1-Lipschitz transformations a p-adic Hilbert space, Theor. Math. Phys. 112(3)
of 2-adic spheres was also solved in [16]. Finally, (1997) 355–374.
[9] V. Anashin and A. Khrennikov, Applied Algebraic
we note that not only polynomial and rational Dynamics, de Gruyter Expositions in Mathematics,
p-adic dynamical systems has been studied: In Vol. 49 (Walter de Gruyter, Berlin, New York,
past decade, a significant progress was achieved 2009).
[10] V. S. Anashin, A. Yu. Khrennikov and E. I. Yurova,
in a study of a very general p-adic dynamical Characterization of ergodicity of p-adic dynamical
systems like non-expansive, locally analytic, shift- systems by using van der Put basis, Doklady Math.
83(3) (2011) 306–308.
like, etc.
[11] V. S. Anashin, Uniformly distributed sequences of
It is also known [33, 41, 43, 56] that a num- p-adic integers, Math. Notes 55(2) (1994) 109–133.
ber of p-adic models in physics cannot be de- [12] V. S. Anashin, Uniformly distributed sequences in
computer algebra, or how to construct program
scribed using ordinary Kolmogorov’s probabil- generators of random numbers, J. Math. Sci. 89(4)
ity theory. In [34] an abstract p-adic probability (1998) 1355–1390.
[13] V. Anashin, Ergodic transformations in the space
theory was developed by means of the theory of p-adic integers, in p-Adic Mathematical Physics.
of non-Archimedean measures. Applications of 2nd Inter. Conference (Belgrade, Serbia and Mon-
the non-Kolmogorov theory of probability can tenegro 15–21 September 2005), eds. Andrei Yu.
Khrennikov, Zoran Rakic, and Igor V. Volovich,
be considered not only in physics, but in many Vol. 826 AIP Conference Proceedings (American
other sciences, especially in biology and possibly Institute of Physics, Melville, New York, 2006),
pp. 3–24.
in sociology. The general principle of statistical [14] V. Anashin, Non-Archimedean ergodic theory and
stabilisation of relative frequencies is a new pos- pseudorandom generators, The Computer J. 53(4)
sibility to find a statistical information in the (2010) 370–392.
[15] V. Anashin, Automata finiteness criterion in terms
chaotic (from the real point of view) sequences of van der Put series of automata functions, p-
of frequencies [31, 1, 2]. Adic Numbers, Ultrametric Anal. Appl. 4(2) (2012)
151–160.
We refer the reader to [30, 24, 44–46] where [16] V. Anashin, A. Khrennikov and E. Yurova, Ergod-
various models of statistical physics in the context icity Criteria for Non-Expanding Transformations of
2-Adic Spheres, preprint, arXiv:1205.0615.
of p-adic fields are studied.
[17] L. Ya. Araf’eva, B. Dragovich, P. H. Frampton and
A non-Archimedean analogue of the Kol- I. V. Volovich, Wave function of the universe and
mogorov theorem was proved in [25]. Such p-adic gravity, Mod. Phys. Lett. A 6 (1991) 4341–
4358.
a result allows to construct wide classes of [18] R. Benedetto, Hyperbolic maps in p-adic dynam-
stochastic processes and the possibility to de- ics, Ergod. Theor. Dynam. Sys. 21 (2001) 1–11.
[19] R. Benedetto, p-Adic dynamics and Sullivan’s no
velop statistical mechanics in the context of p-adic wandering domains theorem, Compositio Math. 122
theory. (2000) 281–298.

4 October 2013, Volume 3 No 4

 Asia Pacific Mathematics Newsletter
5

[20] Z. I. Borevich and I. R. Shafarevich, Number Theory of compact-open subsets of non-archimedean local
(Academic Press, New York, 1966). fields, Trans. Amer. Math. Soc. 361(1) (2009) 61–85.
[21] G. Call and J. Silverman, Canonical height on va- [37] J. Kingsbery, A. Levin, A. Preygel and C. E. Silva,
rieties with morphisms, Compositio Math. 89 (1993) Dynamics of the p-adic shift and applications, Disc.
163–205. Contin. Dyn. Syst. 30(1) (2011) 209–218.
[22] B. Dragovich, A. Yu. Khrennikov, S. V. Kozyrev [38] D. Lin, T. Shi and Z. Yang, Ergodic theory over
and I. V. Volovich, On p-adic mathematical F2 [[X]], Finite Fields Appl. 18 (2012) 473–491.
physics, p-Adic Numbers, Ultrametric Anal. Appl. [39] J. Lubin, Nonarchimedean dynamical systems,
1(1) (2009) 1–17. Composito Math. 94(3) (1994) 321–346.
[23] D. Dubischer, V. M. Gundlach, A. Khrennikov and [40] E. Marinari and G. Parisi, On the p-adic five point
O. Steinkamp, Attractors of random dynamical function, Phys. Lett. B 203 (1988) 52–56.
system over p-adic numbers and a model of ‘noisy’ [41] N. Koblitz, p-Adic Numbers, p-Adic Analysis, and
cognitive process, Phys. D 130 (1999) 1–12. Zeta-Functions (Springer, Berlin, 1977).
[24] G. Gandolfo, U. A. Rozikov and J. Ruiz, On p- [42] Yu. I. Manin and A. A. Panchishkin, Introduction
adic Gibbs measures for hard core model on a to Modern Number Theory (Springer, Berlin, 2007).
Cayley tree, Markov Processes Related Fields 18(4) [43] E. Marinary and G. Parisi, On the p-adic five point
(2012) 701–720. function, Phys. Lett. B 203 (1988) 52–56.
[25] N. N. Ganikhodjaev, F. M. Mukhamedov and U. [44] F. M. Mukhamedov and U. A. Rozikov, On ratio-
A. Rozikov, Phase transitions in the Ising model nal p-adic dynamical systems, Methods Func. Anal.
on Z over the p-adic number field, Uzb. Mat. Zh. Topology 10(2) (2004) 21–31.
4 (1998) 23–29. [45] F. M. Mukhamedov and U. A. Rozikov, On Gibbs
[26] N. N. Ganikhodjaev, F. M. Mukhamedov and U. measures of p-adic Potts model on Cayley tree,
A. Rozikov, Existence of phase transition for the Indag. Math. (N.S.) 15(1) (2004) 85–100.
Potts p-adic model on the set Z, Theor. Math. Phys. [46] F. M. Mukhamedov and U. A. Rozikov, On inho-
130 (2002) 425–431. mogeneous p-adic Potts model on a Cayley tree,
[27] V. M. Gundlach, A. Khrennikov and K. O. Lindahl, Inf. Dim. Anal. Quant. Prob. Rel. Fields 8(2) (2005)
On ergodic behavior of p-adic dynamical systems, 277–290.
Infin. Dimen. Anal. Quantum Probab. Relat. Top. 4 [47] T. Pezda, Polynomial cycles in certain local do-
(2001) 569–577. mains, Acta Arith. 66 (1994) 11–22.
[28] A.-H. Fan and L.-M. Liao, On minimal decomposi- [48] J.-E. Pin, Profinite methods in automata theory,
tion of p-adic polynomial dynamical systems, Adv. in Symposium on Theoretical Aspects of Computer
Math. 228 (2011) 2116–2144. Science — STACS 2009, (Freiburg, 2009), pp. 31–50.
[29] P. G. O. Freund and E. Witten, Adelic string am- [49] A. M. Robert, A Course of p-Adic Analysis (Springer,
pletudes, Phys. Lett. B 199 (1987) 191–194. New York, 2000).
[30] M. Khamraev, F. M. Mukhamedov and U. A. [50] W. Schikhof, Ultrametric Calculas (Cambridge Uni-
Rozikov, On the uniqueness of Gibbs measures for versity, Cambridge, 1984).
p-adic non homogeneous λ-model on the Cayley [51] T. Shi, V. Anashin and D. Lin, Linear weaknesses
tree, Lett. Math. Phys. 70 (2004) 17–28. in T-functions, in SETA 2012, eds. T. Helleseth
[31] A. Yu. Khrennikov, p-Adic Valued Distributions in and J. Jedwab, Lecture Notes Comp. Sci., Vol.
Mathematical Physics (Kluwer, Dordrecht, 1994). 7280 (Springer-Verlag, Berlin, Heidelberg, 2012),
[32] A. Yu. Khrennikov, Mathematical methods of the pp. 279–290.
non-Archimedean physics, Uspekhi Mat. Nauk 45(4) [52] J. Silverman, The Arithmetic of Dynamical Systems,
(1990) 79–110. Graduate Texts in Mathematics, No. 241 (Springer-
[33] A. Yu. Khrennikov, Non-Archimedean Analysis: Verlag, New York, 2007).
Quantum Paradoxes, Dynamical Systems and Biolog- [53] E. Thiran, D. Verstegen and J. Weters, p-adic dy-
ical Models (Kluwer Academic Publishers, Dor- namics, J. Stat. Phys. 54(3/4) (1989) 893–913.
drecht, The Netherlands, 1997). [54] V. S. Vladimirov, I. V. Volovich and E. I. Zelenov, p-
[34] A. Yu. Khrennikov, S. Yamada and A. van Rooij, Adic Analysis and Mathematical Physics [in Russian],
The measure-theoretical approach to p-adic prob- Nauka, Moscow (1994); English transl., World Sci-
ability theory, Ann. Math. Blaise Pascal 6 (1999) entific, Singapore (1994).
21–32. [55] I. V. Volovich, p-adic strings, Class. Quantum Gray.
[35] A. Yu. Khrennikov, The description of Brain’s func- 4 (1987) L83–L87.
tioning by the p-adic dynamical system, preprint [56] V. S. Vladimirov, I. V. Volovich and E. I. Zelenov,
No. 355 (SFB-237), Ruhr Univ. Bochum, Bochum The spectral theory in the p-adic quantum me-
(1997). chanics, Izvestia Akad. Nauk SSSR, Ser. Mat. 54(2)
[36] J. Kingsbery, A. Levin, A. Preygel and C. E. (1990) 275–302.
Silva, On measure-preserving c1 transformations

October 2013, Volume 3 No 4 5

 Asia Pacific Mathematics Newsletter

U A Rozikov
Institute of Mathematics, Tashkent, Uzbekistan
rozikovu@yandex.ru

U A Rozikov is a professor in Institute of Mathematics, Tashkent, Uzbekistan. He gradu-

ated from the Samarkand State University (1993). He otained PhD (1995) and Doctor of
Sciences in physics and mathematics (2001) degrees from the Institute of Mathematics,
Tashkent. He is known for his works on the theory of Gibbs measures of models on trees
of statistical mechanics (recently he published a book: Gibbs Measures on Cayley Trees
(World Scientific, 2013)). He developed a contour method to study the models on trees
and described complete set of periodic Gibbs measures. Rozikov has (with F Mukhamedov)
an important contribution to non-Archimedean theory of phase transitions and dynamical
systems. He has important results on non-Volterra quadratic operators. N Ganikhodjaev and
Rozikov gave a construction of a quadratic operator which connects phases of the models of
statistical mechanics with models of genetics. His most recent works are devoted to evolu-
tion algebras of sex-linked populations. He was invited to several leading universities and
research centres in the UK, France, Italy, Germany, Spain etc. His (more than 100) papers
are published in front-line journals.

6 October 2013, Volume 3 No 4