www.seipub.org/ijc International Journal on Communications (IJC) Volume 3, 2014
16
Voting Procedure and Neural Networks
Ibragim Suleimenov
*1
, Oleg Gabrielyan
2
, Grigoriy Mun
3
, Sergey Panchenko
1,3
, Tungysh Amirzhan
1
,
Kamilya Suleimenov
4
1
Almaty University of Power Engineering and Telecommunications, Almaty, Kazakhstan
2
Crimean University of Culture, Arts and Tourism, Simferopol, Ukraine
3
Al-Faraby Kazakh National University, Almaty, Kazakhstan
4
Birmingham University, Birmingham, United Kingdom
*
esenych@yandex.ru; gabroleg@mail.ru; mungrig@yandex.ru; sergey@panchen.co; greentuns@gmail.com; mlldefer@yandex.ru
Received Jan 21, 2014; Accepted Mar 17, 2014; Published Jun 18, 2014
© 2014 Science and Engineering Publishing Company
Abstract
It is shown that even elementary systems, in which
behaviour determined by the voting procedures often
cannot come to the stabilized conditions. The fact is
determined by different oscillations that can occur in such
systems compliant with polar change of opinions with the
certain frequency. This conclusion has the fundamental
importance; in particular, for theoretic sociology because it
shows that there are no functional, which may be used for
mathematical description of the systems of the considered
type by their maximization of minimization.
Keywords
Voting Procedure; Neural Network; Theoretic Sociology
Introduction
As mentioned in (McKenzie, 1986), nowadays a
number of economical models are created on the basis
of the analogy with classic mechanics. A number of
functional are proposed (for example, a utility
function), whose extremum is compliant with an
equilibrium condition of the system where it
spontaneously comes after some time. Nowadays, a
wide range of such functions is in use (Arrow, 1954
and Balasko, 1989).
As highlighted in (Polterovich, 1997), all the efforts to
create the similar description (i.e. to determine an
evident kind of the functional whose extremum could
be compliant with the system equilibrium condition)
have not been successful yet within sociologic
sciences. Besides, it is supposed in (Polterovich, 1997)
that such functional cannot exist at all. Reasons for
such point of view are based on the Arrow’s
impossibility theorem (Arrow, 1951) and similar
results obtained for the last decades.
Based on the analogy with neuronal networks, it is
proved in this paper that there are sets of examples of
systems of social nature where any stabilized
condition is impossible at all. We consider a certain
example – a Council where voting takes place;obtains
conclusions can be generalized for any voting
procedures.
It is shown in (Suleimenov, 2010) that a feedback
diagram in the voting Council topologically is
completely equivalent to the feedback diagram in the
Hopfield neural processor. However, as determined in
this work, matrix of weight coefficients describing a
feedback system in the voting Council, as usual, is not
symmetrical against a classic diagram of the Hopfield
neuroprocessor. This factor causes instabilities
(oscillation).
It is also shown in this work that the principal
impossibility of the stabile state results in
hypersensitivity of the social system of the considered
type to the external impacts. In particular, there is a
mode of oscillation synchronization when an external
impact with insignificant amplitude can cause
inconvertible changes of the system in whole.
This result provides an opportunity of quantitative
description of such phenomena as influence of
telecommunications on initiation of the revolutionary
situations (such as ‘Arabic Spring’) and the certain
recommendations to manage the processes occurring
in the society.
Basic Model
Background
Fig. 1 provides an example of the classic Hopfield
neuroprocessor diagram. This diagram contains three
formal neurons whose outputs are mutually connected