THE NUMBER OF MONOTONE FUNCTIONS OF TWO-VALUED LOGIC
Main researcher
: VUKOVIĆ, ANTE (53861) Assistants
Duration from: 01/01/95. Papers on project (total): 0
Institution name: Fakultet elektrotehnike,strojarstva i brodogradnje, Split (23) Department/Institute: FESB - Department of Mathematics and Physics Physics Department Address: Ruđera Boškovića b.b. City: 21000 - Split, Croatia
Communication
Phone: 385 (0)21563777
Fax: 385 (0)21563877
Summary: An outstanding problem in the theory of multivalued logic
is the so-called completness. This problem is connected with functions on
ordered sets. Let E_k be such a set, where k is a number of elements of
E_k. Then, the so-called n-th Cartesian product Ečn_k is defined in
natural way; Ečn_k is again an ordered set. Moreover, MONOTONE functions
f(x_1,...,x_n) of n variables from the set E_k are defined. All monotone
functions f(x_1,...,x_n) form one of the so-called PRECOMPLETE CLOSED
classes, which is crucial for k-valued logic completness. Many authors
tried to determine the number M=Psi_E_k(n) of monotone functions of n
variables and the asymptotics for Psi_E_k(n) when n aims to infinite. The
problems was partially solved about thirty years ago for a two-member set
E_2, giving answers to a number of problems in the two-valued logic. For
k-valued logic there are estimates, which in the last twenty years was not
possible to improve. In 1974., the russian mathematician V.B. Alekseev
improved shightly an old result according to which log Psi_E_k(n) behaves
kčn/nč0.5. His methods differ from methods used by the head of this
project in multiple-valued logic stueies; the last also give some
estimates. But, as this theory evolve slowly, the subscribed belives that
the field deserves permanent watch, especially because of increasing
importance of these topics for applications. Applications encourage the
endeavor to throw more light on these problems. Having no round-up and
far-developed theory at disposal, applications use approximate and
experiene methods. In that, more and more appears the connection between
the multiple-valued logic and so-called fuzzy-decision theories. The head
of the project belives that this connection revives the interest for
multiple_valued logics, and that using various approaches, inclusive his
own, the results must be positive. In favour of the above stataments we can
quote the fact, that in two most spread mathematical informative journals,
in the Notices and in the Bulletin of the American Mathematical Society
(AMS), which annually review more then hundred books and publish as much
articles, in the last few years there are almost no reports on
multiple-valued logics. Only one book has been reviewed in the Bulletin
AMS (J.R. Buchi: Finite Automata, 1988.) which touches these problems.
But, in the January number of the Notices AMS for 1993. (p.15) there is an
interesting review on the by L.Wos's books on Automated Reasoning, which
refreshes the whole subject, and in the February number for 1995. of the
same journal (p.297) there is a call for papers for the new journal:
Multiple-Valued Logic (Gordon and Breach publ., Newark, NJ, USA; Editors:
Dan A. Simovici and Ivan Stojmenovic). Let us the four fields of interest
of the new journal, as they are announced: 1. Mathematical and Engeneering
Aspects of Multiple-Valued Logic (MVL); 2. MVL and Automated Reasoning;
3. Computer Science and MVL; 4. Theoretical and Practical Aspectd of Fuzzy
Logic and Philosophical Aspects of MVL. Therefore, the
subscribed strongly belives that research on this topics must be
continued. Also, it is indispensable to attract new researchers to the
project, regardless the fact that results at this moment have not a
definite form. The subscribed proposes a financial support for at least
one institution in Croatia with the aim to acquire the new MVL journal
(there are no possibilityto exchange it for Glasnik Matematicki). There is
an unpleasent possibillity that in the field, where we have results and
current research, in few years there will be no research, and that once
and again we shall find us at the verz begining.