Accession Number : AD0700808

Title :   A BRAIN MODEL USING MULTI-VALUED LOGIC.

Descriptive Note : Final rept.,

Corporate Author : MONTEITH COLL DETROIT MICH DIV OF NATURAL SCIENCES

Personal Author(s) : Stern,Alfred L.

Report Date : NOV 1969

Pagination or Media Count : 56

Abstract : The report discusses the feasibility of using multi-valued logic as a possible model for the central nervous system. The three-valued logic of Lukasiewicz and Tarski was selected for investigation. The development follows that of Post which enables operations to be defined in matrix form. The proof of nonreducibility is outlined. The method of proof selects the matrix classes that produce the same theroems as those of a logic with smaller valuational spread. The partitioning of the functionally complete m-valued logics into a restricted set of similarity classes reveals the embedding of the subordinate logics. This enables a logic at any value level to be manipulated as if it were a totally separable fragment of an infinitely many-valued logic. The use of this method to present ordering principles among the propositional statements and to generalize the notion of designated value is illustrated. The construction of a ternary model for discrimination is discussed. The utility of the approach is suggested from the complexity that is possible with only a three-level array of three-valued elements. Areas for further research are outlined.

Descriptors :   (*CYBERNETICS, MATHEMATICAL LOGIC), (*CENTRAL NERVOUS SYSTEM, MODELS(SIMULATIONS)), BRAIN, BIONICS, COMPUTER LOGIC, ALGEBRAS

Subject Categories : Bionics

Distribution Statement : APPROVED FOR PUBLIC RELEASE