PROBLEMS OF DISCRETE EXTRAPOLATION,
Abstract
The article is a continuation of previous works by the author in which the problem of pattern recognition and the method of incomplete mathematical induction were interpreted respectively as interpolation and extrapolation of arithmetic functions. A general principle was formulated stating that the arithmetic functions should be defined in such a way that the algorithm for computation requires a minimum number of memory cells. Problems relating to formulation of this principle for the case of discrete extrapolation are examined in this paper. A simplified model of a control machine is proposed in which memory organization is described by a single table of addresses. The program for computing the arithmetic function is given by a finite or infinite Moore automaton. The automaton may be binary, ternary or some other type, depending on the computational system. Although only binary automata are considered in this article, the results are valid for any Moore automaton in the proper formulation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 16, 1968
- Accession Number
- AD0685520
Entities
People
- A. Sh. Blokh
Organizations
- National Air and Space Intelligence Center