The notion of partial deduction of a predicate formula from a set of atomic formulas without variables or their negations is introduced. Such a notion allows to solve various pattern recognition problems with incomplete description of a recognizable object. Algorithms based on partial deduction which solve such problems are constructed. These algorithms run upper bounds of step number are proved.
|Journal||ВЕСТНИК САНКТ-ПЕТЕРБУРГСКОГО УНИВЕРСИТЕТА. СЕРИЯ 10: ПРИКЛАДНАЯ МАТЕМАТИКА, ИНФОРМАТИКА, ПРОЦЕССЫ УПРАВЛЕНИЯ|
|State||Published - 2009|