Abstract:
Decision rules (for recognition algorithms) with additional monotonicity constraints are investigated. It is shown that the correctness of an algorithm that uses these decision rules is equivalent to the standard notion of correctness studied in the classical works on the algebraic approach to the recognition problem.