摘要: This paper presents a system-theoretic approach to the analysis of problem training formally relating it control an abstract dynamic system, “adaption automaton” trainee. The utility this formulation and possibility basing real strategies upon are discussed, is argued that further constraints automaton both necessary, available, in so far as theory corresponds practical reality. minimal generate extended which related stability adaption automaton. More lead theoretical foundations for “feedback” or “adaptive” training. Corresponding each set “training theorem” proved demonstrates constraint adequate simple universal strategy.