Stichworte: rewarding successive approximations