The control problem of a production process with output Y(t) = M(t) + e(t) (e(t) being a white noise function with E(e(t)) = σ2 for all time points and M(t) := E(Y(t))) can be considered as the problem of monitoring the unknown expectation function M(t). Because in practice we get observations only at dicrete time points tn, n ≥ 1, we have to control the sequence of expectations Mn := M(tn) = E(Y(tn)), Yn := Y(tn) = Mn + en. Let us (as usual) define the process to be in control with respect to a given target MQ iff Mn ≡ MO. If there exists a time point v such that Mn = MQ ∀n < ν and Mn ≠ MQ ∀n ≥ ν then the process is said to be in control up to time point (ν — 1) and out of control from time point ν and ν is called change point.