Show that a stochastic process $X(t)$ is mean square continuous if and only if its autocorrelation function $R_X(t_1,t_2)$ is continous

$\Rightarrow$ Proof:

We have $E[(X(t)-X(t_0))^2]=R_X(t,t)-R_X(t_0,t)-R_X(t,t_0)+R_X(t_0,t_0)$, so if $t \rightarrow t_0$ the right side of the equation is equal to $0$, therefore $\Rightarrow$ implication is true.

$\Leftarrow$ Proof:

This is where I have trouble.

Is this implication true due to the fact that if $X(t)$ is continuous at $t_0$ in m.s then $\lim_{t \to t_0} m_X(t)=m_X(t_0)$ ??


The whole point of the concept of mean square continuity is to avoid strict assumptions about the continuity of every realization of $X(t)$.

The given proof simply shows that continuity of the autocorrelation function implies that the limit $\lim_{t\rightarrow t_0}E\{[X(t)-X(t_0)]^2\}$ exists and equals zero.

Note that continuity of the mean of $X(t)$ is a consequence of mean square continuity. Since

$$E\{[X(t)-X(t_0)]^2\}\ge E^2\{X(t)-X(t_0)\}\tag{1}$$

it follows from

$$\lim_{t\rightarrow t_0}E\{[X(t)-X(t_0)]^2\}=0$$


$$\lim_{t\rightarrow t_0}E\{X(t)-X(t_0)\}=0$$

must hold, i.e., the mean of $X(t)$ is continuous.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.