I have a discrete-time system y[n] = x[n]*x[n-1]

I need to show that this system is non-linear by using a counter example rather than by disproving with algebra and the properties of additivity and homogeneity.

I think that the system is homogeneous although not additive, but I can't think of an example to show this. Would someone help me out please?

  • $\begingroup$ Hint: try with constant signals. $\endgroup$ – MBaz Sep 30 '15 at 21:46
  • $\begingroup$ It's not even homogeneous. Really, nearly any two signals will work as a counter example. There's nothing complicated about constructing a counter example. Just take something and plug it in. $\endgroup$ – Jazzmaniac Sep 30 '15 at 22:03
  • $\begingroup$ Hmm when I tried to do the algebra to see if it was homogeneous I got a*x[n] -> a*y[n], I guess I screwed up somewhere. I'm really just not familiar with how this works. Say I take a unit impulse and put it in, then I get a signal of 0 out. Then I shift the unit impulse to the right and put it in and still get a signal of 0 out. I wish I had an example to get me started. $\endgroup$ – Austin Sep 30 '15 at 23:27
  • $\begingroup$ You are testing shift invariance with your example, not linearity. $\endgroup$ – Peter K. Sep 30 '15 at 23:43

OK. Let's try homogeneity again: $$ y[n] = x[n]\cdot x[n-1] $$ for input $x[n]$.

For input $a\cdot x[n]$ we get $$ y'[n] = a\cdot x[n] \cdot a x[n-1] = a^2 \cdot x[n]\cdot x[n-1] \not = ay[n] $$ so homogeneity doesn't apply.

Let's try that with an example, as per MBaz's suggestion.

Let's try: $$ x[n] = 2 $$ so that $$ y[n] = 4 $$ Then let's try twice that: $$ x'[n] = 2\cdot x[n] = 4 $$ So that $$ y'[n] = 16 \not= 2\cdot y[n] $$

  • $\begingroup$ Ah, I added instead of multiplying. Thank you I'm starting to get the examples now $\endgroup$ – Austin Oct 1 '15 at 1:27

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.