What is Feedback?
In general feedback is when part of the response of the system is fed back into the system as input. This means that the system’s current behaviour affects its future behaviour. There are two basic types of feedback:
- Positive feedback – the fed back output re-enforces, or amplifies the behaviour of the system leading to ever increasing intensity of the output.
- Negative feedback – the fed back output in some manner reduces the consequent output behaviour, which leads to a regulatory or decaying effect.
A complex system may have many different instances of feedback and feedback can be consciously used to change the behaviour of a system.
A Simple Example of Complex Feedback!
Let’s consider some process ‘A’. This is a complex process with a couple of parameters that we can change. The details of the process are not important but for those who are interested it is just a simple logistics map that generates complex behaviour.
Now let’s feed another process ‘B’ with the output from process ‘A’ and then use the output from process ‘B’ as input to process ‘A’.
Process ‘B’ is fundamentally the same process as process ‘A’ but we can adjust it to have different parameters.
In the graphs below we can see the output from just process ‘A’ in GREEN and the output from process ‘A’ when the input is fed from process B in RED. In both of these cases we start with the same initial conditions for process ‘A’ so that we can compare the effect of behaviour on process ‘A’ of the feedback.
The above graph shows positive feedback. In this case the feedback has reduced the variation in magnitude giving a higher average output. You may also notice that the peaks and troughs from the feedback are not aligned with those that would have been produced by just process ‘A’. This is an important aspect of feedback in that it may not only affect the magnitude of the output but also the phase (relative frequency of the variations).
The second graph is showing negative feedback in that the feedback has caused the otherwise oscillating output to rapidly decay to a constant minimum.
Tuning feedback loops so that we get the required behaviour can be difficult and the graph below shows complex feedback that displays unpredictable patterns of positive and negative feedback.