Already five entries and we didn’t say a single word about feedback. Let’s just introduce the concept and some graphical signs, so that we may complete the blog entries with nice graphs representing what it has been said.
Feedback is the continuous interaction between two parts. Feedback exists between two parts when each affects the other1.
We represent feedback with influence (or causal) links.
A change in A produces a change in B in the SAME direction:
If A increases, B will increase. If A decreases, B will decrease.
A change in A produces a change in B in the OPPOSITE direction:
If A increases, B will decrease. If A decreases, B will increase.
A feedback loop can be positive or negative. If the number of negative links in the loop is an even number (including 0), the loop is positive. If the number of negative links in the loop is an odd number, the loop is negative.
You can also determine the polarity of a loop by tracing the effect of a change in one of the variables as it propagates around the loop. When the feedback effect reinforces the original change, then it’s a positive loop; when it opposes the original change, it’s a negative loop.
Positive feedback loops are self-reinforcing and they produce exponential growth. No real quantity can grow forever, so we assume we are considering only a part of a bigger system in a given period of time.
Negative feedback loops are self-correcting and they counteract change.
One last thing to consider in feedback loops is delay. We find delay when the effect (either positive or negative) between two parts is postponed in time. We represent the delay with the following arrow.
All the systems are a combination of positive and negative links, sometimes deferred with a certain delay.
1. W. Ross Ashby (1957). An introduction to cybernetics (PDF). Chapman & Hall, cited at Wikipedia ‘Feedback’ definition.