What is balancing feedback loop?

A balancing feedback loop compares the current state of a system with a goal and acts to close the gap. A thermostat is the familiar example. Balancing loops explain stability, but also why policies meet resistance, since the system works against the change you are trying to make.

How it works

The loop has four parts: a goal, a current level, a gap between them, and a correcting action that grows with the gap. The larger the gap, the stronger the correction, so the system approaches the goal quickly at first and then slowly. If the correction is delayed, the system can overshoot and oscillate around the goal instead of settling. Many real systems contain both kinds of loops, and which one dominates decides whether the system grows, stabilizes, or declines.

Balancing loops explain why many well-meant interventions seem to have no lasting effect. The system treats the intervention as a disturbance and corrects for it, so lasting change usually needs a change of the goal or of the loop itself.

Example

Say a store wants to hold 500 units of stock and orders 20 percent of the gap each week. With 100 units on hand, the gap is 400, so it orders 80 and holds 180. Next week the gap is 320, so it orders 64 and holds 244. The following week it orders about 51. Orders shrink as the shelf fills, and stock moves smoothly toward 500 without ever overshooting. This is an illustration with round numbers.

How to model it in Machinations

Model the level as a Pool and the target as a Register. Compute the gap in another Register, then use a state connection to set the rate of an inflow from a Source in proportion to that gap. Play the diagram and chart the Pool to see it approach the target. Change the correction percentage with a Lever to see faster or slower settling.