My original education was as an electronics/automation engineer (back when electronics were still largely analog, but that's another story...).
I was trained very early on in the use of control loops.
When automating a process, the common approach is to set up a closed loop with PID (Proportional, Integral, Derivative) control.
The origin of this control technique seems to date back to the 17th century, attributed to Christian Huygens—notably for adjusting the distance between millstones according to the flow of grain to be ground...
If you would like to learn more about this subject, this Wikipedia page is very informative: Proportional–integral–derivative controller - Wikipedia.
The principle of a PID loop is to establish an automatic control system to maintain a variable (temperature, speed, position, pressure, etc.) as close as possible to a desired value called the setpoint.
This is a self-correcting mechanism that is used in all kinds of modern equipment, from our thermostats to drone stabilization and industrial automation control systems.
This control principle incorporates three components: reacting quickly to deviations (P), while eliminating persistent errors (I) and stabilizing behavior (D). It involves reacting quickly but in a controlled manner to filter out noise and avoid unwanted oscillations.
In the field of flow management, the Kanban loop has become standard. Its principle is also self-regulation, which synchronizes manufacturing at an upstream station with the setpoint that is the consumption of the downstream station.
As with a PID loop, the principle is to regulate a system to avoid deviations and stabilize flow.
Both mechanisms react according to the actual state, rather than a pre-established plan, and they ensure automatic correction.
In the Demand Driven methodology, there are two self-regulating mechanisms that work in a similar way.
A DDMRP stock buffer is an evolution of the kanban loop. While in the kanban loop the size of the loop (the number of cards) is fixed, the DDMRP loop automatically adjusts over time and according to market demand. Compared to a kanban loop, a DDMRP buffer incorporates an I (integral) component, which consists of gradually adjusting the thresholds based on the calculated effective daily usage.
In terms of capacity, DBR constraint management controls the throughput of the entire system and manufacturing releases based on the effective capacity of the constraints in the system. This is another self-regulating loop.
The advantage of these self-regulating loops is clear: they automate the pace of operations based on actual market demand, they adapt automatically as things progress, and the role of planners becomes one of monitoring behavior and adjusting control parameters.
Curiously, most companies manage their flows in an open loop, instead of establishing high-frequency self-regulating closed loops.
Sales forecasts are established based on historical data and commercial aspirations.
Net requirements are calculated to meet these projected demands—often without taking constraints into account.
Orders are placed and production is launched.
Things are delivered on time or not, they are manufactured on time or not.
We measure deviations at regular intervals (usually monthly, at best weekly) – forecast accuracy, supplier on time delivery, adherence to the production plan.
We blame each other and start all over again...
Instead of open loops, opt for self-regulating closed loops. This approach dates to the 17th century, is proven, and has been greatly modernized in Intuiflow...