What is system boundary?

A system boundary marks what is inside a model and what is treated as outside, either fixed or ignored. Every model needs one, because nothing can include everything. Drawing it in the wrong place is a common cause of surprising results, since an excluded factor may be the one that pushes back.

How it works

A good boundary includes every feedback loop that drives the behavior being studied. Quantities inside the boundary are explained by the model, while quantities outside are inputs that the modeler sets. If something outside actually responds to what happens inside, the boundary is too tight, and the model will miss that response. If too much is included, the model gets hard to read and hard to check. A practical test is to ask what the model is meant to explain, then include the minimum needed to produce that behavior.

Boundaries can be widened step by step. Start with a small model that explains the main behavior, then add one outside factor at a time and check whether the results change meaningfully. If not, leave the factor outside.

Example

Say a city models its water supply with the reservoir level, household use, and rainfall. It treats the price of water as a fixed number outside the boundary. In reality higher use leads to higher prices, which lower use. The model, with price outside, will overstate how fast the reservoir drains in a dry year. Moving price inside the boundary adds a balancing loop and fixes the forecast. This is an illustration.

How to model it in Machinations

Draw only the Pools, Sources, Drains and Registers that belong inside the boundary. Treat outside factors as Levers or fixed Registers that you set by hand. If a Play shows behavior that looks wrong, ask whether an outside factor should respond, and bring it inside with a state connection.