Breakpoint analysis identifies values of a variable at which a change occurs, e.g. a management option fails, or the preferred solution changes.

When to use

While breakpoint analysis is a general concept, it can specifically be used for decision making in the face of multiple plausible futures to perform vulnerability analysis (which is itself a form of stress testing).

Breakpoint analysis needs to be evaluate success/failure while varying values of a variable, and therefore requires a system model where the relevant variables can be modified by an external algorithm.

How

A range of algorithms are possible depending on the number of variables and the nature of the boundary between success and failure. In general, the boundary forms a lower dimensional "manifold" in model scenario space, such that even with only two variables there are an infinite number of possible breakpoints. Selecting interesting breakpoints is therefore a key issue.

Specific approaches implementing breakpoint analysis include:

Selected examples

  • In a simple flood demonstration problem answering the question "Will regular flooding of ecological assets occur?", breakpoint analysis is performed using POMORE (a variant of MORE), and through optimisation-based hypothesis testing
  • Resources