Definition

Take interconnected systems with jointly admissible behaviour 𝓑_G, and a set of variables y an outside system can reach. The effective boundary constraint is the projection of 𝓑_G onto y, the external behaviour in Willems’s sense. It is the relation the composite imposes on whatever interacts with it only through y (CLAIM-480). It is classed on two independent axes:

  • by memory. It is approximately instantaneous when the interior relaxes fast relative to the boundary: the two-spring chain gives the coupling k₁k₂/(k₁ + k₂) between its ends. Otherwise eliminating the interior leaves a memory kernel, a relation over boundary trajectories. It is a constraint either way.
  • by kind. It is of a new kind when the reduced relation cannot be stated in the parts’ relational vocabulary (TERM-091). Hansen and Ghrist’s star reduces to v₁ = v₂ + v₃, which no pairwise sheaf on the boundary expresses (THEORY-1090). That reduction is static, so a constraint of a new kind need not have memory.

The distinctions it draws

  • A constraint with or without memory. A history-dependent relation is a constraint on admissible trajectories. Separation of time scales decides only whether it is approximately memoryless (CLAIM-480, version 2).
  • Effective, not fundamental. It is realized by the interior and changes when the interior does. It is not a new law (A23 §XII).
  • Interface, not individual. A composite with a well-defined boundary constraint need not be an individual in any of TERM-088’s senses.

Why it is proposed

Four entries filed this evening use the notion, under several names: “effective boundary relation”, “boundary law”, “external behaviour”. It is the clearest place where the record’s general frame, that constraints are realized by organization (CLAIM-477), has a demonstrated case. What would settle it: adoption if the claims cite it, as CLAIM-480 and CLAIM-474 now do. The object is Willems’s external behaviour (LIT-1660, unread); the term is filed for the distinctions above, which that name does not carry, and would be rejected if a reading of Willems showed they are his.