Nonlinear-Boundary Stress Test¶
Stress test — instantiates Coherent Linear Space Design
Pushes a linear model to the edges of its domain to find where superposition and scaling break, and registers those regions as off-limits.
Every linear model has a domain where its linearity is true and a beyond where it is a lie, and the dangerous work happens near the edge. The Nonlinear-Boundary Stress Test deliberately drives an assumed-linear model toward its extremes — large magnitudes, adversarial superpositions of many inputs, boundary and discrete-valued inputs — to locate the point where superposition and proportional scaling stop holding: saturation, thresholds, quantization, path-dependence. Its defining move is finding and recording the boundary rather than assuming there is none: it does not ask whether the model is linear near the calibration data (a linearity check answers that), it asks how far out the linearity extends and where it fails, then writes those failure regions into an exception register that all downstream linear reasoning must respect.
Example¶
A bridge analysis models deck deflection as a linear function of load, so that the effect of many vehicles is just the superposition of each vehicle's effect. That assumption is safe under service loads and catastrophic if extrapolated blindly, so the engineers stress-test the boundary. They scale up a single load well past service range and watch deflection stop tracking proportionally as the structure approaches its yield behavior. They superpose worst-case clusters of heavy loads and find that beyond a combined threshold the members begin to interact nonlinearly rather than adding independently. Each probe marks where the straight-line prediction and the true response diverge past tolerance. The output is an exception register: "single loads above X, or combined loads above Y, leave the linear regime — use the full nonlinear finite-element model there." The linear model keeps its speed and clarity inside the certified region, and the register fences off the region where it would quietly lie.[n1]
How it works¶
- Probe the extremes — increase magnitudes along each direction until the response departs from the linear prediction.
- Probe adversarial combinations — superpose many inputs, including worst-case clusters, to find where independent addition fails.
- Probe the boundaries — test near domain edges, thresholds, and discrete or quantized inputs where linearity is most fragile.
- Locate onset and register it — find where deviation first exceeds tolerance and write the failing region into an exception register that downstream linear reasoning must honor.
Tuning parameters¶
- Probe extremity and step schedule — how far and how fast to push. Aggressive steps find the boundary quickly but may overshoot its precise location.
- Combination adversariality — random superpositions versus deliberately worst-case ones. Worst-case probing finds nearer boundaries but takes more effort.
- Deviation threshold — how much departure from the linear prediction counts as "out of bounds." Tighter thresholds shrink the certified linear region.
- Register granularity — a single safe range versus a detailed map of distinct exception regions. Finer maps are more useful and more expensive to build.
When it helps, and when it misleads¶
Its strength is preventing the most common and costly error with linear models — extrapolating them past the regime where they are valid — by producing an explicit, conservative fence around the safe region.
Its failure mode is that it only finds the boundaries it actually probes: an unexamined direction or combination can hide a nearer nonlinearity than any it located, so the certified region can be overstated. Stopping the probes too early has the same effect. The classic misuse is the opposite error it is meant to guard against — extrapolating a linear calibration far beyond its measured range because "it was linear everywhere we looked," exactly the trap that Hooke's law illustrates: stress is proportional to strain only up to the proportional limit, and beyond it the material yields and the linear relation collapses.[n1] The guarding discipline is to probe multiple directions and adversarial combinations, and to keep the exception register conservative — when in doubt, fence more out, not less.
How it implements the components¶
The stress test fills the archetype's boundary face — it marks where linear reasoning must stop:
nonlinear_or_discrete_exception_register— its core output: the recorded regions (saturation, thresholds, discrete inputs) where linear operations are no longer valid.closure_axiom_checklist— it probes exactly where closure and superposition fail, mapping the frontier of the region in which the linear axioms actually hold.
It does not declare the carrier or operations (the Vector-Space Specification Sheet), verify the axioms across the interior (the Linear-Axiom Verification Checklist), fix a basis (the Basis & Coordinate Table), test a target's reachability (the Linear-Combination Membership Test), or interpret coordinate semantics (the Change-of-Basis Review).
Related¶
- Instantiates: Coherent Linear Space Design — the stress test supplies the exception boundaries that keep linear reasoning honest at the edges.
- Consumes: Vector-Space Specification Sheet — it stress-tests the boundary of exactly the linear space the sheet declares.
- Sibling mechanisms: Vector-Space Specification Sheet · Linear-Axiom Verification Checklist · Basis & Coordinate Table · Linear-Combination Membership Test · Change-of-Basis Review · Zero-Span Linearity Check · Linear Embedding Diagnostics
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Nonlinear-Boundary Stress Test operates as an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation because it pushes a linear model to the edges of its domain to find where superposition and scaling break, and registers those regions as off-limits.
Independent corroboration: The frozen evidence defines Nonlinear-Boundary Stress Test as 'Pushes a linear model to the edges of its domain to find where superposition and scaling break, and registers those regions as off-limits', so its operative form is Experiment, Test & Rehearsal.
Nearest alternative: Assessment, Review & Assurance — Nonlinear-Boundary Stress Test includes features of a bounded evaluation of existing evidence or work that produces a finding or disposition, but its defining operation is an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Engineering & Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Engineering qualification turns those breakdown points into operating envelopes and prohibited extrapolation regions.
Related originating lineages:
- Mathematics — Testing where superposition and scalar homogeneity fail interrogates the defining axioms of a linear mathematical model.
- Physics — Experimental physics established testing a linear law toward regime boundaries, such as the proportional limit beyond which Hooke's law fails.
- Statistics & Experimental Design — Model diagnostics contribute systematic residual checks at domain boundaries.
Review resolution: Authoritative-source research resolves the primary-origin disagreement. Driving a physical or structural model to the limit of its validated linear regime is an engineering stress-test practice grounded in physical nonlinearity and mathematical superposition. Origin breadth is limited to formative lineages; present-day applicability is recorded separately as domain_reach=multi_domain.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
[n1] Hooke's law states that the deformation of an elastic body is proportional to the applied load — but only up to the proportional limit. Beyond it the material yields and the stress-strain relation becomes nonlinear. It is the textbook case of a linear law that is exact within its regime and dangerous when extrapolated past the boundary this stress test is built to find. ↩a ↩b