Seedlabs

Brunn–Minkowski Invariant Test Suite for Nonlinear PDE Solvers

An automated verification-and-validation regression suite that stress-tests nonlinear elliptic (p-Laplace-type) solvers by checking that numerically computed capacities of blended convex geometries satisfy a newly proven exact inequality — any violation flags mesh or convergence error.

MathematicsHistory and Theory of Mathematics
Urban Stormwater and Fluid ManagementEngineering simulation software — automated V&V and regression testing of nonlinear PDE solvers.

Concept

A CI-style V&V product for simulation software vendors and industrial simulation teams. The suite ships a library of convex body pairs and their Minkowski blends; it runs the solver's capacity computation on E1, E2, and λE1+(1−λ)E2, then checks the exact invariant

[Cap_A(λE1+(1−λ)E2)]^{1/(n−p)} ≥ λ[Cap_A(E1)]^{1/(n−p)} + (1−λ)[Cap_A(E2)]^{1/(n−p)}

within a certified discretization tolerance. Any violation flags mesh under-resolution or silent convergence failure — failure modes common in degenerate/singular p-Laplace problems that manufactured-solution tests often miss, because this invariant tests the solver's response to genuine geometry changes across a whole family of cases rather than a single analytic solution. The single wedge: a mathematically exact geometric invariant used as an automated solver-QA oracle.

Why now

Paper [0] proves the Brunn–Minkowski inequality for nonlinear A-capacity, supplying a new, mathematically exact invariant that any converged numerical p-Laplace-type solver must satisfy. Before this result, exact geometric invariants of this kind were available only in the linear (Newtonian) regime, so nonlinear solver QA lacked a geometry-family oracle; the paper hands QA teams a ready-made, checkable inequality spanning an entire family of test geometries.

Written by GLM · 2 Oct 2026

AI assessment

Backed by 1 paper85

A highly specialized V&V tool that leverages a new mathematical invariant to detect silent failures in nonlinear PDE solvers, offering a sharp wedge for high-end CAE vendors.

Evidence strength
5/5
The idea is directly derived from a specific, proven inequality in a peer-reviewed mathematical memoir, providing a concrete formula for the oracle.
Market pull
3/5
The target buyers (CAE vendors) have a clear need for V&V, but the market is small and the specific p-Laplace use case is a niche within the broader simulation industry.
Novelty & moat
4/5
While other catalog ideas use the same inequality for screening/prediction, this is the first to apply it as a regression-testing oracle for solver correctness.
Feasibility
4/5
The MVP is straightforward: a library of geometries and a script to check the inequality against solver outputs.
Wedge clarity
5/5
The wedge is exceptionally sharp: a single, mathematically exact invariant used as a QA oracle for a specific class of solvers.
Simplicity / focus
5/5
The product is focused on one single function—automated verification via a specific invariant—avoiding platform bloat.

Scored by Gemma 4 · 2 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.

Persona discussion

AI personas trained on real people's expertise debate this idea as it evolves.

View the discussion →

Act on this idea

Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.

Who benefits

  • Ansyscompany

    Runs large nonlinear solver regression suites; an exact-geometry invariant family would catch silent convergence failures in p-Laplace-type problems that current tests miss.

  • COMSOLcompany

    Could embed the invariant checks as built-in solver verification, strengthening credibility of its nonlinear PDE modules.

  • MathWorkscompany

    Its PDE Toolbox serves nonlinear elliptic problems; the suite provides objective regression tests for solver releases.

  • NAFEMSorganization

    Publishes engineering-simulation V&V benchmarks; an exact-geometry-invariant test family is a natural addition to its benchmark library for the nonlinear solver community.

  • Dassault Systèmescompany

    Its SIMULIA nonlinear solvers would benefit from third-party-verifiable invariant benchmarks for QA and marketing credibility.

Research it builds on

  1. Memoirs of the American Mathematical Society
    Akman, Murat, Gong, Jasun, Hineman, Jay et al. · 2026 · 662 citations
    All ideas from this paper →
Leave feedback
feasibility