mathias-brandewinder/Quipu

Quipu is a .NET library that implements the Nelder-Mead optimization algorithm for finding function minima or maxima without requiring derivatives. It provides both F\# and C\# APIs to define objective functions, configure tolerances, and specify starting points, while handling edge cases like unbounded objectives and invalid inputs. The system includes comprehensive unit tests and benchmarks to validate solver performance on standard test functions.

This entry marks the initial release of Quipu, a .NET library implementing the Nelder-Mead optimization method for finding function minima or maxima without requiring derivatives. The release includes the core library (Quipu.fsproj), a solution file (Quipu.sln) organizing the project structure, and an MIT license. It provides APIs for both F\# (pipeline-based) and C\# (fluent interface) to define objective functions, set tolerances, and specify starting points (e.g., Start.around), with examples demonstrating usage via NuGet package version 1.0.0.

(repo-wide) · high confidence

The Quipu library now includes a Nelder-Mead optimization algorithm, providing both an F\# pipe-forward interface and a C\# fluent interface for minimizing, maximizing, or goal-seeking functions. Users can configure the solver with custom tolerances, maximum iterations, and various starting point strategies (such as starting around a specific value or providing an explicit simplex). The implementation includes robust validation of the initial simplex to detect unbounded objectives or invalid inputs, and supports functions of varying dimensions through automatic vectorization.

src/Quipu · high confidence

The Quipu library has restructured its core API to improve clarity and maintainability. The \IObjective\ interface has been renamed to \IVectorFunction\, and the \Candidate\ type has been replaced by \Evaluation\ to better represent function evaluation results. Configuration is now centralized in a new \Configuration.fs\ module, introducing a \Configuration\ record that groups \UpdateParameters\ (alpha, gamma, rho, sigma) and \Tolerance\ settings. The \Solution\ type in \Core.fs\ now includes an \Iterations\ count and a \Status\ field (Optimal, Suboptimal, Unbounded) to provide more detailed feedback on the solver's outcome. Additionally, the \Simplex\ type has been extracted into its own file with a private constructor and explicit creation methods, and the \Vectorize\ utility class now supports converting functions of up to 3 arguments into \IVectorFunction\ instances.

Quipu · high confidence

Added C\# unit tests for the Nelder-Mead optimizer; Added Nelder-Mead solver benchmarks for Beale and Booth functions; Added comprehensive unit tests for the Nelder-Mead solver.

The Quipu Nelder-Mead function minimization solver is introduced with its core library targeting .NET 6.0 and pinned to FSharp.Core version 5. The test and benchmark infrastructure is established, with unit tests (both F\# and C\# variants) and benchmarks targeting .NET 8.0. Key testing dependencies include Expecto 10.2.1, xUnit 2.9.2, BenchmarkDotNet 0.14.0, and Microsoft.NET.Test.Sdk 17.11.1.

(dependencies) · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 58 → 60 (+1.9)
• Rubric changed (rubric-2026.09.15 → rubric-2026.10.1) — scores are not directly comparable.

• Code Health 98 → 98 (+0.0)
• Architecture 53 → 60 (+7.2)
• Maturity 57 → 57 (+0.0)
• Readiness 56 → 56 (+0.0)
• Security 75 → 70 (-5.0)
• Performance 80 (new)

• Documentation: no architecture or design documentation (README.md)
• Documentation: no installation or build instructions (README.md)

• Unchanged — 0 containers · 1 contexts · 0 edges

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

mathias-brandewinder/Quipu was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.