Changes of the Raku package "Math::NIntegrate".
0.0.10 2026-10-18
- Implemented Quasi Monte Carlo rules and strategies
- Both crude and adaptive
- Implemented Double Exponential rule (1D only)
0.0.9 2026-10-15
- Implemented Adaptive Monte Carlo integration strategy
- Performance improvements of Multidimensional rule
- ≈ 1.5 times speed up with Rakudo
- ≈ 3.5 times speed up with Rakupp / Raku++
0.0.8 2026-10-13
- Performance related implementations and refactoring using Infinite variable transform
- Not hooked-up yet in main integration process; only profiled and documented
- Unit cube rescaling strategy decorator implementation for numeric ranges (finite and infinite)
- For example: ,
- Can provide between 1.5 to 5 times speedup for both 1D and nD integrals
0.0.7 2026-10-10
- Performance related implementations and refactoring:
- ≈ 5 times speed up on multidimensional crude Monte Carlo integration
- ≈ 4.5 times speed up for Cartesian rules
- ≈ 2 times speed up for one-dimensional integrals
0.0.6 2026-10-07
- Performance related refactoring, ≈ 25% speed-up
- MonteCarlo strategy improvements: better error handling, processing partitioning and random seed options
- Implemented Romberg quadrature for TrapezoidalRule
- Grammar and interpreter enhancements to:
- Take a list of integers a value for the "partitioning" option
- Process "romberg-quadrature" option
0.0.5 2026-10-05
- Fixed a bug in the handling of integrals with range (-Inf, Inf)
- Implemented max-points stopping criteria for the Global-Adaptive and Local-Adaptive strategies
- Not working properly yet for MultiDimensionalRule
0.0.4 2026-10-05
- Streamlined computations with user defined integration rules and strategies
- Implemented (crude) Monte-Carlo strategy
- Integrals can have functional boundaries
- I.e. this works: nintegrate({$^x + $^y}, ['x', 0, 1], ['y', {$^x.sqrt}, {$^x}])
0.0.3 2026-10-02
- Process user specified "singular points" given in the range specs
- For example:
- nintegrate({1/(1-$^x).abs.sqrt}, )
- nintegrate({1/($^x + $^y).abs.sqrt}, , )
- Fixes for both concise- and multi-component Cartesian rule specifications
- Facilitation of stepping away from singular points during integrand evaluation
- Special representation methods -- gist, Hash, Str -- for strategies and rules
0.0.2 2026-10-01
- Implemented LocalAdaptive integration strategy
- Related region division implementation and grammar/interpreter extensions.
0.0.1 2026-09-28
- First version with:
- GlobalAdaptive integration strategy
- Can be given an integration monitor routine as an argument.
- The integration rules:
- Cartesian, ClenshawCurtis, GaussKronrod, MonteCarlo, MultiDimensional, NewtonCotes, Trapezoidal
- Singularity handler by Iri, Moriguti, & Takesawa ("IMT rule").
- Comprehensive integrator spec parsing and interpretation
- Kahaner integrator tests
- CLI script