StapletonMathPackage is a C# numerical methods library providing tools for linear algebra, optimization, root-finding, interpolation, and ODE solving. It is pretty old, and I mostly started it as a wrapper around other math packages because I wanted to stick with double arrays in my program rather than specialized matrix classes.
This version wraps around Math.Numerics for many of the linear algebra operations.
using RandomMath;
double[,] A = { { 1, 2 }, { 3, 4 } };
double[,] B = { { 5, 6 }, { 7, 8 } };
double[,] C = MatrixMath.Multiply(A, B);
double detA = MatrixMath.Determinant(A);
Cubic Spline Interpolation
using RandomMath.CubicSplineFit;
var xs = new double[] { 0, 1, 2, 3 };
var ys = new double[] { 0, 1, 0, 1 };
var spline = new CubicSplineFit(xs, ys);
double dy = 0, d2y = 0;
double yInterp = spline.Interpolate(1.5, ref dy, ref d2y);
Genetic Algorithm
using RandomMath.GeneticAlgorithm;
public class SphereFunction : IOptimize {
public int nX => 3;
public double Eval(double[] X) => X.Sum(x => x*x);
public IOptimize DeepCopy() => new SphereFunction();
}
var ga = new GeneticAlgorithm(new SphereFunction());
ga.Run();
Console.WriteLine($"Best fitness: {ga.Best.Fitness}");
Newton-Raphson (Jacobian supplied)
using RandomMath.NewtonRaphson;
public class SimpleFunc : IMatrixFunctionAndDerivative {
public double[] Eval(double[] X) => new[] { X[0]*X[0] - 2 };
public double[,] DEval(double[] X) => new double[,] { { 2*X[0] } };
}
var solver = new NewtonRaphsonJacobian(
new double[] { 0 }, // target y
new double[] { 1.0 }, // initial guess
new SimpleFunc(),
1e-10, 50
);
solver.Solve();
Console.WriteLine($"Root ≈ {solver.X[0]}");
Matrix Exponential ODE Solve
using RandomMath.ODESystemSolver;
double[,] A = { { 0, 1 }, { -1, 0 } };
var mexp = new MatrixExponential(A, 1.0);
var expAx = mexp.Solve(1.0); // exp(A*1.0)
Documentation
All public APIs include XML comments visible in IntelliSense.
Example usage included in this README.
For advanced methods (ODE solvers, GA tuning), see source files in RandomMath.* namespaces.
Contributing
Contributions are welcome!
Please open an issue or submit a PR with improvements, bug fixes, or new numerical routines.
Guidelines:
Follow existing naming conventions.
Provide XML documentation for new public APIs.
Add unit tests where possible.
License
This project is licensed under the MIT License — see the LICENSE file for details.
