NLoptNet

This is a C# wrapper around the NLopt C library. It includes both 32 and 64 bit DLLs. It will work with "Any CPU" projects, but this is not a "portable" project (yet). This particular wrapper is licensed with an MIT license, but the enclosed NLopt library carries an LGPL license. Thanks to ASI for sponsoring some time on this project.

Example:

[TestMethod]
public void FindParabolaMinimum()
{
	using (var solver = new NLoptSolver(NLoptAlgorithm.LN_COBYLA, 1, 0.001, 100))
	{
		solver.SetLowerBounds(new[] { -10.0 });
		solver.SetUpperBounds(new[] { 100.0 });
				
		solver.SetMinObjective(variables =>
		{
			return Math.Pow(variables[0] - 3.0, 2.0) + 4.0;
		});
		double? finalScore;
		var initialValue = new[] { 2.0 };
		var result = solver.Optimize(initialValue, out finalScore);

		Assert.AreEqual(NloptResult.XTOL_REACHED, result);
		Assert.AreEqual(3.0, initialValue[0], 0.1);
		Assert.AreEqual(4.0, finalScore.Value, 0.1);
	}
}

You can use the derivative-based optimizers like this:

solver.SetMinObjective((variables, gradient) =>
{
	if (gradient != null)
		gradient[0] = (variables[0] - 3.0) * 2.0;
	return Math.Pow(variables[0] - 3.0, 2.0) + 4.0;
});

And add constraints like this:

solver.AddEqualZeroConstraint((variables, gradient) =>
{
	if (gradient != null)
		gradient[0] = 1.0;
	return variables[0] - 100.0;
});

To use the Augmented Lagrangian:

using(var solver = new NLoptSolver(NLoptAlgorithm.LN_AUGLAG, 2, 0.001, 1000, NLoptAlgorithm.LN_SBPLX))
{
	solver.SetLowerBounds(new[] { 1.0, 1.0 });
	solver.SetUpperBounds(new[] { 200.0, 200.0 });
	solver.AddLessOrEqualZeroConstraint(Constrain);
	solver.SetMinObjective(Score);
	var result = solver.Optimize(best, out finalScore);
}

Contributions are welcome. Please read through the NLopt documentation before posting questions/issues here.

eriove/NLoptNet

NLoptNet