A rigorous interval arithmetic library for .NET
This section will give some examples on how to use IntSharp. For a quick introduction to interval arithmetic, take a look at the corresponding Wikipedia article. For further, advanced functions of IntSharp, it is recommended to study the corresponding unit tests. They cover almost every aspect of the implementation and give insights on how to use specific features.
- Create intervals with one of the several factory methods and calculate with them:
var i1 = Interval.FromInfSup(3.2,7); // [3.2 , 7]
var i2 = Interval.FromMidRad(4.8,2.2); // [2.6 , 7]
var i3 = Interval.FromDoublePrecise(9.34); // [9.34 , 9.34]
var res = (i1 * i2 + i3) / 2;
- Make verified calculations with included error propagation. No approximations, no derivations.
// Init interval with MidRad factory
var height = Interval.FromMidRad(3.8, 0.2);
// Init interval with InfSup factory
var accelerationOfGravity = Interval.FromInfSup(9.8, 9.82);
// Calculate the verified result
var velocity = Math.Sqrt(2 * height * accelerationOfGravity);
// = [8.4 , 8.863]
- Solve determined or overdetermined linear equation systems and obtain verified results:
// Init items for the matrix A.
var aItems = new[,]
{
{Interval.FromMidRad(2,0.2), Interval.FromDoublePrecise(1)},
{Interval.FromMidRad(7,0.2), Interval.FromDoublePrecise(1)}
};
// Init items for the rhs vector b.
var bItems = new[]
{
Interval.FromMidRad(4,0.2),
Interval.FromMidRad(1,0.2)
};
// Init matrix A and vector b.
var a = new IntervalMatrix(aItems);
var b = new IntervalVector(bItems);
// Solve the system with or withought error weighing:
var result = LinearEquationSystem.Solve(a, b);
Find verified enclosures of a function's root:
private static Interval MyTestFunction(Interval x)
{
return 2 * Math.Pown(x, 3) - 9;
}
private static Interval MyTestDerivative(Interval x)
{
return 6 * Math.Pown(x, 2);
}
public void FindTestRoot()
{
var myTestRange = Interval.FromInfSup(0, 100);
var root = RootFinding.FindRoot(
MyTestFunction,
MyTestDerivative,
myTestRange);
}
Quickstart for Mono (Linux)
- Clone or download IntSharp and change to it's root directory
git clone https://github.com/selmaohneh/IntSharp.git
cd /path/to/IntSharp
- Install the required packages mono-complete and nuget from your distribution package manager.
apt-get install mono-complete nuget
- Use the NuGet package mangager to obtain the remaining dependencies MathNet.Numerics, NUnit, and NUnit.Console:
mono /path/to/NuGet.exe update -self
mono /path/to/NuGet.exe install NUnit -OutputDirectory packages -Version 3.7.1
mono /path/to/NuGet.exe install NUnit.Console -OutputDirectory packages -Version 3.6.1
mono /path/to/NuGet.exe install MathNet.Numerics -OutputDirectory packages -Version 3.19.0
- Build the project:
xbuild IntSharp.sln
- Finally, run the tests:
mono packages/NUnit.ConsoleRunner.3.6.1/tools/nunit3-console.exe \
IntSharpTests/bin/Debug/IntSharpTests.dll --noresult
NUnit Console Runner 3.6.1
Copyright (C) 2017 Charlie Poole
Runtime Environment
OS Version: Linux 4.4.74.18
CLR Version: 4.0.30319.42000
Test Files
IntSharpTests/bin/Debug/IntSharpTests.dll
Run Settings
DisposeRunners: True
WorkDirectory: /workspace/IntSharp
ImageRuntimeVersion: 4.0.30319
ImageTargetFrameworkName: .NETFramework,Version=v4.6.1
ImageRequiresX86: False
ImageRequiresDefaultAppDomainAssemblyResolver: False
NumberOfTestWorkers: 4
Test Run Summary
Overall result: Passed
Test Count: 93, Passed: 93, Failed: 0, Warnings: 0, Inconclusive: 0, Skipped: 0
Start time: 2017-07-09 21:31:31Z
End time: 2017-07-09 21:31:32Z
Duration: 1.041 seconds
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent