RombergIntegrator<T> Class

public sealed class RombergIntegrator<T> : NumericalIntegrator<T>
where T : Object, INumber<T>

Public NotInheritable Class RombergIntegrator(Of T As {Object, INumber(Of T)})
	Inherits NumericalIntegrator(Of T)

generic<typename T>
where T : Object, INumber<T>
public ref class RombergIntegrator sealed : public NumericalIntegrator<T>

[<SealedAttribute>]
type RombergIntegrator<'T when 'T : Object and INumber<'T>> = 
    class
        inherit NumericalIntegrator<'T>
    end

Inheritance

Object  →  ManagedIterativeAlgorithm<T, T, SolutionReport<T, T>>  →  IterativeAlgorithm<T>  →  NumericalIntegrator<T>  →  RombergIntegrator<T>

Type Parameters

T

Romberg integration is an enhancement of the basic trapezoid rule. The number of points is doubled in each iteration. The successive approximations are then used to extrapolate the exact value of the integral as the width of each subinterval approaches zero. Romberg integration has an effective order of 2k after k steps of the algorithm. Even though the number of function evaluations is still fairly high, it is a major improvement over Simpson's rule, and a reasonable choice for integrands that are sufficiently smooth.

Reference