Options
All
  • Public
  • Public/Protected
  • All
Menu

Class EndCriteria

Criteria to end optimization process:

  • maximum number of iterations AND minimum number of iterations around stationary point
  • x (independent variable) stationary point
  • y=f(x) (dependent variable) stationary point
  • stationary gradient

Hierarchy

  • EndCriteria

Implements

Index

Enumerations

Interfaces

Constructors

Properties

Methods

Constructors

constructor

  • new EndCriteria(maxIterations: Size, maxStationaryStateIterations: Size, rootEpsilon: Real, functionEpsilon: Real, gradientNormEpsilon: Real): EndCriteria

  • Parameters

    • maxIterations: Size
    • maxStationaryStateIterations: Size
    • rootEpsilon: Real
    • functionEpsilon: Real
    • gradientNormEpsilon: Real

    Returns EndCriteria

Properties

Protected _functionEpsilon

_functionEpsilon: Real

Protected _gradientNormEpsilon

_gradientNormEpsilon: Real

Protected _maxIterations

_maxIterations: Size

Maximum number of iterations

Protected _maxStationaryStateIterations

_maxStationaryStateIterations: Size

Maximun number of iterations in stationary state

Protected _rootEpsilon

_rootEpsilon: Real

root, function and gradient epsilons

Methods

checkMaxIterations

  • checkMaxIterations(iteration: Size, ref: byRef): boolean

  • Test if the number of iteration is below MaxIterations

    Parameters

    Returns boolean

checkStationaryFunctionAccuracy

  • checkStationaryFunctionAccuracy(f: Real, positiveOptimization: boolean, ref: byRef): boolean

  • Test if the function value is below functionEpsilon

    Parameters

    • f: Real
    • positiveOptimization: boolean
    • ref: byRef

    Returns boolean

checkStationaryFunctionValue

  • checkStationaryFunctionValue(fxOld: Real, fxNew: Real, ref: byRef): boolean

  • Test if the function variation is below functionEpsilon

    Parameters

    Returns boolean

checkStationaryPoint

  • checkStationaryPoint(xOld: Real, xNew: Real, ref: byRef): boolean

  • Test if the root variation is below rootEpsilon

    Parameters

    Returns boolean

checkZeroGradientNorm

  • checkZeroGradientNorm(gradientNorm: Real, ref: byRef): boolean

  • Test if the gradient norm value is below gradientNormEpsilon

    Parameters

    Returns boolean

f

  • f(iteration: Size, positiveOptimization: boolean, fold: Real, normgold: Real, fnew: Real, normgnew: Real, ref: byRef): boolean
  • operator

    () Test if the number of iterations is not too big and if a minimum point is not reached

    Parameters

    Returns boolean

functionEpsilon

  • functionEpsilon(): Real

  • Returns Real

gradientNormEpsilon

  • gradientNormEpsilon(): Real

  • Returns Real

maxIterations

  • maxIterations(): Size

  • Returns Size

maxStationaryStateIterations

  • maxStationaryStateIterations(): Size

  • Returns Size

rootEpsilon

  • rootEpsilon(): Real

  • Returns Real