Class GoldsteinLineSearch

Hierarchy

LineSearch
- GoldsteinLineSearch

Implements

QuaternaryFunction<Problem, byRef, EndCriteria, Real, Real>

Index

Constructors

constructor

Properties

Accessors

searchDirection

Methods

Constructors

constructor

new GoldsteinLineSearch(eps?: Real, alpha?: Real, beta?: Real, extrapolation?: Real): GoldsteinLineSearch

Overrides LineSearch.constructor
- Defined in ql/math/optimization/goldstein.ts:14
Parameters
- Default value eps: Real = 1e-8
- Default value alpha: Real = 0.05
- Default value beta: Real = 0.65
- Default value extrapolation: Real = 1.5
Returns GoldsteinLineSearch

Properties

Private _alpha

_alpha: Real

Private _beta

_beta: Real

Private _extrapolation

_extrapolation: Real

Protected _gradient

_gradient: Real[]

Protected _qpt

_qpt: Real

Protected _qt

_qt: Real

Protected _searchDirection

_searchDirection: Real[]

Protected _succeed

_succeed: boolean

Protected _xtd

_xtd: Real[]

Accessors

searchDirection

get searchDirection(): Real[]
set searchDirection(v: Real[]): void

Inherited from LineSearch.searchDirection
- Defined in ql/math/optimization/linesearch.ts:48
current value of the search direction

Returns Real[]
Inherited from LineSearch.searchDirection
- Defined in ql/math/optimization/linesearch.ts:52
current value of the search direction

Parameters
- v: Real[]
Returns void

Methods

f

f(P: Problem, ecType: byRef, ec: EndCriteria, init: Real): Real

Overrides LineSearch.f
- Defined in ql/math/optimization/goldstein.ts:25
operator

()

Parameters
- P: Problem
- ecType: byRef
- ec: EndCriteria
- init: Real
Returns Real

lastFunctionValue

lastFunctionValue(): Real

Inherited from LineSearch.lastFunctionValue
- Defined in ql/math/optimization/linesearch.ts:29
return last cost function value

Returns Real

lastGradient

lastGradient(): Real[]

Inherited from LineSearch.lastGradient
- Defined in ql/math/optimization/linesearch.ts:34
return last gradient

Returns Real[]

lastGradientNorm2

lastGradientNorm2(): Real

Inherited from LineSearch.lastGradientNorm2
- Defined in ql/math/optimization/linesearch.ts:39
return square norm of last gradient

Returns Real

lastX

lastX(): Real[]

Inherited from LineSearch.lastX
- Defined in ql/math/optimization/linesearch.ts:24
return last x value

Returns Real[]

succeed

succeed(): boolean

Inherited from LineSearch.succeed
- Defined in ql/math/optimization/linesearch.ts:43
Returns boolean

update

update(params: Real[], direction: Real[], beta: Real, constraint: Constraint): Real

Inherited from LineSearch.update
- Defined in ql/math/optimization/linesearch.ts:64
Parameters
- params: Real[]
- direction: Real[]
- beta: Real
- constraint: Constraint
Returns Real