fr.cnes.sirius.patrius.math.optim.joptimizer.optimizers

Class LPStandardConverter

java.lang.Object

fr.cnes.sirius.patrius.math.optim.joptimizer.optimizers.LPStandardConverter

public class LPStandardConverter
extends Object

Converts a general LP problem stated in the form (1):
min(c) s.t.
G.x < h
A.x = b
lb <= x <= ub

to the (strictly)standard form (2)
min(c) s.t.
A.x = b
x >= 0

or to the (quasi)standard form (3)
min(c) s.t.
A.x = b
lb <= x <= ub

Setting the field strictlyStandardForm to true, the conversion is in the (strictly) standard form (2).
Note 1: (3) it is not exactly the standard LP form (2) because of the more general lower and upper bounds terms.
Note 2: if the vector lb is not passed in, all the lower bounds are assumed to be equal to the value of the field unboundedLBValue
Note 3: if the vector ub is not passed in, all the upper bounds are assumed to be equal to the value of the field unboundedUBValue
Note 4: unboundedLBValue is the distinctive value of an unbounded lower bound. It must be one of the values:

1. Double.NaN (the default)
2. Double.NEGATIVE_INFINITY

Note 5: unboundedUBValue is the distinctive value of an unbounded upper bound. It must be one of the values:

1. Double.NaN (the default)
2. Double.POSITIVE_INFINITY

See Converting LPs to standard form, "Convex Optimization", p 147

Since:

4.6

Author:

alberto trivellato (alberto.trivellato@gmail.com)

Field Summary

Fields

Modifier and Type	Field and Description
static double	DEFAULT_UNBOUNDED_LOWER_BOUND Default unbounded lower bound
static double	DEFAULT_UNBOUNDED_UPPER_BOUND Default unbounded upper bound

Constructor Summary

Constructors

Constructor and Description
LPStandardConverter() Constructor, set the strictlyStandardForm to false
LPStandardConverter(boolean strictlySF) Constructor, set the default unbounded lower and upper bounds
LPStandardConverter(boolean strictlySF, double unboundedLBVal, double unboundedUBVal) Constructor
LPStandardConverter(double unboundedLBVal, double unboundedUBVal) Constructor

Method Summary

Modifier and Type	Method and Description
int	getOriginalN() Get the original number of variables
RealMatrix	getStandardA() Get the final equalities constraints coefficients
RealVector	getStandardB() Get the final equalities constraints limits
RealVector	getStandardC() Get the final objective function
double[]	getStandardComponents(double[] x) Express a vector in the original variables in the final standard variable form
RealVector	getStandardLB() Get the final lower bounds This makes sense only if strictlyStandardForm = false (otherwise all lb are 0).
int	getStandardN() Get the final number of variables
int	getStandardS() Get the final number of slack variables for inequalities constraints
RealVector	getStandardUB() Get the final upper bounds This makes sense only if strictlyStandardForm = false (otherwise all ub are unbounded).
double	getUnboundedLBValue() Get the unbounded lower bound value
double	getUnboundedUBValue() Get the unbounded upper bound value
boolean	isLbUnbounded(Double lb) Is the lower bound unbounded?
boolean	isUbUnbounded(Double ub) Is the upper bound unbounded?
double[]	postConvert(double[] x) Get back the vector in the original components.
void	toStandardForm(double[] originalC, double[][] originalG, double[] originalH, double[][] originalA, double[] originalB, double[] originalLB, double[] originalUB) Transforms the problem from a general form to the (quasi) standard LP form.
void	toStandardForm(RealVector originalC, RealMatrix originalG, RealVector originalH, RealMatrix originalA, RealVector originalB, RealVector originalLB, RealVector originalUB) Transforms the problem from a general form to the (quasi) standard LP form.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

DEFAULT_UNBOUNDED_LOWER_BOUND

public static final double DEFAULT_UNBOUNDED_LOWER_BOUND

Default unbounded lower bound

See Also:

Constant Field Values

DEFAULT_UNBOUNDED_UPPER_BOUND

public static final double DEFAULT_UNBOUNDED_UPPER_BOUND

Default unbounded upper bound

See Also:

Constant Field Values

Constructor Detail

LPStandardConverter

public LPStandardConverter()

Constructor, set the strictlyStandardForm to false

LPStandardConverter

public LPStandardConverter(boolean strictlySF)

Constructor, set the default unbounded lower and upper bounds

Parameters:

strictlySF - true/false

LPStandardConverter

public LPStandardConverter(double unboundedLBVal,
                           double unboundedUBVal)

Constructor

Parameters:

unboundedLBVal - unbounded lower bound value

unboundedUBVal - unbounded upper bound value

LPStandardConverter

public LPStandardConverter(boolean strictlySF,
                           double unboundedLBVal,
                           double unboundedUBVal)

Constructor

Parameters:

strictlySF - true/false

unboundedLBVal - unbounded lower bound value

unboundedUBVal - unbounded upper bound value

Method Detail

toStandardForm

public void toStandardForm(double[] originalC,
                           double[][] originalG,
                           double[] originalH,
                           double[][] originalA,
                           double[] originalB,
                           double[] originalLB,
                           double[] originalUB)

Transforms the problem from a general form to the (quasi) standard LP form.

Parameters:

originalC - original C array

originalG - original G matrix

originalH - original H array

originalA - original A matrix

originalB - original B array

originalLB - if null, all lower bounds default to this.unspecifiedLBValue

originalUB - if null, all upper bounds default to this.unspecifiedUBValue See Converting LPs to standard form, "Convex Optimization", p 147

toStandardForm

public void toStandardForm(RealVector originalC,
                           RealMatrix originalG,
                           RealVector originalH,
                           RealMatrix originalA,
                           RealVector originalB,
                           RealVector originalLB,
                           RealVector originalUB)

Transforms the problem from a general form to the (quasi) standard LP form.

Parameters:

originalC - original C array

originalG - original G matrix

originalH - original H array

originalA - original A matrix

originalB - original B array

originalLB - if null, all lower bounds default to this.unspecifiedLBValue

originalUB - if null, all upper bounds default to this.unspecifiedUBValue See Converting LPs to standard form, "Convex Optimization", p 147

postConvert

public double[] postConvert(double[] x)

Get back the vector in the original components.

Parameters:

x - vector in the standard variables

Returns:

the original component

getStandardComponents

public double[] getStandardComponents(double[] x)

Express a vector in the original variables in the final standard variable form

Parameters:

x - vector in the original variables

Returns:

vector in the standard variables

getOriginalN

public int getOriginalN()

Get the original number of variables

Returns:

num variables

getStandardN

public int getStandardN()

Get the final number of variables

Returns:

num variables

getStandardS

public int getStandardS()

Get the final number of slack variables for inequalities constraints

Returns:

num variables

getStandardC

public RealVector getStandardC()

Get the final objective function

Returns:

objective function

getStandardA

public RealMatrix getStandardA()

Get the final equalities constraints coefficients

Returns:

coefficients

getStandardB

public RealVector getStandardB()

Get the final equalities constraints limits

Returns:

limits

getStandardLB

public RealVector getStandardLB()

Get the final lower bounds This makes sense only if strictlyStandardForm = false (otherwise all lb are 0).

Returns:

lower bounds

getStandardUB

public RealVector getStandardUB()

Get the final upper bounds This makes sense only if strictlyStandardForm = false (otherwise all ub are unbounded).

Returns:

upper bounds

isLbUnbounded

public boolean isLbUnbounded(Double lb)

Is the lower bound unbounded?

Parameters:

lb - lower bound

Returns:

true/false

isUbUnbounded

public boolean isUbUnbounded(Double ub)

Is the upper bound unbounded?

Parameters:

ub - upper bound

Returns:

true/false

getUnboundedLBValue

public double getUnboundedLBValue()

Get the unbounded lower bound value

Returns:

value

getUnboundedUBValue

public double getUnboundedUBValue()

Get the unbounded upper bound value

Returns:

value