tech.nmfin.portfoliooptimization.markowitz.constraints

Class QPWeightsLimit

java.lang.Object

tech.nmfin.portfoliooptimization.markowitz.constraints.QPWeightsLimit

All Implemented Interfaces:

QPConstraint

public class QPWeightsLimit
extends Object
implements QPConstraint

One bound per asset

minW and maxW must have the same asset order and dimension as the optimizer's return vector. They encode componentwise lower and upper bounds; a negative lower bound allows short selling. This object supplies no budget equality. MarkowitzByQP adds the sum-to-one constraint itself.

For a fully invested portfolio, each lower bound must be no greater than its upper bound, the sum of lower bounds must be at most one, and the sum of upper bounds must be at least one. These checks are necessary; additional constraints can still make the problem infeasible. Check bounds before solving rather than relying on late solver failures to validate input data.

Constructor Summary

Constructors

Constructor and Description
QPWeightsLimit(Vector minW, Vector maxW)

Method Summary

Modifier and Type	Method and Description
LinearEqualityConstraints	linearEqualityConstraints()
LinearGreaterThanConstraints	linearGreaterThanConstraints()
LinearLessThanConstraints	linearLessThanConstraints()

Methods inherited from class java.lang.Object

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

Constructor Detail

QPWeightsLimit

public QPWeightsLimit(Vector minW,
                      Vector maxW)

Method Detail

linearGreaterThanConstraints

public LinearGreaterThanConstraints linearGreaterThanConstraints()

Specified by:

linearGreaterThanConstraints in interface QPConstraint

linearLessThanConstraints

public LinearLessThanConstraints linearLessThanConstraints()

Specified by:

linearLessThanConstraints in interface QPConstraint

linearEqualityConstraints

public LinearEqualityConstraints linearEqualityConstraints()

Specified by:

linearEqualityConstraints in interface QPConstraint