org.akutan.optimization
Class UnconstrainedSolver

java.lang.Object
  org.akutan.optimization.UnconstrainedSolver

All Implemented Interfaces:

Solver

public class UnconstrainedSolver
extends java.lang.Object
implements Solver

Created on Nov 3, 2005 Unconstrained solution to mean variance portfolio optimization. Started with code from the applet at Campbell Harvey's website.

Field Summary

protected cern.colt.matrix.DoubleMatrix1D	_Er
protected cern.colt.matrix.DoubleMatrix2D	_V
protected java.text.DecimalFormat	df

Constructor Summary

UnconstrainedSolver(cern.colt.matrix.DoubleMatrix1D e_r, cern.colt.matrix.DoubleMatrix2D V)
Constructs

Method Summary

java.util.List<SolvedPoint>	solve(java.util.List<Constraint> extraConstraints, org.akutan.optimization.ProgressIndicator progress) Solves for the efficient frontier given the returns and covariances specified.
cern.colt.matrix.DoubleMatrix1D	solveForReturn(double r_e) This method is called to solve the mostly unconstrained optimization problem for a specific return.

Methods inherited from class java.lang.Object

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

Field Detail

df

protected java.text.DecimalFormat df

_Er

protected cern.colt.matrix.DoubleMatrix1D _Er

_V

protected cern.colt.matrix.DoubleMatrix2D _V

Constructor Detail

UnconstrainedSolver

public UnconstrainedSolver(cern.colt.matrix.DoubleMatrix1D e_r,
                           cern.colt.matrix.DoubleMatrix2D V)

Constructs

Parameters:

e_r - Vector of asset mean expected returns

V - Matrix of asset covariance of returns

Method Detail

solve

public java.util.List<SolvedPoint> solve(java.util.List<Constraint> extraConstraints,
                                         org.akutan.optimization.ProgressIndicator progress)

Solves for the efficient frontier given the returns and covariances specified.

Specified by:

solve in interface Solver

Parameters:

extraConstraints - Ignored (required for Solver)

Returns:

List of points on the efficient frontier

solveForReturn

public cern.colt.matrix.DoubleMatrix1D solveForReturn(double r_e)

This method is called to solve the mostly unconstrained optimization problem for a specific return.

Parameters:

r_e - Expected return to solve for.

Returns:

Vector of the asset weights for the portfolio with return r_e.