com.numericalmethod.suanshu.optimization.constrained.linearprogramming.simplex.standard
Class StandardSimplex

java.lang.Object
  com.numericalmethod.suanshu.optimization.constrained.linearprogramming.simplex.standard.StandardSimplex

All Implemented Interfaces:

LpSolver

public class StandardSimplex
extends java.lang.Object
implements LpSolver

In our implementation, we follow the convention in "Linear Programming with MATLAB," Michael C. Ferris, Olvi L. Mangasarian, Stephen J. Wright, that the inequalities have this form: A * x ≥ b, , b ≤ 0 iff the problem is feasible.

Constructor Summary

StandardSimplex(StandardLpProblem1 problem)

StandardSimplex(StandardLpProblem2 problem)

Method Summary

Tableau	getResultantTableau() Get a copy of the Tableau after the solve operation.
int	lambdaCol() Get the column index for which there is no row that passes the ratio test.
LpSolution	solve() Solve the Linear Programming (LP) problem.

Methods inherited from class java.lang.Object

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

Constructor Detail

StandardSimplex

public StandardSimplex(StandardLpProblem2 problem)

StandardSimplex

public StandardSimplex(StandardLpProblem1 problem)
                throws LpProblem.EmptyCostVector,
                       LpProblem.NoConstraint,
                       LpProblem.DimensionNotMatched

Throws:

LpProblem.EmptyCostVector

LpProblem.NoConstraint

LpProblem.DimensionNotMatched

Method Detail

solve

public LpSolution solve()
                 throws LpProblem.Infeasible

Description copied from interface: LpSolver

Solve the Linear Programming (LP) problem.

Specified by:

solve in interface LpSolver

Returns:

a (partial) solution

Throws:

LpProblem.Infeasible

lambdaCol

public int lambdaCol()

Description copied from interface: LpSolver

Get the column index for which there is no row that passes the ratio test. That is, the problem is unbounded.

To check whether a problem is bounded, we can do:

 lambdaCol() == 0;
 

Specified by:

lambdaCol in interface LpSolver

Returns:

0 if LP problem is bounded; +ve column index otherwise

getResultantTableau

public Tableau getResultantTableau()

Description copied from interface: LpSolver

Get a copy of the Tableau after the solve operation. The table can be passed to other operations, e.g., other LpSolver, for further processing.

Specified by:

getResultantTableau in interface LpSolver

Returns:

a Tableau as a result of the solve operation