Package org.ssclab.pl.milp

Class LP

java.lang.Object

org.ssclab.pl.milp.LP

All Implemented Interfaces:

FormatTypeInput

public final class LP
extends java.lang.Object
implements FormatTypeInput

This class allows executing and solving formulations of linear programming problems. The method used for solving such optimization problems is the simplex method.

Version:

4.0

Author:

Stefano Scarioli

See Also:

SSC Software www.sscLab.org

Nested Class Summary

Nested classes/interfaces inherited from interface org.ssclab.pl.milp.FormatTypeInput

FormatTypeInput.FormatType

Field Summary

Fields

Modifier and Type	Field	Description
static double	NaN

Constructor Summary

Constructors

Constructor	Description
LP(java.lang.String path)
LP(java.util.ArrayList<java.lang.String> inequality)
LP(LinearObjectiveFunction fo, java.util.ArrayList<Constraint> constraints)	Constructor Creates an LP object for solving problems expressed in matrix format.
LP(LinearObjectiveFunction fo, ListConstraints constraints)	Constructor of an LP object for solving problems expressed in matrix format.
LP(org.ssclab.ref.Input input_natural)	Constructor of an LP object for solving problems expressed in coefficient format.
LP(org.ssclab.ref.Input input_natural, org.ssclab.context.Session session)	Constructor of an LP object for solving problems expressed in coefficient format.
LP(org.ssclab.ref.Input input, org.ssclab.context.Session session, FormatTypeInput.FormatType format)	Constructor of an LP object for solving problems expressed in sparse format or coefficient format.
LP(org.ssclab.ref.Input input, FormatTypeInput.FormatType format)	Constructor of an LP object for solving problems expressed in either sparse or coefficient format.

Method Summary

Modifier and Type	Method	Description
int	getNumMaxIteration()	This method returns the maximum number of simplex iterations
Solution	getSolution()	If the problem has an optimal solution, this method returns that optimal solution in the form of an object of the Solution class.
double[][]	getStandartMatrixA()	This method returns the matrix A obtained after the process of reduction to standard form (max z , Ax + s=b, x ≥ 0, b ≥ 0) of the original linear programming problem.
double[]	getStandartVectorB()	This method returns the vector b of the rhs values obtained after the process of reduction to standard form(max z , Ax+s=b, x ≥ 0, b ≥ 0) of the original linear programming problem.
double[]	getStandartVectorC()	This method returns the vector c of the coefficients of the objective function after the process of reduction to standard form (max z , Ax+s=b, x ≥ 0, b ≥ 0) of the original linear programming problem.
LPThreadsNumber	getThreadsNumber()
double[]	getValuesSolution()	If the problem has an optimal solution, this method returns that optimal solution in the form of an array with the values of the variables.
boolean	isJustTakeFeasibleSolution()	Returns true if only phase 1 execution is set to obtain a feasible solution.
boolean	isParallelSimplex()
SolutionType	resolve()	Executes the simplex (phase 1 + phase 2).
void	setCEpsilon(EPSILON epsilon)	This method allows setting the epsilon value relative to the tolerance in determining if an optimal solution expressed by phase 1 is close to or equal to zero and thus gives rise to feasible solutions for the initial problem.
void	setEpsilon(EPSILON epsilon)	This method allows setting the epsilon value relative to the tolerance that intervenes in various contexts.
void	setJustTakeFeasibleSolution(boolean isStopPhase2)	Setting to true allows interrupting the simplex at the end of phase 1, in order to determine not an optimal solution but only a feasible solution of the problem.
void	setNumMaxIteration(int num_max_iteration)	This method allows limiting the maximum number of simplex iterations (phase 1 iterations + phase 2 iterations)
void	setParallelSimplex(boolean isParallelSimplex)	If the number of physical cores of the host on which SSc is running is greater than 4, the performance of the simplex can be improved by executing the optimization processes in parallel on multiple threads.
void	setThreadsNumber(LPThreadsNumber threadsNumber)	If the set value is AUTO, the system decides the number of threads to use.

Methods inherited from class java.lang.Object

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

Field Detail

NaN

public static double NaN

Constructor Detail

LP

public LP(java.util.ArrayList<java.lang.String> inequality)
   throws java.lang.Exception

Parameters:

inequality - An ArrayList (of String objects) containing the problem formulation in the form of inequalities

Throws:

java.lang.Exception - An exception is thrown if the problem is not correctly formulated

LP

public LP(java.lang.String path)
   throws java.lang.Exception

Parameters:

path - The path where the file containing the LP problem formulated with the inequality format is located

Throws:

java.lang.Exception - An exception is thrown if the problem is not correctly formulated or if the file does not exist

LP

public LP(LinearObjectiveFunction fo,
          java.util.ArrayList<Constraint> constraints)
   throws java.lang.Exception

Constructor Creates an LP object for solving problems expressed in matrix format.

Parameters:

fo - A LinearObjectiveFunction object representing the objective function

constraints - The list of constraints expressed as an ArrayList of Constraint objects

Throws:

java.lang.Exception - An exception is thrown if the problem is not correctly formulated

LP

public LP(LinearObjectiveFunction fo,
          ListConstraints constraints)
   throws java.lang.Exception

Constructor of an LP object for solving problems expressed in matrix format.

Parameters:

fo - A LinearObjectiveFunction object representing the objective function

constraints - The list of constraints as a ListConstraints object

Throws:

java.lang.Exception - An exception is thrown if the problem is not correctly formulated

LP

public LP(org.ssclab.ref.Input input,
          org.ssclab.context.Session session,
          FormatTypeInput.FormatType format)
   throws java.lang.Exception

Constructor of an LP object for solving problems expressed in sparse format or coefficient format.

Parameters:

input - The problem formulated in sparse format or coefficient format

session - An SSC working session

format - Constant to express in which format the problem is formulated (FormatType.SPARSE or FormatType.COEFF)

Throws:

java.lang.Exception - An exception is thrown if the problem is not formulated correctly

LP

public LP(org.ssclab.ref.Input input,
          FormatTypeInput.FormatType format)
   throws java.lang.Exception

Constructor of an LP object for solving problems expressed in either sparse or coefficient format.

Parameters:

input - The problem formulated in sparse format

format - Constant to express in which format the problem is formulated (FormatType.SPARSE or FormatType.COEFF)

Throws:

java.lang.Exception - An exception is thrown if the problem is not formulated correctly

LP

public LP(org.ssclab.ref.Input input_natural)
   throws java.lang.Exception

Constructor of an LP object for solving problems expressed in coefficient format.

Parameters:

input_natural - The problem formulated in coefficient format

Throws:

java.lang.Exception - An exception is thrown if the problem is not formulated correctly

LP

public LP(org.ssclab.ref.Input input_natural,
          org.ssclab.context.Session session)
   throws java.lang.Exception

Constructor of an LP object for solving problems expressed in coefficient format.

Parameters:

input_natural - The problem formulated in coefficient format

session - An SSC working session

Throws:

java.lang.Exception - An exception is thrown if the problem is not formulated correctly

Method Detail

setEpsilon

public void setEpsilon(EPSILON epsilon)

This method allows setting the epsilon value relative to the tolerance that intervenes in various contexts. It is used in the following cases:
During phase one, both in determining the entering variable and in determining the exiting variable with or without the Bland rule; it is also used to determine if the base is degenerate. It is also used at the end of phase one: if there is an auxiliary variable in the base, epsilon is used to determine if it is possible to eliminate the rows and columns of these on the extended table.
During phase two, both in determining the entering variable and in determining the exiting variable with or without the Bland rule; it is also used to determine if the base is degenerate.

Parameters:

epsilon - Tolerance used in various phases of the simplex. Default value 1E-10

setCEpsilon

public void setCEpsilon(EPSILON epsilon)

This method allows setting the epsilon value relative to the tolerance in determining if an optimal solution expressed by phase 1 is close to or equal to zero and thus gives rise to feasible solutions for the initial problem.

Parameters:

epsilon - Tolerance of phase 1 solution with respect to zero. Default value 1E-8

setNumMaxIteration

public void setNumMaxIteration(int num_max_iteration)
                        throws LPException

This method allows limiting the maximum number of simplex iterations (phase 1 iterations + phase 2 iterations)

Parameters:

num_max_iteration - The maximum number of iterations to be executed. Default value 10,000,000.

Throws:

LPException - If an incorrect number (zero or negative) is set

getNumMaxIteration

public int getNumMaxIteration()

This method returns the maximum number of simplex iterations

Returns:

The maximum number of iterations

resolve

public SolutionType resolve()
                     throws java.lang.Exception

Executes the simplex (phase 1 + phase 2).

Returns:

The type of solution found

Throws:

java.lang.Exception - If the execution process generates an error

getStandartMatrixA

public double[][] getStandartMatrixA()
                              throws org.ssclab.pl.milp.simplex.SimplexException,
                                     java.io.IOException

This method returns the matrix A obtained after the process of reduction to standard form (max z , Ax + s=b, x ≥ 0, b ≥ 0) of the original linear programming problem.

Returns:

The coefficient matrix A

Throws:

org.ssclab.pl.milp.simplex.SimplexException - If null matrix

java.io.IOException - if the problem has not been reduced to standard form

getStandartVectorB

public double[] getStandartVectorB()

This method returns the vector b of the rhs values obtained after the process of reduction to standard form(max z , Ax+s=b, x ≥ 0, b ≥ 0) of the original linear programming problem.

Returns:

The vector of RHS coefficients

getStandartVectorC

public double[] getStandartVectorC()

This method returns the vector c of the coefficients of the objective function after the process of reduction to standard form (max z , Ax+s=b, x ≥ 0, b ≥ 0) of the original linear programming problem.

Returns:

The vector c of the coefficients of the objective function.

getSolution

public Solution getSolution()
                     throws org.ssclab.pl.milp.simplex.SimplexException

If the problem has an optimal solution, this method returns that optimal solution in the form of an object of the Solution class.

Returns:

The optimal solution of the problem

Throws:

org.ssclab.pl.milp.simplex.SimplexException - If the optimal solution is not present

getValuesSolution

public double[] getValuesSolution()
                           throws org.ssclab.pl.milp.simplex.SimplexException

If the problem has an optimal solution, this method returns that optimal solution in the form of an array with the values of the variables.

Returns:

The optimal solution of the problem as an array of double values

Throws:

org.ssclab.pl.milp.simplex.SimplexException - If the optimal solution is not present

isParallelSimplex

public boolean isParallelSimplex()

Returns:

true if the parallelization with multiple threads of the simplex is active.

setParallelSimplex

public void setParallelSimplex(boolean isParallelSimplex)

If the number of physical cores of the host on which SSc is running is greater than 4, the performance of the simplex can be improved by executing the optimization processes in parallel on multiple threads.

Parameters:

isParallelSimplex - True to activate parallelization

getThreadsNumber

public LPThreadsNumber getThreadsNumber()

Returns:

the number of threads used in the execution. If the value is AUTO, the system decides the number of threads to use.

setThreadsNumber

public void setThreadsNumber(LPThreadsNumber threadsNumber)

If the set value is AUTO, the system decides the number of threads to use.

Parameters:

threadsNumber - Sets the number of threads to use in the execution.

isJustTakeFeasibleSolution

public boolean isJustTakeFeasibleSolution()

Returns true if only phase 1 execution is set to obtain a feasible solution.

Returns:

true if active

setJustTakeFeasibleSolution

public void setJustTakeFeasibleSolution(boolean isStopPhase2)

Setting to true allows interrupting the simplex at the end of phase 1, in order to determine not an optimal solution but only a feasible solution of the problem.

Parameters:

isStopPhase2 - true to interrupt the simplex before phase 2.