Interfaces Linear Programming algorithms.

Unit
MtxVecTools

Hierarchy
TMtxComponent <--TMtxLP

Subclasses
None

The component can be used to solve several linear programming problems.

• Drop a TMtxLP component on the form.
• Set the Mimimize property to minimize or maximize objective function f(x).
• By setting Algorithm select which algorithm will be used to solve LP problem.
• Define A,b, c and optionaly Relations , keeping in mind correct size of matrix, vectors and string.
• Call the Recalculate method to solve LP problem.

The results of solving LP are returned as:

• Final tableaux AFinal,
• Legitimate variables, stored in vector x,
• Indexes of legitimate variables, stored in vector indexes IValues,
• Solution type, stored in SolutionType,
• Objective function z, evaluated at minimum or maximum.

Linear Programming

Properties

Name	Summary
A	Defines A in A*x<=b equation.
AFinal	Returns the final tableaux.
Algorithm	Defines LP solving algorithm.
AutoUpdate
b	Defines b in A*x<=b equation.
c	Defines c in f=c(T)*x equation.
Dirty
Indexes	Returns indices (stored in IValues) of basic variables in final tableaux.
Minimize	Find minimum or maximum of objective function.
Relations	Defines relations for two phase algorithm.
SolutionType	Returns LP solution type.
Verbose	If not nil then the optimization method uses it for logging each optimization step.
x	Returns values of the legitimate variables (optimal solution).
z	Returns objective function, evaluated at minimum or maximum.

Methods

Name	Summary
CheckDimensions	Checks A,b,c and relations dimensions.
Recalculate	Triggers TMtxLP recalculation.

Copyright 2008 Dew Research

http://www.dewresearch.com