LinearFixPointEquationSystem.java

// FILE. . . . . d:/hak/hlt/src/hlt/math/matrix/sources/LinearFixPointEquationSystem.java
// EDIT BY . . . Hassan Ait-Kaci
// ON MACHINE. . Hak-Laptop
// STARTED ON. . Sat Nov 16 15:03:29 2019


Copyright:  © by the author
Author:  Hassan Aït-Kaci
Version:  Last modified on Mon Dec 16 09:52:22 2019 by hak


package hlt.math.matrix;

package hlt.math.matrix documentation listing



This implements a system of linear fix-point equations with double coefficients of the form: A × X + B = X where A is an m × n Matrix, B is an m × 1 Matrix (i.e., a column vector of order m), and X is a column vector of order n of variables Xi, i=1, ... n. For example, if m = 3, n = 4 and:

1.1 1.2 1.3 1.4 1.0 A = 2.1 2.2 2.3 2.4 B = 2.0 3.1 3.2 3.3 3.4 3.0

then:

1.1 × X1 + 1.2 × X2 + 1.3 × X3 + 1.4 × X4 + 1.0 = X1 2.1 × X1 + 2.2 × X2 + 2.3 × X3 + 2.4 × X4 + 2.0 = X2 3.1 × X1 + 3.2 × X2 + 3.3 × X3 + 3.4 × X4 + 3.0 = X3

which gives the equivalent non-fix-point formulation (A-I) × X = -B:

0.1 × X1 + 1.2 × X2 + 1.3 × X3 + 1.4 × X4 = -1.0 2.1 × X1 + 1.2 × X2 + 2.3 × X3 + 2.4 × X4 = -2.0 3.1 × X1 + 3.2 × X2 + 2.3 × X3 + 3.4 × X4 = -3.0

and conversely, from the more conventional non-fix-point formulation of the form A' × X = B', one recovers the corresponding equivalent fix-point form as: (A'+I) × X + (-B') = X, and then use this class. Non square equational systems may only have parametric solutions if there are more variables than equations (expressed in terms of the missing variables), or redundant solutions if there are more equations than variables. For a square matrix, a solution will exist iff it has an inverse (i.e., iff it has a non-zero determinant).

See also:  Matrix

public class LinearFixPointEquationSystem
{
}


This file was generated on Wed Dec 18 03:37:41 PST 2019 from file LinearFixPointEquationSystem.java
by the hlt.language.tools.Hilite Java tool written by Hassan Aït-Kaci