Está en la página 1de 16

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1

MAESTRA EN MANUFACTURA AVANZADA

Dr. Roberto Armando Hernndez Gmez

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


2 Sistemas Lineales

Mtodos Directos.

Sistema de ecuaciones lineales en forma matricial.


Solucin del sistema de ecuaciones lineales. Mtodos
directos e iterativos.
Matrices de forma especial. Identidad, diagonal, triangular.
Eliminacin de Gauss y Gauss-Jordan.
Estrategias de pivoteo.
Factorizacin A = LU. Costo de Computacin.
Matriz Permutacin. Factorizacin PA=LU.

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


2 Sistemas Lineales

Mtodos Iterativos.

Normas de vectores.
Convergencia. Tasas de convergencia.
Aplicacin (mapeo) contractiva.
Iteracin de punto fijo.
Error absoluto, error relativo y cifras significativas
Error de convergencia. Criterios de terminacin.
Mtodo Jacobi.
Mtodo Gauss-Seidel.

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1

Mtodos Directos.

Sistema de ecuaciones lineales en forma matricial.

Una matriz es un arreglo de nmeros dispuestos


sistemticamente en filas y columnas.

Se puede visualizar como


1 i M, 1 j N

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Una combinacin lineal de las variables x1, x2, , xN es
una suma
a1x1 + a2x2 + + aNxN
donde ak es el coeficiente de xk para k = 1, 2, , N
Una ecuacin lineal es la combinacin lineal igualada a un
valor b
a1x1 + a2x2 + + aNxN = b

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Sistemas de ecuaciones lineales surgen con frecuencia de modo
que cuando tenemos M ecuaciones lineales con N incgnitas
escribimos
a11x1 + a12x2 + + a1NxN = b1
a21x1 + a22x2 + + a2NxN = b2
.
.
.
aM1x1 + aM2x2 + + aMNxN = bM

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Una solucin al sistema de M ecuaciones con N
incgnitas es un conjunto de valores x1, x2, , xN tales
que satisfacen las M ecuaciones simultneamente, y se
dice que el vector N-dimensional
X = (x1, x2, , xN)
el sistema de ecuaciones se expresa en trminos de una
matriz A y los vectores X y B
AX = B

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Matrices Especiales

La matriz de MxN con todos sus elementos cero se


denomina la matriz cero de dimensin MxN
La matriz identidad de orden N es la matriz cuadrada dada
por
IN= [ij]NxN donde
Una matriz A de NxN se dice que es no-singular o
invertible si existe una matriz B de NxN tal que
AB = BA = I

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Matrices Especiales
si tal matriz B no es encontrada entonces se dice que A es
singular.
a la matriz B se le denomina la inversa de A y se le denota
como A-1.
Entonces A A-1= A-1A = I si A es no-singular
Una matriz A = es llamada triangular superior si sus
elementos cumplen que = 0 cuando i > j
Se llama triangular inferior si = 0 cuando i < j.

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Cuando se tiene un sistema de ecuaciones lineales triangular
superior se pueden obtener las soluciones desarrollando el
mtodo de substitucin hacia atrs. El cual consiste en:
La ltima ecuacin solo involucra a xN, asi que se resuelve
primero
Ahora que conocemos xN subtituimos en la penltima ecuacin
y

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Con xN y xN-1 procedemos a despejar xN-2 de la
antepenltima ecuacin y asi hasta la primera ecuacin
el trmino general es
Si A es una matriz triangular superior o inferior su determinante
es

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Eliminacin Gaussiana y pivoteo
Cuando se tiene un sistema de ecuaciones de N ecuaciones con
N incgnitas el sistema general AX= B se resuelve por
eliminacin gaussiana, que consiste en transformar el sistema
general en un sistema triangular superior UX=Y y entonces
resolverlo por substitucin hacia atrs.

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


Gaussian Elimination and Pivoting
In this section we develop a scheme for solving a general system AX = B of N equations and N unknowns. The goal
is to construct an equivalent upper-triangular system UX = Y that can be solved by the method of Section 3.3.
Two linear systems of dimension N x N are said to be equivalent provided that tlieir solution sets are the same.
Theorems from linear algebra show that when certain transformations are applied to a given system the solution sets
do not change.
Theorem 3.7 (Elementary Transformations). The following operations applied to a linear system yield an equivalent
system:
Interchanges: The order of two equations can be changed.
Scaling:
Multiplying an equation by a nonzero constant.
Replacement: An equation can be replaced by the sum of itself and
a nonzero multiple of any other equation.
It is common to use (3) by replacing an equation with the difference of that equation and a multiple of another
equation. These concepts are illustrated in the next example.
Example 3.15. Find the parabola y = A + Bx + Cx2 that passes through the three points (1, 1), (2,-1), and (3, 1).
For each point we obtain an equation relating the value of x to the value of y. The result is the linear system
A+ B+ C= 1 at (1,1)
A + 2B+4C = -\ at (2,-1)
A + 3B + 9C = 1 at (3,1).

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


The variable A is eliminated from the second and third equations by subtracting the first equation
from them. This is an application of the replacement transformation (3), and the resulting equivalent
linear system is
A ~h B -f- C ~ 1

S + 3C 2
2B + SC = 0.
The variable B is eliminated from the third equation in (5) by subtracting from it tv, o times the
second equation. We arrive at the equivalent upper-triangular system:
A+B+ C= 1

B +3C = 2

C= 4.
The back-substitution algorithm is now used to find the coefficients C = 4/2 = 2, B = 2 3(2) =
8, and A = 1 (8) 2 = 7, and the equation of the parabola y - 1 - ix + 2x7. 1
It is efficient to store all the coefficients of the linear system AX = B in an arr; \ of dimension N x (N
+ 1). The coefficients of B are stored in column W + 1 of tlu array (i.e., a*jv+i = &*) Each row
contains all the coefficients necessary to represeni an equation in the linear system. The augmented
matrix is denoted \A\B] and tlv

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


The system AX = B, with augmented matrix given in (7), can be solved by performing row
operations on the augmented matrix [A\B]. The variables xt are placeholders for the coefficients and
can be omitted until the end of the calculation.
Theorem 3.8 (Elementary Row Operations). The following operations applied to the augmented
matrix (7) yield an equivalent linear system.
(S)Interchanges: The order of two rows can be changed.
Scaling: Multiplying a row by a nonzero constant.
Replacement1, The row can be replaced by the sum of that row and
a nonzero multiple of any other row; that is: row r = row, mrp x row p.
It is common to use (10) by replacing a row with the difference of that row and a multiple of
another row.
Definition 3.3 (Pivot). The number arr in the coefficient matrix A that is used to eliminate a* r,
where k r + 1, f + 2, ..., N, is called the rthpivotal element, and the rth row is called the pivot
row. A
The following example illustrates how to use the operations in Theorem 3.8 to obtain an
equivalent upper-triangular system VX =Y from a linear system AX B where A is an N x N
matrix.

POSGRADO EN MANUFACTURA AVANZADA

ANLISIS NUMRICO APLICADO 1


The system AX = B, with augmented matrix given in (7), can be solved by performing row
operations on the augmented matrix [A\B]. The variables xt are placeholders for the coefficients and
can be omitted until the end of the calculation.
Theorem 3.8 (Elementary Row Operations). The following operations applied to the augmented
matrix (7) yield an equivalent linear system.
(S)Interchanges: The order of two rows can be changed.
Scaling: Multiplying a row by a nonzero constant.
Replacement1, The row can be replaced by the sum of that row and
a nonzero multiple of any other row; that is: row r = row, mrp x row p.
It is common to use (10) by replacing a row with the difference of that row and a multiple of
another row.
Definition 3.3 (Pivot). The number arr in the coefficient matrix A that is used to eliminate a* r,
where k r + 1, f + 2, ..., N, is called the rthpivotal element, and the rth row is called the pivot
row. A
The following example illustrates how to use the operations in Theorem 3.8 to obtain an
equivalent upper-triangular system VX =Y from a linear system AX B where A is an N x N
matrix.

También podría gustarte