SYSTEMS OF LINEAR EQUATIONS

Similar documents
FOURIER SERIES. Series expansions are a ubiquitous tool of science and engineering. The kinds of

3.4 Properties of the Stress Tensor

ERDOS-SMARANDACHE NUMBERS. Sabin Tabirca* Tatiana Tabirca**

minimize c'x subject to subject to subject to

CBSE , ˆj. cos CBSE_2015_SET-1. SECTION A 1. Given that a 2iˆ ˆj. We need to find. 3. Consider the vector equation of the plane.

1. Stefan-Boltzmann law states that the power emitted per unit area of the surface of a black

Section 5.1/5.2: Areas and Distances the Definite Integral

LECTURE 6 TRANSFORMATION OF RANDOM VARIABLES

CHAPTER 4. FREQUENCY ESTIMATION AND TRACKING

More Statistics tutorial at 1. Introduction to mathematical Statistics

Introduction to logistic regression

Inner Product Spaces INNER PRODUCTS

Machine Learning. Principle Component Analysis. Prof. Dr. Volker Sperschneider

Unbalanced Panel Data Models

Order Statistics from Exponentiated Gamma. Distribution and Associated Inference

CHAPTER 5d. SIMULTANEOUS LINEAR EQUATIONS

Lecture 3-4 Solutions of System of Linear Equations

Numerical Method: Finite difference scheme

Quantum Circuits. School on Quantum Day 1, Lesson 5 16:00-17:00, March 22, 2005 Eisuke Abe

Linear Prediction Analysis of Speech Sounds

CBSE SAMPLE PAPER SOLUTIONS CLASS-XII MATHS SET-2 CBSE , ˆj. cos. SECTION A 1. Given that a 2iˆ ˆj. We need to find

Advanced Algorithmic Problem Solving Le 3 Arithmetic. Fredrik Heintz Dept of Computer and Information Science Linköping University

Department of Mathematics and Statistics Indian Institute of Technology Kanpur MSO202A/MSO202 Assignment 3 Solutions Introduction To Complex Analysis

Linear Prediction Analysis of

Chapter 3 Fourier Series Representation of Periodic Signals

Binary Choice. Multiple Choice. LPM logit logistic regresion probit. Multinomial Logit

A general N-dimensional vector consists of N values. They can be arranged as a column or a row and can be real or complex.

COLLECTION OF SUPPLEMENTARY PROBLEMS CALCULUS II

COMPLEX NUMBERS AND ELEMENTARY FUNCTIONS OF COMPLEX VARIABLES

Lecture 1: Empirical economic relations

MODEL QUESTION. Statistics (Theory) (New Syllabus) dx OR, If M is the mode of a discrete probability distribution with mass function f

TOTAL LEAST SQUARES ALGORITHMS FOR FITTING 3D STRAIGHT LINES

ITERATIVE METHODS FOR SOLVING SYSTEMS OF LINEAR ALGEBRAIC EQUATIONS

Line Matching Algorithm for Localization of Mobile Robot Using Distance Data from Structured-light Image 1

Bayesian Shrinkage Estimator for the Scale Parameter of Exponential Distribution under Improper Prior Distribution

page 11 equation (1.2-10c), break the bar over the right side in the middle

Vtusolution.in FOURIER SERIES. Dr.A.T.Eswara Professor and Head Department of Mathematics P.E.S.College of Engineering Mandya

Second Handout: The Measurement of Income Inequality: Basic Concepts

A note on Kumaraswamy Fréchet distribution

Complex Numbers. Prepared by: Prof. Sunil Department of Mathematics NIT Hamirpur (HP)

IFYFM002 Further Maths Appendix C Formula Booklet

SOLVING SYSTEMS OF EQUATIONS, DIRECT METHODS

Ordinary Least Squares at advanced level

T and V be the total kinetic energy and potential energy stored in the dynamic system. The Lagrangian L, can be defined by

CIVL 8/ D Boundary Value Problems - Rectangular Elements 1/7

How much air is required by the people in this lecture theatre during this lecture?

Handout 11. Energy Bands in Graphene: Tight Binding and the Nearly Free Electron Approach

Course 10 Shading. 1. Basic Concepts: Radiance: the light energy. Light Source:

Entropy Equation for a Control Volume

Soo King Lim Figure 1: Figure 2: Figure 3: Figure 4: Figure 5: Figure 6: Figure 7: Figure 8: Figure 9: Figure 10: Figure 11:

Counting the compositions of a positive integer n using Generating Functions Start with, 1. x = 3 ), the number of compositions of 4.

Aotomorphic Functions And Fermat s Last Theorem(4)

Computer Programming

Minimum Spanning Trees

Gilbert the Green Tree Frog

CHAPTER 7d. DIFFERENTIATION AND INTEGRATION

COMPLEX NUMBERS AND DE MOIVRE S THEOREM

Integration by Guessing

Worksheet: Taylor Series, Lagrange Error Bound ilearnmath.net

Chapter Gauss-Seidel Method

Chp6. pn Junction Diode: I-V Characteristics II

European Journal of Mathematics and Computer Science Vol. 3 No. 1, 2016 ISSN ISSN

Research on the Massive Data Classification Method in Large Scale Computer Information Management huangyun

Linear Algebra Existence of the determinant. Expansion according to a row.

The Theory of Small Reflections

Fundamentals of Continuum Mechanics. Seoul National University Graphics & Media Lab

A NEW GENERALIZATION OF THE EXPONENTIAL-GEOMETRIC DISTRIBUTION

Chapter Discrete Fourier Transform

Overview. Splay trees. Balanced binary search trees. Inge Li Gørtz. Self-adjusting BST (Sleator-Tarjan 1983).

On Estimation of Unknown Parameters of Exponential- Logarithmic Distribution by Censored Data

ASSERTION AND REASON

A Study of Fuzzy Linear Regression

CS537. Numerical Analysis

Bayesian belief networks: Inference

Integration Continued. Integration by Parts Solving Definite Integrals: Area Under a Curve Improper Integrals

Estimating the Variance in a Simulation Study of Balanced Two Stage Predictors of Realized Random Cluster Means Ed Stanek

Different types of Domination in Intuitionistic Fuzzy Graph

Introduction to logistic regression

Planar convex hulls (I)

Total Prime Graph. Abstract: We introduce a new type of labeling known as Total Prime Labeling. Graphs which admit a Total Prime labeling are

Using Nonlinear Filter for Adaptive Blind Channel Equalization

signal amplification; design of digital logic; memory circuits

CURVE FITTING LEAST SQUARES METHOD

INTEGRALS. Chapter 7. d dx. 7.1 Overview Let d dx F (x) = f (x). Then, we write f ( x)

A Monotone Process Replacement Model for a Two Unit Cold Standby Repairable System

Stats & Summary

THE TRANSMUTED GENERALIZED PARETO DISTRIBUTION. STATISTICAL INFERENCE AND SIMULATION RESULTS

D. Bertsekas and R. Gallager, "Data networks." Q: What are the labels for the x-axis and y-axis of Fig. 4.2?

Area and the Definite Integral. Area under Curve. The Partition. y f (x) We want to find the area under f (x) on [ a, b ]

Repeated Trials: We perform both experiments. Our space now is: Hence: We now can define a Cartesian Product Space.

Statistical Thermodynamics Essential Concepts. (Boltzmann Population, Partition Functions, Entropy, Enthalpy, Free Energy) - lecture 5 -

ASYMPTOTIC AND TOLERANCE 2D-MODELLING IN ELASTODYNAMICS OF CERTAIN THIN-WALLED STRUCTURES

RAKE Receiver with Adaptive Interference Cancellers for a DS-CDMA System in Multipath Fading Channels

Chapter 2 Reciprocal Lattice. An important concept for analyzing periodic structures

CHAPTER 7. X and 2 = X

ENGI 3424 Appendix Formulæ Page A-01

12 Iterative Methods. Linear Systems: Gauss-Seidel Nonlinear Systems Case Study: Chemical Reactions

LE230: Numerical Technique In Electrical Engineering

under the curve in the first quadrant.

III Z-Plane Analysis

Transcription:

SYSES OF INER EQUIONS Itroducto Emto thods Dcomposto thods tr Ivrs d Dtrmt Errors, Rsdus d Codto Numr Itrto thods Icompt d Rdudt Systms

Chptr Systms of r Equtos /. Itroducto h systm of r qutos s formd y th ddto of th products of vr wth coffct, whch s so costt. h systm of r quto c sovd v mtr pproch. h gr form of st of r quto hvg r qutos d ukows s O (.) whr r vrs or ukows,,,, K j d j r coffct or costt (r or comp). Eq. (.) c wrtt mor compct form: } { } { ] [ j j (.) whr s mtr [ j ] of sz, s vr vctor { j } d s rght-hd sd vctor { j }. h procss of sovg Eq. (.) yd thr poss soutos:. Uqu souto.g.:. No souto.g.:. Ift soutos.g.:

Chptr Systms of r Equtos /. Emto thods h most popur mthod s th Guss mto mthod, whch comprss of two stps:. Forwrd mto to form uppr trgur systm v rowsd trsformto procss,. Bck susttuto to produc th souto of j. Cosdr th foowg systm: O If, for,,,, sustrct th -th quto wth th product of wth th frst quto to produc th frst trsformd systm: () () () () () () O whr () j j j for, j,,, () for,,, h procss c rptd for () tms ut th ()-th trsformd systm s formd s foowd, whch compts th forwrd mtos: (.) () () () () ( ) ( ) ( ) ( ) ( ) O

Chptr Systms of r Equtos / whr ( k) ( k ) ( k ) k j j ( k) kk ( ) ( ) ( k k k ) k ( k) ( k kj ( k k kk ) ) for, j k,, for k,, (.) (.) Bck susttutos c th cutd so tht j r sovd: ( ) (.) ( ) k ( k ) ( k ) ( ) k k kj kk j k j for k,, (.) h ov mthod c f f kk, th row hs to trchgd, whch s rfrrd to s pvotg: Pvotg kk whr th w dgo mt s cd pvot, whch c sctd mog th mmum sout vu of h pvot Guss mto gvs mor ccurt soutos,.g. cosdr ths systms (vus to roudd up to sgfct fgurs): k. Org Guss mto:.6...8 Pvot Guss mto:..8 ( ) () ( ).6.6.6.....8.8.8.6..8.6 ( ) () ( )...6..8..8. Ect souto:.8.8.6.

Chptr Systms of r Equtos / Emp. Sov th foowg systm usg th Guss mto mthod: Souto h systm c rwrtt mtr form s: or [ ] Frst stp of forwrd mto: 6 () () () () () () Scod stp of forwrd mto: () () () Hc, th trsformd uppr trgur systm s: Bck susttutos r s foows

Chptr Systms of r Equtos / 6 Emp. Prform th pvot Guss mto to th systm gv Emp.. Souto h pvot Guss mto c prformd s foowd: () ( ) ( ) () ( ) ( ) () ( ) ( ) ( ) ( ) ( ) ( ) Hc, th uppr trgur systm s: h, ck susttuto c prformd: () ( ) ( ).,,

Chptr Systms of r Equtos /. Dcomposto thods I som css, th ft-hd sd mtr s frquty usd wh th rght-hd sd vctor s chgd dpdg o th cs. h ovr systm c trsformd to uppr trgur form so tht t c usd rptdy for dffrt, thus mtr hs to dcomposd. For gr o-symmtrc systm, th popur mthod s th Doot or U dcomposto: U (.6) whr d U r th owr d uppr trgur mtrcs, rspctvy: u u u u u u u u u u u ( mmory) u h souto stps of th systm r s foowd: By tkg trmdt vctor y: Hc, U U y (.) y (.8) h mts for d U c otd from th Guss mto: U () () ( ) () ()

Chptr Systms of r Equtos / 8 othr vrto of th U dcomposto s th Crout dcomposto, whch mts u for,,, U std of : For th frst row d coum: For j,,,: for,,, (.) j u j for j,,, (.) j j j k k u kj for j, j,, (.c) u jk jk j jj j u k for k j, j,, (.d) d, k k u k (.) If th systm s symmtrc, th Chosky dcomposto c usd, whr mtr c dcomposd such tht: For th k-th row: (.) k k j j kj for,,,k (.) k kk j kk kj (.) hs mthod optmss th us of computr mmory storg th dcomposd form of.

Chptr Systms of r Equtos / Emp. Dcompos th foowg mtr usg th Doott U dcomposto: Souto Wth rfrc to th mtr mts drvd Emp.:., U Emp. Dcompos th foowg mtr usg th Chosky dcomposto: Souto By usg Eq..:.,,,,, k,.88..88..

Chptr Systms of r Equtos /. tr Ivrs d Dtrmt h Guss mto c usd to grt th vrs of squr mtr y rpcg th ft-hd sd vctor wth dtty mtr I. By usg th foowg dtty: I (.) ( ) ( ) ( ) If coums of r wrtt s,, K, d th coums of th I s () ( ) (,, K, ), rspctvy, thus Eq. (.) c rwrtt s: () ( ) ( ) ( ) ( ) ( ) (,, K, ),, K ( ), h, st of r systms c ssmd: ( ) ( ) ( ) ( ) ( ) ( ) (.) Cosquty, th dtrmt of mtr c smpy ccutd usg: dt (.) p () ( ) ( ) p ( ) ( ) ( ) K ( ) whr p s th umr of row trchg oprto durg pvotg.

Chptr Systms of r Equtos / Emp.8 Dtrm th vrs of th foowg mtr usg th Guss mto: Souto h comto of d I c rprstd ugmtd form: mto forwrd Guss Upo ck susttuto: () 8 () () Hc, th vrs of s 8 Emp. Ccut th dtrmt of th mtr gv Emp.8. Pys I Emp., thr s o row trchg prformd, thus p. Hc, ( ) ( ) ( )()() dt

Chptr Systms of r Equtos /. Errors, Rsdus d Codto Numr If s ppromt souto of r systm, th th systm rror s dfd s (.6) O th othr hd, th systm rsdu r s dfd s r (.) or, r For w-codtod systm, th rsdu c rprst th rror. orovr, for comprso, mtr or vctor c prssd form of scr kow s orm. For vctor (, K ), th p-orm s dfd s,, p p p ( ) p p p p (.) If p, t s kow s -orm: (.) If p, t s kow s Eucd orm: (.8) If p, t s kow s mmum orm: {,, K, } m (.) m For mtr [ j ] of sz m, th Frous orm, whch s quvt to th Eucd orm for vctors, s dfd s m j j (.)

Chptr Systms of r Equtos / d, th quvt -orm d mmum orm for mtr r dfd s m mmum sum of coums j (.) j m mmum sum of rows j (.) j h proprts of orms of vctor or mtr r s foowd:. d f, d oy f,.. c c whr c s scr qutty.. B B rgur quty, whr B s vctor or mtr of th sm dmso of.. B B Schwrz quty, whr B s vctor or mtr whch forms vd product wth. h cocpt of orms c usd to ccut th codto umr rprsts th hth of r systm, thr - or w-codtod. If s th rror for th systm, from th rtos r d r, th foowg quty c stshd: r d r r r so, from d : d hus, th comto of oth quty rtos yds th rg of th rtv rror,.. r ( ) r Hc, th codto umr s dfd s ( ) κ (.)

Chptr Systms of r Equtos / whr th rg of th rtv rror s. κ ( ) r κ ( ) r (.6) h chrctrstcs of th codto umr r tht: κ th smr th ttr, d othrws.. ( ) κ, th rtv rsdu r c rprst th rtv. If ( ) rrors. If th rror s soy cotrutd y mtr, th quty coms: E κ ( ) (.) O th othr hd, f th rror s soy cotrutd y vctor, th quty coms: κ ( ) (.8) hrfor, from Eqs. (.6-8), t c s tht th codto umr c dtrm th rg of rror d thus th hth of systm.

Chptr Systms of r Equtos /. Itrto thods For rg systms (sz > ), th mto d dcomposto mthods r ot ffct du to crsg umr of rthmtc oprtos. h umr of rthmtc oprtos c rducd v trto mthods, such s th Jco trto d th Guss-Sd trto mthods. I th Jco trto, Eq. (.) c wrtt for from th -th quto: ( ) ( ) ( ), (.) () () () () Eq. (.) ds t vus ( ),, K, () () () () (, K, ),, whch yd, d th computto cotus s foowd: ( k ) ( k ) ( k ) ( k ) ( k ) ( k ) ( ) ( k ) ( k ) ( k ) ( ) ( k ) ( k ) ( k ) ( ), (.) For k, vctor (k) covrgs to ts ct souto f th dgo dom codto s foowd,.. > j for,, K, j j (.) d th mtr whch foows ths codto s cd dgo dom mtr.

Chptr Systms of r Equtos / 6 o trmt th trto procss, covrgc or trmto crtro c spcfd,.. ( k ) ( k ) < ε (.) h Guss-Sd trto mthod uss th most currt kow souto ftr ch rthmtc oprto ordr to spd up covrgc: ( k ) ( k ) ( k ) ( k ) ( k ) ( k ) ( ) ( k ) ( k ) ( k ) ( ) ( k ) ( k ) ( k ) ( ), (.) s of th Jco mthod, th Guss-Sd mthod must so osrv th dgo dom codto for covrgc to poss (s Fg..). ( 6, ) ( 6, ) () h off-dgo dom systm () h dgo dom systm FIG.. Dvrgc d covrgc th Guss-Sd mthod

Chptr Systms of r Equtos / Emp. Us th Jco trto mthod to sov th foowg systm up to dcm pots: Souto 6 Frst of, form dgo dom systm: 6 h, rwrt th systm ccordg to Eq. (.): ( k ) ( k ) ( k ) 6 ( k ) ( k ) ( k ) ( ) ( k ) ( k ) ( k ) ( ) ( ) By tkg t vus () (,, ), thus th mthod covrgs wth trtos: Itrto o. : () (.8,.6,.), Itrto o. : () (.,.6,.), Itrto o. : () (.,.66,.6), Itrto o. : () (.,.66,.6), Itrto o. : () (.,.66,.6).

Chptr Systms of r Equtos / 8 Emp. Rpt prom gv Emp. usg th Guss-Sd trto mthod. Souto Frst of, form dgo dom systm: 6 By tkg t vus () (,, ), th frst souto th frst trto: () 6 [ ( ) ]. 8 () () Us to ccut d so o,.. [ (.8) ]. 6 () (.8.6 ). () Hc, th mthod covrgs wth trtos: () (.8,.6,.), () (.,.66,.6), () (.,.66,.6), () (.,.66,.6).

Chptr Systms of r Equtos /.8 Icompt d Rdudt Systms If, thr w two stutos: m. m < compt systm.. m > rdudt systm. For compt systm, o souto s poss sc ddto ( m) qutos from othr dpdt sourcs r rqurd ut m. For rdudt systm, uqu souto s ot poss, d th systm hs to optmsd v st squr mthod (so kow s r rgrsso): ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ).,,, S Usg th dtty ( ) : B B S msg S: S forms ppromt systm of qutos,.. (.) whr th ft-hd sd mtr s symmtry d th stdrd dvto σ c ccutd from th Eucd orm of,..: m m S σ (.)

Chptr Systms of r Equtos / Emp. Ccut th st ppromt souto for th foowg systm: so, ccut th rsutg stdrd dvto. Souto h ov systm c rwrttd form of s: By usg Eq. (.): 6 8 whr ts soutos r...6,.,

Chptr Systms of r Equtos / h stdrd dvto c otd from th Eucd orm of th rror :.8..68.6...6. ( ) ( ) ( ).8.,.8..68.6. hrfor,.6.8 σ

Chptr Systms of r Equtos / Ercss. Cosdr th foowg systm:..8.. 8. Us th Guss mto mthod to ot th souto of.. Ccut th dtrmt for th ft-hd sd mtr. c. Grt th owr d uppr trgur mtrcs usg th Doott fctorsto.. Cosdr th foowg systm of comp qutos: z z By wrtg, sov th quto usg th Guss-Sd trto mthod usg crosoft Ec ut t covrgs up to dcm pots. y z k k k. Cosdr th foowg st of rdudt qutos:. Drv ppromt systm of r qutos d sov t v th Guss mto.. Ccut th corrspodg stdrd dvto.