P. 1
The.Manga.Guide.to.Linear.Algebra.pdf

The.Manga.Guide.to.Linear.Algebra.pdf

|Views: 2,504|Likes:
Published by Iliyan Milen
The.Manga.Guide.to.Linear.Algebra
The.Manga.Guide.to.Linear.Algebra

More info:

Published by: Iliyan Milen on Feb 22, 2013
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

08/20/2013

pdf

text

original

Th e M a n g a G u i de to

Comics inside!

LINEAR ALGEBRA
Shin Takahashi Iroha Inoue TREND-PRO Co., Ltd.

Praise for the Manga Guide series
“Highly recommended.” —choice magazine on the manga guide to databases “Stimulus for the next generation of scientists.” — scientific computing on the manga guide to molecular biology “A great fit of form and subject. Recommended.” — otaku usa magazine on the manga guide to physics “The art is charming and the humor engaging. A fun and fairly painless lesson on what many consider to be a less-than-thrilling subject.” — school library journal on the manga guide to statistics “This is really what a good math text should be like. Unlike the majority of books on subjects like statistics, it doesn’t just present the material as a dry series of pointlessseeming formulas. It presents statistics as something fun, and something enlightening.” — good math, bad math on the manga guide to statistics “I found the cartoon approach of this book so compelling and its story so endearing that I recommend that every teacher of introductory physics, in both high school and college, consider using it.” — american journal of physics on the manga guide to physics “The series is consistently good. A great way to introduce kids to the wonder and vastness of the cosmos.” —discovery.com on the manga guide to the universe “A single tortured cry will escape the lips of every thirtysomething biochem major who sees The Manga Guide to Molecular Biology: ‘Why, oh why couldn’t this have been written when I was in college?’” —the san francisco examiner “Scientifically solid . . . entertainingly bizarre.” — chad orzel , author of how to teach physics to your dog,
on the manga guide to relativity

Wow!

“A lot of fun to read. The interactions between the characters are lighthearted, and the whole setting has a sort of quirkiness about it that makes you keep reading just for the joy of it.” — hack a day on the manga guide to electricity

“The Manga Guide to Databases was the most enjoyable tech book I’ve ever read.” — rikki kite, linux pro magazine “The Manga Guides definitely have a place on my bookshelf.” — smithsonian’s “surprising science” “For parents trying to give their kids an edge or just for kids with a curiosity about their electronics, The Manga Guide to Electricity should definitely be on their bookshelves.” — sacramento book review “This is a solid book and I wish there were more like it in the IT world.” —slashdot on the manga guide to databases “The Manga Guide to Electricity makes accessible a very intimidating subject, letting the reader have fun while still delivering the goods.” — geekdad blog, wired.com “If you want to introduce a subject that kids wouldn’t normally be very interested in, give it an amusing storyline and wrap it in cartoons.” — make on the manga guide to statistics “A clever blend that makes relativity easier to think about—even if you’re no Einstein.” — stardate, university of texas, on the manga guide to relativity “This book does exactly what it is supposed to: offer a fun, interesting way to learn calculus concepts that would otherwise be extremely bland to memorize.” — daily tech on the manga guide to calculus “The art is fantastic, and the teaching method is both fun and educational.” — active anime on the manga guide to physics “An awfully fun, highly educational read.” — frazzleddad on the manga guide to physics “Makes it possible for a 10-year-old to develop a decent working knowledge of a subject that sends most college students running for the hills.” — skepticblog on the manga guide to molecular biology “This book is by far the best book I have read on the subject. I think this book absolutely rocks and recommend it to anyone working with or just interested in databases.” — geek at large on the manga guide to databases “The book purposefully departs from a traditional physics textbook and it does it very well.” — dr. marina milner-bolotin, ryerson university on the manga guide to physics “Kids would be, I think, much more likely to actually pick this up and find out if they are interested in statistics as opposed to a regular textbook.” — geek book on the manga guide to statistics

The Manga Guide™ to Linear Algebra

.

The Manga Guide™ to LineAr ALGebrA Shin Takahashi. Ltd. iroha inoue. . and Trend-Pro Co..

) -. living or dead. Inoue. etc. and Ohmsha. Ltd. QA184.. with no intention of infringement of the trademark. recording. electronic or mechanical. No part of this work may be reproduced or transmitted in any form or by any means.50222--dc23 2012012824 No Starch Press and the No Starch Press logo are registered trademarks of No Starch Press. Inc. [Manga de wakaru senkei daisu. of Tokyo. info@nostarch.9900. ISBN 978-1-59327-413-9 (pbk. Trend-pro Co.nostarch. directly: No Starch Press. shall have any liability to any person or entity with respect to any loss or damage caused or alleged to be caused directly or indirectly by the information contained in it. cm.The Manga Guide to Linear Algebra. CA 94103 phone: 415. II. This English edition is co-published by No Starch Press. Ltd. and any resemblance to real persons. Ltd. English] The manga guide to linear algebra / Shin Takahashi. Other product and company names mentioned herein may be the trademarks of their respective owners. http://www. . Inc.863. While every precaution has been taken in the preparation of this work. Manga de wakaru senkeidaisuu. 2.863. All characters in this publication are fictitious. p. Iroha. Inc. Ltd. fax: 415. 38 Ringold Street. First printing 16 15 14 13 12 123456789 ISBN-10: 1-59327-413-0 ISBN-13: 978-1-59327-413-9 Publisher: William Pollock Author: Shin Takahashi Illustrator: Iroha Inoue Producer: TREND-PRO Co. Copyright © 2012 by Shin Takahashi and TREND-PRO Co. Shin. Graphic novels. III. The information in this book is distributed on an “As Is” basis.com/ Library of Congress Cataloging-in-Publication Data Takahashi.. Iroha Inoue. we are using the names only in an editorial fashion and to the benefit of the trademark owner. © 2008 by Shin Takahashi and TRENDPRO Co. published by Ohmsha. without warranty. Ltd. Production Editor: Alison Law Developmental Editor: Keith Fancher Translator: Fredrik Lindh Technical Reviewer: Eric Gossett Compositor: Riley Hoffman Proofreader: Paula L.2. Fleming Indexer: BIM Indexing & Proofreading Services For information on book distributors or translations. The Manga Guide to Linear Algebra is a translation of the Japanese original.9950. Rather than use a trademark symbol with every occurrence of a trademarked name.ISBN 1-59327-413-0 (pbk. Inc. including photocopying.com.T3513 2012 512’. strips. Title. Algebras. Trend-pro Co.) 1. Inc. San Francisco. Linear--Comic books. I. or by any information storage or retrieval system. All rights reserved. is purely coincidental.. without the prior written permission of the copyright owner and the publisher. Ltd. please contact No Starch Press. Japan. neither the author nor No Starch Press.

. . . . . . . . . . . . . . . . 30 Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 . . . . . . . . . . . 44 Onto and One-to-One Functions . . . . . . . . 9 An Overview of Linear Algebra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Images . . . . . . 25 Implication and Equivalence . . . . . . . . . . . Zero Matrices. . . . 27 Propositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2 The Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 What is Linear Algebra? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 . 28 Equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Number Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Inverse Functions . 50 Combinations and Permutations. . . Matrix Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Set Symbols. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subtraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 . . Matrix Multiplication. . . . . . . . . . . . . . . . . . . . . 71 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Contents Preface . 48 Linear Transformations. . . . . . . . . 55 Not All “Rules for Ordering” Are Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Domain and Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 . . . . . . . . . . . . . 27 Implication . . . . . . . . . . . . . . Special Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3 intro to Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 What Is a Matrix? . . . xi Prologue Let the Training begin! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Set Theory . . . . . . . . . . . . . . . . . . . . . . . . . Scalar Multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Subsets . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Calculating Inverse Matrices Using Cofactors . . . . . . . . . .Transpose Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 3-D Projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Upper Triangular and Lower Triangular Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 5 introduction to Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Special Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 . . . . . . . . . . . . . . 166 Why We Study Linear Transformations. . . . . . . . . . . . . . 86 Calculating Inverse Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 Scaling . . . . . . . . . . . . . . . . . . . . . . . 131 Linear Independence . . . 111 Solving Linear Systems with Cramer’s Rule . . . . . 149 . . . . . . . . . . . . . . . . . . . . . . 113 What Are Vectors? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Using Determinants . . . . . . . . . . . . . . . . . 78 Symmetric Matrices . . . . 109 Calculating Inverse Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 4 More Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Geometric Interpretations . . . . . . . . . . . . . . . . Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 viii Contents . . . . . . Bases . . . . . . . . . . . . . . . . . . . . . . . 88 Determinants . . . 163 What Is a Linear Transformation?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Identity Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 6 More Vectors . . . . . . . . . . . . 161 7 Linear Transformations. . . Basis and Dimension Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 . . . . . . . . . 108 Cij . 108 Mij . . . . . . . . . . 79 Diagonal Matrices . . . . . . . . . . . . . . . . . . . . . . . 116 Vector Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Inverse Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Calculating Determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 Translation . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Image. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . and the Dimension Theorem for Linear Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 Kernel. . . . . . . . . . . . . 243 The Appendixes . 243 index . . . . . . . . . . . . . . . . . . . . . . Calculating the pth Power of an nxn Matrix . . . . . . . 189 Rank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 epilogue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multiplicity and Diagonalization . . . . . . . . . . . . . . . . . . . . . . . . . Calculating Eigenvalues and Eigenvectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Contents ix . . . 231 Online resources . . . . . 243 Updates . . . . . . . A Diagonalizable Matrix with an Eigenvalue Having Multiplicity 2 A Non-Diagonalizable Matrix with a Real Eigenvalue Having Multiplicity 2. . . . . . . . . . . . . . 205 What Are Eigenvalues and Eigenvectors? . . . . . . . . . . . . . . . . . . 196 The Relationship Between Linear Transformations and Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Some Preliminary Tips . . . . . . . . . . . . . . 216 . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 . . . . . . . . . . . . . . . . . . . . 211 . . . . . 203 8 eigenvalues and eigenvectors . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 . . . . . . . . . . 219 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 Calculating the Rank of a Matrix . . . . . . . .

.

for giving me the opportunity to write this book. the book’s illustrator. I would also like to express my gratitude towards re_akino. I recommend that you read both parts and then review each subject in more detail for maximal effect. Shin Takahashi november 2008 . I would like to thank my publisher. While skipping the text parts and reading only the manga will give you a quick overview of each subject.Preface This book is for anyone who would like to get a good overview of linear algebra in a relatively short amount of time. as well as Iroha Inoue. and everyone at Trend Pro who made it possible for me to convert my manuscript into this manga. or those who are already taking the course and need a helping hand Students who have taken linear algebra in the past but still don’t really understand what it’s all about High school students who are aiming to enter a technical university Anyone else with a sense of humor and an interest in mathematics! The book contains the following parts: Chapter 1: What Is Linear Algebra? Chapter 2: The Fundamentals Chapters 3 and 4: Matrices Chapters 5 and 6: Vectors Chapter 7: Linear Transformations Chapter 8: Eigenvalues and Eigenvectors Most chapters are made up of a manga section and a text section. Ohmsha. more comprehensive literature. who created the scenario. This book is meant as a complement to other. Those who will get the most out of The Manga Guide to Linear Algebra are: • • • • University students about to take linear algebra. I thank you all. I also received plenty of good advice from Kazuyuki Hiraoka and Shizuka Hori. not as a substitute.

.

Prologue Let the Training begin! .

Se aa! ya Hanamichi University Ei! nothing to be afraid of.. Okay! * it's now or never! * Hanamichi Karate Club ..

Ba-du m excu— Bump Wha— e Rattl Ba-du m what do we have here? Let the Training begin! 3 .

the karate club captain? indeed. but i think i— Are you serious? My students would chew you up and spit you out. ? i can't back down now! i want to join the karate club! ank Y ow B 4 Prologue i don't have any experience.. Would you by any chance be Tetsuo ichinose.i... the Hanamichi Hammer himself! U-um. My name is reiji Yurino.. ... Wow.. i'm a freshman..

. ak e Hmm? Haven't i seen your face somewhere? Ah a! Uhh.. Let the Training begin! 5 ..Please! i— i want to get stronger! Sh ak e Sh ..

Wha—? Y-yes.. i see. hmm i might consider letting you into the club. 6 Prologue .. you've seen my book? So it is you! i may not be the strongest guy.... but i've always been a whiz with numbers.Aren't you that guy? The one on my sister's math book? Mathe matic sf A uth o or every on e r: reij i Yurin o by st ud For s ents— tuden ts Oh..

You have to tutor my little sister in math. err So if i tutor your sister you'll let me in the club? Let the Training begin! 7 ..really?! . She's never been that good with numbers..Under one condition! And she complained just yesterday that she's been having trouble in her linear algebra class..... you see.

... We won't go easy on you.i suppose i should give you fair warning... Crack i. follow me. Would that be acceptable to you? Of course! if you try to make a pass at her. ya know.. even once... wouldn't think of it! Snap in that case. We'll start right away! Of course! i'm in! 8 Prologue ...

1 What is Linear Algebra? Vectors Matrices .

i somehow doubt you guys have met.Okay! That's all for today! Ossu!* Pa n t Pa n t bow! Ossu! Thank you! Yurinooo! Still alive. i told her to wait for you at. but since there seem to be a lot of you this year.. ... Gr ab You're free to start tutoring my sis after you've cleaned the room and put everything away. alright? She's also a freshman here. b Wo e bl a ke Sh * Ossu is an interjection often used in Japanese martial arts to enhance concentration and increase the power of one's blows. eh? O-ossu..

. i'm Misa ichinose... She looks a little like Tetsuo. Thu mp i wonder what she'll be like.. Umm. are you reiji Yurino? Wha— Pleased to meet you...mmh.. Oww.. excuse me..

an autograph? no.Sorry about my brother asking you to do this all of a sudden.... would it be awkward if i asked you to sign this? Like. no problem! i don't mind at all! but.. To u c h Only if you want to.. 12 Chapter 1 . i've been looking forward to this a lot! err. How could this girl possibly be his sister?! i had no idea that i went to school with the famous reiji Yurino! Well.... if it’s too weird— Gulp. i wouldn’t exactly say famous. it's my pleasure.

So— . Sor— —ry Sorry Misa. big brother! Yurinooo You remember our little talk? Yep! What is Linear Algebra? 13 .. but i think i'll head home a bit early today..

... it is true that linear algebra is a pretty abstract subject.. i don't really understand the concept of it all.. when would you like to start? How about right now? Yes.An Overview of Linear Algebra Well then. Let’s see.. but! Linear nce independe and there are some hardto-understand concepts. Your brother said that you were having trouble with linear algebra? And the calculations seem way over my head. Subsp a basis ce .

but i'll give it my best shot. that's a relief. the math behind it is actually very simple.. You said it! That's a tough question to answer properly.The calculations aren't nearly as hard as they look! i wouldn't say it's middle school level. Oh! Well. really? Why? err. really? And once you understand the basics.. Well... An Overview of Linear Algebra 15 . what is linear algebra exactly? Um.. but it's not far off. but i still don't understand.. it's pretty abstract stuff...

.... Matrices Vectors And vectors. Oh! We’ll learn to work with matrices. 16 Chapter 1 What is Linear Algebra? . Linear algebra is about translating something residing in an m-dimensional space into a corresponding shape in an n-dimensional space.From three to two dimensions From two to three dimensions From two to the same two dimensions broadly speaking. with the goal of understanding the central concepts of: • Linear transformations • eigenvalues and eigenvectors L in e a r tra n s fo r m atio n s eigenvalues and eigenvectors Vectors Matrices i see...

? You just had to ask me the dreaded question. such as earthquake-proofing architecture.. eh? While it is useful for a multitude of purposes indirectly. fighting diseases. protecting marine wildlife. Oh? And mathematicians and physicists are the only ones who are really able to use the subject to its fullest potential.. What exactly is it good for? Outside of academic interest.. Aww! An Overview of Linear Algebra 17 .So. didn't you? .. and generating computer graphics. it doesn't stand that well on its own. to be completely honest.. of course..

it won't do me any good in the end? That's not what i meant at all! For example. The same relationship holds for math and science students and linear algebra. lot of There is also a re that you aca demic literatu d if you don't won't understan ebra. it's just considered common knowledge. know linear alg i'll try! 18 Chapter 1 What is Linear Algebra? . Just buckle down and study. it's just one of those things you've got to know. he has to know how to fillet a fish. i see. Like it or not. we should all know how to do it... and you'll do fine.So even if i decide to study. best not to fight it. for an aspiring chef to excel at his job.

i'll try to avoid that as much as possible..and focus on explaining the basics as best i can.regarding our future lessons. Whew.. .... Which leads to. Vectors tra n Lin ea r sfo rm ation s eig en eig va lue env s ect a n d or s Matr ice s Most books and courses in the subject deal with long calculations and detailed proofs. i think it would be best if we concentrated on understanding linear algebra as a whole.. Great! An Overview of Linear Algebra 19 ...

Ri ng Oops. uh. “Come home now ” Oh? hard to imagine..Well. that's me! Ri ng R in g Ri ng The sensei texts?! it's from my brother. Well.. i don't want to overwhelm you. tell him i said hello! 20 Chapter 1 What is Linear Algebra? .. Okay.. Why don't we call it a day? We'll start with some of the fundamentals next time we meet.

2 The Fundamentals .

you seem pretty out of it today. Are you okay? reiji. i thought we'd start off with.97 You've got to know your basics! 98 99 Gghh Wh um 100… p Ddone! You wish! After you're done with the pushups.. i-i'll be fine... i want you to start on your legs! That means squats! Go go go! Hey.. take a look at thi— 22 Chapter 2 .

Don't worry.. Om n om n om i don't mind.R r r r u m b le Ah— Sorry.. Just give me five minutes. i guess i could use a snack.. take your time. Take a look at this. Well then. The Fundamentals 23 . let's begin. pushing your body that hard has its consequences..

you'll be fine. Course layout Fundamentals Matrices Vectors We'll start off slow and build our way up to the more abstract parts. okay? Don't worry.Course layout basics Fundamentals i took the liberty of making a diagram of what we're going to be talking about. Prep Matrices Vectors Wow! Main Linear Transformations eigenvalues and eigenvectors i thought today we'd start on all the basic mathematics needed to understand linear algebra. Sure. 24 Chapter 2 The Fundamentals .

.333.i've never really understood the meaning of i. . They're organized like this. real numbers integers rational numbers* (not integers) irrational numbers imaginary numbers • Positive natural numbers • 0 • Negative natural numbers • Terminating decimal numbers like 0... Let's talk about number systems first. Complex numbers.number Systems Complex numbers Complex numbers are written in the form a+b·i where a and b are real numbers and i is the imaginary unit. where b is a nonzero real number * numbers that can be expressed in the form q / p (where q and p are integers. • Numbers like  and √2 whose decimals do not follow a pattern and repeat forever • Complex numbers without a real component.. and p is not equal to zero) are known as rational numbers. integers are just special cases of rational numbers.3 • Non-terminating decimal numbers like 0.... i don't know for sure.. defined as i = √−1. like 0 + bi.. number Systems 25 . but i suppose some mathematician made it up because he wanted to solve equations like ? x2 + 5 = 0 Well..

. but complex numbers appear pretty frequently in a variety of areas. Thanks! 26 Chapter 2 The Fundamentals . these previously unsolvable problems suddenly became approachable. i'll just have to get used to them. i understand where you're coming from.. i suppose... Sigh Don't worry! i think it'd be better if we avoided them for now since they might make it harder to understand the really important parts. Using this new symbol. Why would you want to solve them in the first place? i don't really see the point.So.

ambiguous sentences that produce different reactions depending on whom you ask are not propositions. by the way." Umm Let's look at a few examples. “one plus one equals two” or “japan's population does not exceed 100 people. My mom says i’m the most handsome guy in school.. implication and equivalence 27 T T T T T F T F ..implication and equivalence Propositions i thought we'd talk about implication next. A proposition is a declarative sentence that is either true or false.. but a sentence like “reiji Yurino is handsome” is not.. T F That kind of makes sense. let’s discuss propositions.. like. but first.” 0 < 10 “That is either true or false. A sentence like “reiji Yurino is male” is a proposition. To put it simply. “reiji Yurino is female” is also a proposition..

i hope not! in situations where we know that “if P then Q” is true. “if this dish contains pork then it is a schnitzel” .it is no longer necessarily true. Pigs ’ feet! but if we look at its Converse. Yeah. like this: P⇒Q if P then Q i think i get it. This is a schnitzel This dish contains pork 28 Chapter 2 The Fundamentals . it is common to write it with the implication symbol.. we say that “P entails Q” and that “Q could entail P..” it is a schnitzel it is a schnitzel it contains pork Could entail it is a schnitzel Tr u e entails it contains pork n ne ce ot ss tr aril ue y it contains pork When a proposition like “if P then Q” is true... but don't know anything about its converse “if Q then P ”. The statement “if this dish is a schnitzel then it contains pork” is always true.implication Let's try to apply this knowledge to understand the concept of implication...

Then P and Q are equivalent. So it’s like the implication symbols point in both directions at the same time? And this is the symbol for equivalence. All right. implication and equivalence 29 .. You’re due for a growth spurt. Don’t worry. if both “if P then Q” and “if Q then P ” are true. P ⇒ Q as well as Q ⇒ P.equivalence That is. Tetsuo is taller than reiji. Tetsuo is taller than reiji.. reiji is shorter than Tetsuo. reiji is shorter than Tetsuo. exactly! it's kind of like this.

This might give you a good idea of what i mean. a set is a collection of things. Hehe. okay. Another important field of mathematics is set theory. Probably.. The things that make up the set are called its elements or objects.Set Theory Sets Oh yeah.i think we covered that in high school. but let's review it anyway. 30 Chapter 2 The Fundamentals . Slide Just as you might think..

consists of these four elements: • • • • Kagawa-ken1 Ehime-ken Kouchi-ken Tokushima-ken Kagawa Tokushima Ehime Kouchi example 2 The set consisting of all even integers from 1 to 10 contains these five elements: • • • • • 2 4 6 8 10 1. A Japanese ken is kind of like an American state. Set Theory 31 .example 1 The set “Shikoku.” which is the smallest of Japan’s four islands.

ehime-ken Shikoku 32 Chapter 2 ..Set Symbols e n nu m b e r s A ll ev e t w ee n 1 an d 1 0 b To illustrate. With that in mind. for example. X.” Okay. it's also convenient to give the set a name.. our definition now looks like this: X marks the set! This is a good way to express that “the element x belongs to the set X. For example. the set consisting of all even numbers between 1 and 10 would look like this: These are two common ways to write out that set: Mmm.

Shikoku Japan Set Theory 33 . Let's say that all elements of a set X also belong to a set Y.Subsets And then there are subsets. i see. Set Y (Japan) X is a subset of Y Hokkaidou Aomori-ken iwate-ken Miyagi-ken Akita-ken Yamagata-ken Fukushima-ken ibaraki-ken Tochigi-ken Gunma-ken Saitama-ken Chiba-ken Toukyou-to Kanagawa-ken niigata-ken Toyama-ken ishikawa-ken Fukui-ken Yamanashi-ken nagano-ken Gifu-ken Shizuoka-ken Aichi-ken Mie-ken Shiga-ken Kyouto-fu Oosaka-fu Hyougo-ken nara-ken Wakayama-ken Tottori-ken Shimane-ken Okayama-ken Hiroshima-ken Yamaguchi-ken Fukuoka-ken in this case. For example. Set X (Shikoku) Kagawa-ken ehime-ken Kouchi-ken Tokushima-ken Saga-ken nagasaki-ken Kumamoto-ken Ooita-ken Miyazaki-ken Kagoshima-ken Okinawa-ken And it's written like this.

10 } Y = { 4.. X is no longer a subset of Y. 4. example 2 X 4 10 Y 2 6 8 Suppose we switch the sets: X = { 2. 10 } Y = { 2.. 4. 8. 6. since all elements in X also exist in Y. 6. 4. So X is a subset of Y. example 3 X 2 6 8 X Y 4 10 Suppose we have two equal sets instead: X = { 2. and Y is a subset of X. 8.example 1 Suppose we have two sets X and Y: X = { 4. 8 } In this case neither X nor Y is a subset of the other. 4. both sets are subsets of each other. 10 } In this case. 10 } Y = { 4. example 4 Y 2 4 6 8 10 X 2 6 10 Y 4 8 Suppose we have the two following sets: X = { 2. Are you still hanging in there? You know it! 34 Chapter 2 The Fundamentals . 6. 10 } Y = { 2. 6. 6. i think we're about halfway done for today. 10 } X is a subset of Y. 8. 8. 10 } Since all elements in X don’t exist in Y.

but you'll be fine as long as you take your time and think hard about each new idea. Let's start by defining the concept itself. Got it.Functions i thought we'd talk about functions and their related concepts next. Sounds good. Fu n c ti o n s it's all pretty abstract. Functions 35 .

imagine the following scenario: Captain ichinose. decides to treat us freshmen to lunch. So we follow him to restaurant A. he gets a say in any and all orders. Udon 500 yen Curry 700 yen breaded pork 1000 yen broiled eel 1500 yen but there is a catch. in a pleasant mood. Kind of like this: ? What do you mean? 36 Chapter 2 The Fundamentals . Follow me! This is the restaurant menu. of course. Since he's the one paying.

right? Udon for everyone! Yurino Udon Yoshida Cu y Yajima Breaded pork Tomiyama Broiled l Or say.We wouldn't rea y be able to say no if he told us to order the cheapest dish. to order di Order erent stu ! Yurino Udon Yoshida Cu y Yajima Breaded pork Tomiyama Broiled l Functions 37 . if he just told us a di erent things.

but that doesn't change the fact that we have to obey him.Even if he told us to order our favorites. Order what you want! Yurino Udon Yoshida Cu y Yajima Breaded pork Tomiyama Broiled l You could say that the captain's ordering guidelines are like a “rule” that binds elements of X to elements of Y. we wouldn't rea y have a choice. X Rule! ? Y Yurino Udon ? Yoshida Cu y ? Yajima Breaded pork ? Tomiyama Broiled l 38 Chapter 2 The Fundamentals . This might make us the most ha y.

And that is why. Gotcha. Functions A rule that binds elements of the set X to elements of the set Y is called “a function from X to Y. just like the captain’s rules for how we order lunch! ! This is how we write it: rule Menu or Club member or rule Club member Menu f is completely arbitrary. g or h would do just as well..” X is usually called the domain and Y the co-domain or target set of the function. Functions 39 . function! We define a “function from X to Y ” as the rule that binds elements in X to elements in Y..

Okay! 40 Chapter 2 The Fundamentals . xi 's image under f in Y Also..” The element in Y that corresponds to x i when put through f. it's not uncommon to write “xi 's image under f in Y”..images Let’s assume that xi is an element of the set X. As f(xi).. images? is called “x i 's image under f in Y. next up are images..

. X f Y Yurino Udon Yoshida Cu y Yajima Breaded pork Tomiyama Broiled l Like this: i hope you like udon! f (Yurino) = udon f (Yoshida) = broiled eel f (Yajima) = breaded pork f (Tomiyama) = breaded pork image This is the element in Y that corresponds to xi of the set X.And in our case. when put through the function f.. Functions 41 .

yeah. i'm supposed to write f(2) and.they always used this weird symbol f(x) where they could have used something much simpler like y instead? “Like whatever! Anyways.by the way. Didn't you ever wonder why.... i have! M 42 Chapter 2 The Fundamentals ... sure. do you remember this type of formula from your high school years? ? Oh...” ide in s b r a i n ’s  isa Actually. so if i want to substitute with 2 in this formula.. ...

f(2) implies this: i think i'm starting to get it. really means is this: f(x) = 2x − 1 What The function f is a rule that says: “The element x of the set X goes together with the element 2x − 1 in the set Y. So we were using functions in high school too? The image of 2 under the function f is 2 ∙ 2 − 1.” Oh! So that's what it meant! Similarly. here’s why.Well. Functions 43 . exactly.

Domain and range

X
On to the next subject.

f

Y

Yurino

Udon

in this case...

Yoshida

Cu

y

Yajima

Breaded pork

Tomiyama

Broiled

l

We're going to work with a set {udon, breaded pork, broiled eel} which is the image of the set X under the function f. * Udon broiled eel breaded pork

This set is usually called the range of the function f, but it is sometimes also called the image of f.

Kind of confusing...

44

Chapter 2

* The term image is used here to describe the set of elements in Y that are the image of at least one element in X.

X
And the set X is denoted as the domain of f.
Yurino

Domain

Y
Range

co-Domain

Udon Yoshida
Breaded pork

Yajima
Broiled l

Tomiyama

Cu

y

We could even have described this function as

Y = { f (Yurino), f (Yoshida), f (Yajima), f (Tomiyama)}
if we wanted to.

Hehe.

u! Oss

range and Co-domain The set that encompasses the function f ’s image { f(x1), f(x2), … , f(xn)} is called the range of f, and the (possibly larger) set being mapped into is called its co-domain. The relationship between the range and the co-domain Y is as follows:

{ f(x1), f(x2), … , f(xn)} ⊂ Y
In other words, a function’s range is a subset of its co-domain. In the special case where all elements in Y are an image of some element in X, we have

{ f(x1), f(x2), … , f(xn)} = Y
Functions 45

Onto and One-to-One Functions

next we’ll talk about onto and one-to-one functions.

right.

Let's say our karate club decides to have a practice match with another club...

And that the captain’s mapping function f is “Fight that guy.”

Hanamichi University

University A

X

f

Hanamichi University

University B

Y

X

f

Hanamichi University

University C

Y

X

f

Y

Yurino

Yurino

Yurino

Yoshida

Yoshida

Yoshida

Yajima

Yajima

Yajima

Tomiyama

Tomiyama

Tomiyama

You're already doing practice matches?

n-not really. This is just an example.

Still working on the basics!

46

Chapter 2

The Fundamentals

Onto Functions
Hanamichi University University A

X
Yurino

f

Hanamichi University

University B

Y

X
Yurino

f

Y

Yoshida

Yoshida

A function is onto if its image is equal to its co-domain. This means that a the elements in the co-domain of an onto function are being ma ed onto.

Yajima

Yajima

Tomiyama

Tomiyama

One-to-One Functions
Hanamichi University University A

X

f

Y

Hanamichi University

University C

X

f

Y

Yurino

Yurino

Yoshida

Yoshida

If xi ≠ xj leads to f(xi) ≠ f(xj), we say that the function is oneto-one. This means that no element in the co-domain can be ma ed onto more than once.

Yajima

Yajima

Tomiyama

Tomiyama

One-to-One and Onto Functions
Hanamichi University University A

X

f

Y

Yurino

Yoshida

It’s also po ible for a function to be both onto and one-to-one. Such a function creates a “bu y system” betw n the elements of the domain and co-domain. Each element has one and only one “partner.”

Yajima

Tomiyama

Functions

47

inverse Functions
Han This time we're Univamichi ersity going to look at X the other team captain's orders as well.
Yurino

University A

g

now we have inverse functions.

Y

Yoshida

Yajima

inverse?

Tomiyama

Hanamichi University

University A

X

f

Hanamichi University

University A

Y

X

g

Y

Yurino

Yurino

Yoshida

Yoshida

Yajima

Yajima

Tomiyama

Tomiyama

i see.

We say that the function g is f 's inverse when the two captains' orders coincide like this.

48

Chapter 2

The Fundamentals

right? The function f has an inverse. f is an inverse of g if these two relations hold.. and vice versa. There is also a connection between one-to-one and onto functions and inverse functions.To specify even further. Got it! Functions 49 .. it’s like the functions undo each other! This is the symbol used to indicate inverse functions. Have a look at this. You raise it to –1. Oh. The function f is one-to-one and onto. it has an inverse. or So if it’s one-toone and onto.

Linear transformations? basics Fundamentals We're already there? Prep Matrices Oh right. one of the main subjects. it's going to get pretty abstract from now on! O-okay! 50 Chapter 2 The Fundamentals . Main Linear Transformations eigenvalues and eigenvectors no.Linear Transformations i know it's late. into more We'll go r on. but i'd also like to talk a bit about linear transformations if you're okay with it. detail late but don't be fooled and let your guard down. we're just going to have a quick look for now.

.. Functions 51 . so that means. f is called a linear transformation from X to Y if it satisfies the following two conditions: u v f(xi)+ f(xj) = f(xi + xj) cf(xi) = f(cxi) Hmm.. and f be a function from X to Y.linear transformations! Gah. Linear Transformations Let xi and xj be two arbitrary elements of the set X. i think we'd better draw a picture.. This should clear things up a bit. That’s a little easier to understand... c be any real number.. What do you say? And condition v is when the product of this and a scalar c equals this...behold. Condition u is when the sum of these two equals this..

. as you can see in the table below. Yeah. today . reiji. . 52 Chapter 2 The Fundamentals P h ew My pleasure. This is because it satisfies both  and .. f(xi) + f(xj) = (2xi − 1) + (2xj − 1) = 2xi + 2xj − 2 f(xi + xj) = 2(xi + xj) − 1 = 2xi + 2xj − 1 Condition  Condition  cf(xi) = c(2xi − 1) = 2cxi − c f(cxi) = 2(cxi) − 1 = 2cxi − 1 Are you all right? Thanks.Let's have a l k at a couple of examples. This is because it satisfies neither  nor . don't worry! That's all for . An example of a linear transformation The function f(x) = 2x is a linear transformation. as you can see in the table below. f(xi) + f(xj) = 2xi + 2xj f(xi + xj) = 2(xi + xj) = 2xi + 2xj Condition  Condition  cf(xi) = c(2xi) = 2cxi f(cxi) = 2(cxi) = 2cxi An example of a function that is not a linear transformation The function f(x) = 2x − 1 is not a linear transformation.

Um... next time we meet you won't have to.. yeah. reiji? Do you always eat lunch in the school cafeteria? i live alone.. and i'm not that good at cooking. i'm making you lunch! Don't worry about it! Tetsuo is still helping me out and all! You don't want me to? Functions 53 . so most of the time. Well.

.no. you know—stamina lunches.... i’d love for you to... On second thought..Make me lunch. Great! . that's not it.How lovely.. Uh.. i make a lot of them for my brother too.. it's just.. Oh.

F. B. C. But he couldn’t take all of the songs. and finally I’ll present a  ! Solution . Let’s call the songs A. and G. Questions: 1. 2. since his taste in music wasn’t very compatible with Nemoto’s. ? Problem Reiji bought a CD with seven different songs on it a few days ago. these are the first two questions. then I’ll establish a good * way of  thinking . he decided to make a new CD with only three songs on it from the original seven.Combinations and Permutations I thought the best way to explain combinations and permutations would be to give a concrete example. 2. Choose an order in which to play them. while packing for a car trip he had planned with his friend Nemoto. where three songs are chosen from a pool of seven? Combinations and Permutations 55 . Choose three songs out of the seven possible ones. In how many ways can Reiji select three songs from the original seven? In how many ways can the three songs be arranged? In how many ways can a CD be made. The solution to question 3. The following day. 3. is as follows: Solution to Question 1 · Solution to Question 2 = Solution to Question 3 In how many ways can Reiji select three songs from the original seven? In how many ways can the three songs be arranged? In how many ways can a CD be made. it struck him that it might be nice to take the songs along to play during the drive. E. I’ll start by explaining the  ? Problem . then. D. where three songs are chosen from a pool of seven? * Way of Thinking It is possible to solve question 3 by dividing it into these two subproblems: 1. As you may have realized. After some deliberation.

7 3 = 35 56 Chapter 2 The Fundamentals .! Solution 1. Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 A A A A A A A A A A A A A A A and and and and and and and and and and and and and and and B and C B and D B and E B and F B and G C and D C and E C and F C and G D and E D and F D and G E and F E and G F and G Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern Pattern 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 B and C and D B and C and E B and C and F B and C and G B and D and E B and D and F B and D and G B and E and F B and E and G B and F and G C and D and E C and D and F C and D and G C and E and F C and E and G C and F and G D and E and G D and E and G D and F and G E and F and G Choosing k among n items without considering the order in which they are chosen is called a combination. In how many ways can Reiji select three songs from the original seven? All 35 different ways to select the songs are in the table below. The number of different ways this can be done is written by using the binomial coefficient notation: n k which is read “n choose k. Feel free to look them over.” In our case.

2. When we’re picking our second song. and G instead: Song 1 B B E E G G Song 2 E G B G B E Song 3 G E G B E B Trying a few other selections will reveal a pattern: The number of possible arrangements does not depend on which three elements we choose—there are always six of them. Here’s why: Our result (6) can be rewritten as 3 · 2 · 1. 3. We start out with all three songs and can choose any one of them as our first song. B. This table illustrates the 6 different ways in which they can be arranged: Song 1 A A B B C C Song 2 B C A C A B Song 3 C B C A B A Suppose we choose B. and C. which we get like this: 1. This gives us 3 possibilities · 2 possibilities · 1 possibility = 6 possibilities. For our last song. In how many ways can the three songs be arranged? Let’s assume we chose the songs A. 2. E. we’re left with only one choice. only two remain to choose from. Combinations and Permutations 57 .

.3. Choosing three from seven items in a certain order creates a permutation of the chosen items.. we could use this if we wanted to know how many different ways seven objects can be arranged. In how many ways can a CD be made. The number of possible permutations of k objects chosen among n objects is written as n Pk In our case. where three songs are chosen from a pool of seven? The different possible patterns are The number of ways to choose three songs from seven = 7 3 6 The number of ways the three songs can be arranged = 35 6 = 210 This means that there are 210 different ways to make the CD. The answer is 7! = 7 · 6 · 5 · 4 · 3 · 2 · 1 = 5040 58 Chapter 2 The Fundamentals . · 2 · 1 For instance. this comes to 7P3 = 210 The number of ways n objects can be chosen among n possible ones is equal to n-factorial: n Pn = n! = n · (n − 1) · (n − 2) · .

F B . B. B.. F. Pattern Pattern 1 2 3 30 31 60 61 90 91 120 121 150 151 180 181 209 210 A A A ... A B . Pattern Pattern . D. F G . G A . D. F B .... F B ..... B C ... F C ..... Combinations and Permutations 59 .. E F .. We can choose any of the 7 songs A.. analogous to the previous example. rewrite our problem of counting the different ways in which to make a CD as 7 ∙ 6 ∙ 5 = 210... Song 1 Pattern Pattern Pattern .. G A . And finally we choose any of the now 5 remaining songs as our last song....... 3.I’ve listed all possible ways to choose three songs from the seven original ones (A.. C.. E. and G) in the table below. Pattern Pattern ... E F Song 3 C D E ... G A .... Here’s how we get those numbers: 1....... F B . 2. C. F E We can.. and G as our first song. G A . Pattern Pattern . E.. G A . Pattern Pattern . F... C D ... Pattern Pattern ... We can then choose any of the 6 remaining songs as our second song.. E B . G G Song 2 B B B . Pattern Pattern ... D E . G A ...

The definition of the binomial coefficient is as follows:  n  n ⋅ ( n − 1)  ( n − ( r − 1) ) n ⋅ ( n − 1)  ( n − r + 1) =  = r ⋅ ( r − 1)1 r ⋅ ( r − 1)1 r  Notice that  n  n ⋅ ( n − 1)  ( n − ( r − 1) )  = r ⋅ ( r − 1)1 r  (n − r ) ⋅ (n − r + 1)1 r ⋅ ( r − 1)1 (n − r ) ⋅ (n − r + 1)1 n ⋅ ( n − 1)( n − ( r − 1) ) ⋅ ( n − r ) ⋅ ( n − r + 1)1 = (r ⋅ (r − 1)1) ⋅ ( (n − r ) ⋅ (n − r + 1)1) = ⋅ = n! r! ⋅ ( n − r ) ! n ⋅ ( n − 1)  ( n − ( r − 1) ) Many people find it easier to remember the second version: n  n!  =  r  r! ⋅ ( n − r ) ! We can rewrite question 3 (how many ways can the CD be made?) like this: 7 7 7 7! 7! 7 ⋅ 6 ⋅ 5 ⋅ 4 ⋅ 3 ⋅ 2 ⋅ 1 P3 =   ⋅ 6 =   ⋅ 3! = = 7 ⋅ 6 ⋅ 5 = 210 ⋅ 3! = = 3! ⋅ 4! 4! 4 ⋅ 3 ⋅ 2 ⋅1 3 3 60 Chapter 2 The Fundamentals .

It is important to note. because there are several different ways to obey that command. Yurino Udon Yurino Udon Yoshida Cu y Yoshida Cu y Yajima Breaded pork Yajima Breaded pork Tomiyama Broiled l Tomiyama Broiled l not All “rules for Ordering” Are Functions 61 . that “Order different stuff!” isn’t actually a function in the strictest sense.not All “rules for Ordering” Are Functions We talked about the three commands “Order the cheapest one!” “Order different stuff!” and “Order what you want!” as functions on pages 37–38. however.

.

3 intro to Matrices .

ei! ei! Yurino! Put your backs into it! Don't rely on your hands. heheh i thought he'd quit right away. Use your waist! Ossu! i guess i was wrong.. 64 Chapter 3 ..

thank you... intro to Matrices 65 .i could never eat something so beautiful! Hehe..d a ! You must be really tired after all that exercise! Wow! but..thank you! Om m no no m no m Awesome! So good! Thanks.. don't be silly. really.Ta . Misa. Joy i don't know what to say. Don't worry about it..

i don't think you'll have any problems with the basics this time around either.Ah.. Are you ready to begin? Sure. why not. bounded by parentheses. and those can be a bit tricky. row M 66 Chapter 3 intro to Matrices . Course layout Fundamentals Pre Matrices Vecto Main And i'd really like to take my time on this one since they appear in most parts of linear algebra. basics Linear Transformations eigenvalues eigenvectors Okay. like this. A matrix is a collection of numbers arranged in m rows and n columns. We'll be talking about matrices today.. i feel a lot better now. What is a Matrix? row 1 row 2 column 1 column 2 column n These are called subscripts. but i'll talk a little about inverse matrices toward the end.

" 2×3 matrix 4×1 matrix m ×n matrix Ah... Square matrix with n rows The grayed out elements in this matrix are part of what is called its main diagonal. i've marked the (2. col 1 col col col 2 3 1 row 1 row 2 row 3 row 4 row m row 1 row 2 col 1 col 2 i see.A matrix with m rows and n columns is called an “m by n matrix. col n Element row 1 row 2 A matrix that has an equal number of rows and columns is called a square matrix. What is a Matrix? 67 . The items inside a matrix are called its elements. 1) elements of these three matrices for you. Square matrix with two rows Uh huh.

these are some of the advantages. Well. Um.Hmm... exciting or not. Just numbers.. they also help to make mathematical literature more compact. huh? Yep.. Yeah. Matrices aren't as exciting as they seem in the movies. Why is that? So people use them because they're practical.. no Keanu. matrices are very useful! • They're great for writing linear systems more compactly. 68 Chapter 3 intro to Matrices . • Since they make for more compact systems. • And they help teachers write faster on the blackboard during class..

........ + a1nxn • a21x1 + a22x2 + ... + amnxn = bm a11x1 + a12x2 + ........ becomes this? exactly! not bad! Writing Systems of Equations as Matrices a11x1 + a12x2 + . ch Skrit ch Skrit it does look a lot cleaner................. + amnxn What is a Matrix? 69 . + a2nxn .......... is written a11 a21 am1 a11 a21 am1 a12 a22 am2 a12 a22 am2 a1n a2n amn a1n a2n amn x1 x2 xn x1 x2 xn = b1 b2 bm am1x1 + am2x2 + ............. using matrices................ We could write it like this..................... + a2nxn = b2 ... So this........ is written am1x1 + am2x2 + ......instead of writing this linear system like this........... + a1nxn = b1 • a21x1 + a22x2 + ...

10) + (−3. • (10. 10 + (−6)) = (7. −6) = (10 + (−3). 4) 10 10 −3 −6 10 + (−3) 7 = 10 + (−6) 4 • + = 70 Chapter 3 intro to Matrices . • Subtraction • Scalar multiplication • Matrix Multiplication A ition Addition 1 Let's a the 3×2 matrix 2 4 6 6 5 3 1 5 3 4 2 3 5 to this 3×2 matrix 1 That is: 2 6 3 5 4 + 4 The elements would be a elementwise.Matrix Calculations The four relevant operators are: • Addition now let's look at some calculations. like this: ed 6 1+6 3+4 5+2 2 1 2+5 4+3 6+1 examples Examples 1 • 2 6 6 2 5 1 1+6 5+2 2+5 6+1 7 7 7 7 7 3 5 4 + 4 3 = 3+4 4+3 = 7 Note that A ition and subtraction work only with matrices that have the same dimensions.

−6) = (10 − (−3). 16) 10 10 − −3 −6 10 − (−3) 13 = 10 − (−6) 16 • = Matrix Calculations 71 . 10 − (−6)) = (13. 3−4 like this: 5−2 Examples examples • 1−6 3 4 − 4 3 = 3−4 5 6 2 1 5−2 1 2 6 5 2−5 4−3 6−1 −5 −3 = −1 3 1 5 • (10.Subtraction Subtraction 6 5 Let's subtract the 3×2 matrix 4 3 2 1 1 2 from this 3×2 matrix 3 4 5 6 1 2 6 5 2 1 2−5 4−3 6−1 That is: 3 4 − 4 3 5 6 The elements would similarly 1 − 6 be subtracted elementwise. 10) − (−3.

1) = (2 · 3. like this: examples Examples 1 2 • 10 · 1 10 · 2 10 20 10 3 4 = 10 · 3 10 · 4 = 30 40 50 60 5 6 10 · 5 10 · 6 • 2 (3. 2 · 1) = (6. That is: 10 3 4 5 6 10 · 1 10 · 2 10 · 3 10 · 4 10 · 5 10 · 6 The elements would each be multiplied by 10. 2) 3 1 2·3 2·1 6 2 • 2 = = 72 Chapter 3 intro to Matrices .Scalar Multiplication Scalar Multiplication 1 2 Let's multiply the 3×2 matrix 3 4 5 6 1 2 by 10.

forming the two products 1 3 5 2 4 6 x1 x2 1x1 + 2x2 = 3x1 + 4x2 5x1 + 6x2 and 1 3 5 2 4 6 y1 y2 1y1 + 2y2 = 3y1 + 4y2 5y1 + 6y2 and then rejoining the resulting columns: 1x1 + 2x2 3x1 + 4x2 5x1 + 6x2 1y1 + 2y2 3y1 + 4y2 5y1 + 6y2 example Example 1 2 • 3 4 5 6 x1 x2 y1 y2 1x1 + 2x2 = 3x1 + 4x2 5x1 + 6x2 1y1 + 2y2 3y1 + 4y2 5y1 + 6y2 There's more! Matrix Calculations 73 .Matrix Multiplication Matrix Multiplication 1 2 The product 3 4 5 6 x1 y1 x2 y2 1x1 + 2x2 = 3x1 + 4x2 5x1 + 6x2 1y1 + 2y2 3y1 + 4y2 5y1 + 6y2 Can be derived by temporarily separating the two columns x1 x2 and y1 y2 .

As you can s from the example below. Matrices can be multiplied only if the number of columns in the left factor matches the number of rows in the right factor. an m×n matrix times an n×p matrix yields an m×p matrix. • 8 2 −3 3 1 1 1 2 = 8 · 3 + (−3) · 1 2·3+1·1 8 · 1 + (−3) · 2 2·1+1·2 = 24 − 3 6+1 8−6 2+2 = 21 7 2 4 • 3 1 1 2 8 2 −3 1 = 3·8+1·2 1·8+2·2 3 · (−3) + 1 · 1 1 · (−3) + 2 · 1 = 24 + 2 −9 + 1 8 + 4 −3 + 2 = 26 −8 12 −1 And you have to watch out. changing the order of factors usua y results in a completely di erent product. 74 Chapter 3 intro to Matrices .

Well. We run into a problem here: there are no elements corresponding to these positions! Oops.. really? is the same as and which is the same as Product of 3×2 and 2×2 factors in the same matrix.This means we wouldn't be able to calculate the product if we switched the two matrices in our first example. One more thing. it's okay to use exponents to express repeated multiplication of square matrices.. nothing stops us from trying. Product of 2×2 and 3×2 factors is the same as and which is the same as and in the same matrix. factors Matrix Calculations 75 . Huh.

but how am i supposed to calculate three of them in a row? y prett g usin co n f Well. so.Oh. Um. Like this: Oh... The easiest way would be to just multiply them from left to right. This is all right then? Yeah... Like this. of course! 1 3 7 15 2 4 3 = 1 3 1 3 2 4 2 4 = 1 3 2 4 1 3 2 4 = 1·1+2·3 3·1+4·3 7 · 2 + 10 · 4 15 · 2 + 22 · 4 1·2+2·4 3·2+4·4 = 37 81 54 118 1 3 2 4 10 22 7 · 1 + 10 · 3 15 · 1 + 22 · 3 76 Chapter 3 intro to Matrices ...

Okay!  Zero Matrices Zero matrices ① ①①①① A zero matrix is a matrix where all elements are equal to zero.. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Special Matrices 77 . Diagonal matrices identity matrices inverse matrices Let's look at them in order.Special Matrices So we'll look at only these eight today.. Transpose matrices Symmetric matrices Upper triangular matrices Lower triangular matrices To explain them all would take too much time. Zero matrices There are many special types of matrices.

T for transpose. If we transpose the 2×3 matrix 1 we get the 3×2 matrix 3 5 2 4 6 1 2 3 4 5 6 As you can see. a11 a21 am1 The transpose of the n×m matrix a12 a1n a11 is consequently a21 am1 a12 a22 am2 a1n a2n amn a22 a2n am2 amn The most common way to indicate a transpose is to add a small T at the top-right corner of the matrix. TransposeMatrices Transpose matrices The easiest way to understand transpose matrices is to just look at an example. the transpose operator switches the rows and columns in a matrix. i see. 1 = 3 5 2 4 6 78 Chapter 3 intro to Matrices . a11 a21 am1 For example: 1 2 3 4 5 6 T a12 a22 am2 a1n a2n amn T Ah.

1 This is an upper triangular matrix. since all elements below the main diagonal are zero. 0 0 0 1 This is a lower triangular matrix—all elements above the main diagonal are zero. a symmetric matrix is always equal to its transpose.  U er Triangular and Upper triangular and  lowerTriangular Matrices Lower triangular matrices Triangular matrices are square matrices in which the elements either above the main diagonal or below it are all equal to zero. Sy etric Matrices Symmetric matrices Symmetric matrices are square matrices that are symmetric around their main diagonals. 1 5 6 7 5 2 8 9 6 8 3 10 7 9 10 4 Because of this characteristic. 5 6 7 5 2 0 0 0 2 8 9 6 8 0 0 0 3 10 7 9 4 0 0 0 4 3 10 Special Matrices 79 .

1 For example. Why? 80 Chapter 3 intro to Matrices . Diagonal Matrices Diagonal matrices A diagonal matrix is a square matrix in which all elements that are not part of its main diagonal are equal to zero.4). Multiplying diagonal matrices by themselves is really easy. Note that this matrix could also be written as diag(1. 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 is a diagonal matrix.2.3.

huh? 2 0 p 0 3 p Special Matrices 81 .. 2 0 2 2 3 0 2 02 0 2 0 3 30 30 and 30 3 to see why. Try calculating Hmm.See for yourself! Uh. 2 0 2 0 p p p p 0 2 02 0 2 0 = p = p 3 0 30 3 0 3 0 2 3 2 0 0 3 3 Like this? You're right! 2 0 0 3 p = Weird....

they are square matrices with n rows in which all elements on the main diagonal are equal to 1 and all other elements are 0..1).1. just like you said! 82 Chapter 3 intro to Matrices .. In other words.. What do you mean? it's like the number 1 in ordinary multiplication. it stays the same. For example.1. Unchanged Try multiplying 1 0 0 1 x1 x2 if you'd like.. an identity matrix with n = 4 would look like this: 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 Multiplying with the identity matrix yields a product equal to the other factor. IdentityMatrices identity matrices Identity matrices are in essence diag(1.

Let's try a few other examples. 1 • 0 1 0 0 0 1 x1 x2 xn = 1 · x1 + 0 · x 2 + 0 · x1 + 1 · x 2 + 0 · x1 + 0 · x 2 + xn1 xn2 + 0 · xn + 0 · xn + 1 · xn = x1 x2 xn 1 · xn1 + 0 · xn2 0 · xn1 + 1 · xn2 0 0 • 1 0 0 1 x11 x21 x12 x22 = 1 · x11 + 0 · x12 0 · x11 + 1 · x12 x11 x21 x12 x22 xn1 xn2 1 · x21 + 0 · x22 0 · x21 + 1 · x22 = x11 • x12 x22 xn2 1 0 0 1 = x11 · 1 + x12 · 0 x21 · 1 + x22 · 0 xn1 · 1 + xn2 · 0 x11 · 0 + x12 · 1 x21 · 0 + x22 · 1 xn1 · 0 + xn2 · 1 = x11 x21 xn1 x12 x22 xn2 x21 xn1 Were you able to follow? Want another look? no way! Piece of cake! Special Matrices 83 .

Let's take a break. Fine by me. but they're a bit more complex than this. i wasn't trying to get you to— Thanks again for lunch. We still have inverse matrices left to look at. A h! no problem! i'll make you another tomorrow if you'd like. . no. What do you say? ready for more matrices? Sure. it'd be a pleasure. Th-thanks. the best part of cooking something is seeing someone enjoy it.. Don't worry about it. i had no idea you were such a good cook..

.. And sw p ... .4 More Matrices Gather 'em up .

explaining them without using examples isn't all that easy. but let's start with the definition anyway. then the two factor matrices are inverses of each other. This means that x11 x21 x12 x22 x12 x22 is an inverse matrix to 1 2 3 4 if 1 3 2 4 x11 x21 = 1 0 0 1 Ooh.inverse Matrices inverse matrices are very important and have a lot of different applications. 86 Chapter 4 More Matrices .  Inverse matrices If the product of two square matrices is an identity matrix.

And that's it. Should we get right down to business? Sure! inverse Matrices 87 . Didn't you say something about examples? We're done already? Don't worry. i thought we'd go into more detail on this one. that was only the definition! Huh? Since they're so important. i'll teach you how to identify whether an inverse exists or not—and also how to calculate one.

Let's have a look at that. Gaussian elimination can also be used to solve linear systems.Calculating inverse Matrices Ga u elim ssian inat io n There are two main ways to calculate an inverse matrix: Using Cofactors or using Gaussian elimination. Shh Sh shh Cool! * The japanese term for Gaussian elimination is Hakidashihou. Gaussian elimination is easy both to understand and to calculate.. in contrast. ignore it as long as you're not expecting it on a test. Anyway. hod met r cto ofa C Can do. so. Gotcha. in fact. which roughly translates to ”The sweeping out method. cofactor method The calculations involved in the cofactor method can very easily become cumbersome. i won't talk about cofactors at all today.. it's as easy as sweeping the floor!* in addition to finding inverse matrices.” Keep this in mind as you're reading this chapter! .

5 1 0 2 x1 x2 = 2 0 Gather 'em up and sweep. Hmm. 3x1 + 1x2 = 1 1x1 + 2x2 = 0 Start by multiplying the top equation by 2..gaussian elimination ? Problem ? Problem Solve the following linear system: 3x1 + 1x2 = 1 1x1 + 2x2 = 0 ! Solution Solution The co on method The co on method expre ed with matrices Gau ! K p comparing the rows on the left to s how it works. 6x1 + 2x2 = 2 1x1 + 2x2 = 0 Subtract the bottom equation from the top equation.. 5 1 0 2 2 0 5x1 + 0x2 = 2 5x1 + 10x2 = 0 Subtract the top equation from the bottom equation. ian elimination OKay. 5 0 x1 x2 0 10 = 2 −2 5 0 0 10 2 −2 2 5 1 − 5 2 5 1 0x1 + 1x2 = − 5 And we’re done! 1x1 + 0x2 = 1 0 0 1 x1 x2 2 5 = 1 − 5 1 0 0 1 Done! So you just rewrite the equations as matrices and calculate as usual? Well.. Gaussian elimination is about trying to get this part here to approach the identity matrix. 89 Calculating inverse Matrices . not about solving for variables.. 5 0 x1 x2 5 10 = 2 0 5 5 0 10 2 0 5x1 + 0x2 = 2 0x1 + 10x2 = −2 Divide the top equation by 5 and the bottom by 10. 3 1 1 2 x1 x2 = 1 0 3 1 1 2 1 0 Gather 'em up and sweep. 6 1 2 2 x1 x2 = 2 0 6 1 2 2 2 0 5x1 + 0x2 = 2 1x1 + 2x2 = 0 Multiply the bottom equation by 5.

Skr it skr ch itch We’re trying to find the inverse of 3 1 1 2 Ah. Okay. Problem ? Problem ? Find the inverse of the 2×2 matrix 3 1 1 2 Think about it like this.Let's try to find an inverse next. 3 1 1 2 x11 x12 x21 x22 = 1 0 0 1  x11 x12 x21 x22 We need to find the matrix that satisfies  3 1 or x11 x21 and x12 x22 that satisfy 1 2 3 1 1 2 x11 x21 x12 x22 = 1 0 0 1 = 3x11 + 1x21 = 1 1x11 + 2x21 = 0  3x12 + 1x22 = 0 1x12 + 2x22 = 1 We need to solve the systems and Let's do the math. 90 Chapter 4 More Matrices . right.

we’re done! 1x11 + 0x21 = 1 0 0 1 x11 x21 2 1 − x12 5 5 = 1 3 x22 − 5 5 1 0 0 1 2 1 − 5 5 1 3 − 5 5 Done.. So the inverse we want is 2 5 − 1 5 − 1 5 3 5 Yay! That was a lot easier than i thought it would be. 2 1 1x12 + 0x22 = − 5 5 1 3 0x11 + 1x21 = − 0x12 + 1x22 = 5 5 This is our inverse matrix.... 5x11 + 0x21 = 2 0x11 + 10x21 = −2 5x12 + 0x22 = −1 0x12 + 10x22 = 6 5 0 0 10 x11 x21 x12 2 −1 = x22 −2 6 5 0 0 10 2 −2 −1 6 Divide the top by 5 and the bottom by 10. Great. Calculating inverse Matrices 91 . but. 5x11 + 0x21 = 2 5x11 + 10x21 = 0 5x12 + 0x22 = −1 5x12 + 10x22 = 5 5 0 5 10 x11 x21 x12 2 −1 = x22 0 5 Huff 5 0 5 10 2 0 −1 5 Subtract the top equation from the bottom. 6x11 + 2x21 = 2 1x11 + 2x21 = 0 6x12 + 2x22 = 0 1x12 + 2x22 = 1 6 1 2 2 x11 x21 x12 2 = x22 0 0 1 6 1 2 2 2 0 0 1 Huff Subtract the bottom equation from the top. 5x11 + 0x21 = 2 1x11 + 2x21 = 0 5x12 + 0x22 = −1 1x12 + 2x22 = 1 5 1 0 2 x11 x21 x12 2 −1 = x22 0 1 5 1 0 2 2 0 −1 1 Multiply the bottom equation by 5.! ! Solution The co on method The co on method expre ed with matrices Gau ian elimination 3x11 + 1x21 = 1 1x11 + 2x21 = 0 3x12 + 1x22 = 0 1x12 + 2x22 = 1 3 1 1 2 x11 x21 x12 1 = x22 0 0 1 3 1 1 2 1 0 0 1 Multiply the top equation by 2.

. Chapter 4 More Matrices . the product is always the identity matrix! Remembering this test is very useful. so. The symbol used to denote inverse matrices is the same as any inverse in mathematics.... The product of the original and inverse matrix is 3 • 1 2 1 2 5 1 − 5 − 1 5 3 5 2 +1· 5 = 2 1· +2· 5 3· 1 5 1 − 5 − 1 3 +1· 5 5 1 3 1· − +2· 5 5 3· − 1 = 0 1 0 The product of the inverse and original matrix is 2 5 1 − 5 − 1 5 3 5 3 1 1 = 2 2 ·3+ 5 1 − ·3+ 5 − 1 ·1 5 3 ·1 5 2 ·1+ 5 1 − ·1+ 5 − 1 ·2 5 3 ·2 5 1 = 0 1 0 • It s ms like they both become the identity matrix. got it... the inverse of a11 a12 a1n a21 an1 92 is wri a11 a12 a21 an1 a22 an2 en as a1n a2n ann −1 a22 an2 a2n ann To the power of minus one. You should use it as often as you can to check your calculations. by the way.Let's make sure that the product of the original and calculated matrices rea y is the identity matrix. That’s an important point: the order of the factors doesn't ma er.

.. well..this formula right here.. Calculating inverse Matrices 93 .... Huh? Let's apply the formula to our previous example: Why even bother with the other method? 3 1 1 2 We got the same answer as last time. if you want to find the inverse of a bigger matrix. i'm afraid you're going to have to settle for Gaussian elimination.. This formula only works on 2×2 matrices... Hmm That's too bad. a21 a22 . Ah.Actually.we also could have solved −1 a11 a12 with.

some matrices lack an inverse? Yeah..... also invertible. of course. Oh. i thought i'd show you how to determine whether a matrix has an inverse or not. invertible One last thing: the inverse of an invertible matrix is. So. not invertible Makes sense! 94 Chapter 4 More Matrices . the denominator becomes zero. Try to calculate the inverse of this one with the formula i just showed you.next. 3 1 6 2 −1 Let's see. i guess you're right.

like this: We'll be using this function. det? it's short for determinant. it's also written with straight bars. .Determinants now for the test to see whether a matrix is invertible or not. Hmm. Does a given matrix have an inverse? a11 det a21 an1 a12 a22 an2 a1n a2n ann ≠0 means that a11 a21 an1 a12 a22 an2 a1n a2n ann −1 exists. The inverse of a matrix exists as long as its determinant isn't zero.

just substitute the expression like this. Which one's best depends on the size of the matrix. To find the determinant of a 2×2 matrix. Oh. Holding your fingers like this makes for a good trick to remember the formula.Calculating Determinants Let's start with the formula for twodimensional matrices and work our way up. Sounds good. cool! 96 Chapter 4 More Matrices . There are several different ways to calculate a determinant.

a21) The point (a12. Incidenta y. • • • • The origin The point (a11. since det 3 0 0 2 ≠ 0..coincides with the absolute value of det a11 a21 a12 a22 3 0 2 y y So l 0 ks like this? a22 2 6 a21 O a12 a11 x O 3 x Calculating Determinants 97 . det 3 0 0 2 =3·2−0·0=6 It does. a22) The point (a11 + a12. the area of the para elogram spa ed by the fo owing four points..Let's s whether 3 0 0 2 has an inverse or not.. a21 + a22) ..

i'm supposed to memorize this? Don't worry. third column. giving you a total of five columns. This is sometimes called Sarrus' Rule. and it’s much easier to remember! a11 a21 a31 a12 a22 a32 a13 a23 a33 a11 a21 a31 a12 a22 a32 P h ew 98 Chapter 4 More Matrices . Add the products of the diagonals going from top to bottom (indicated by that's Well. from bottom to top (indicated by dotted lines). Sarrus’ rule Write out the matrix. the solid lines) and subtract the products of the diagonals going a relief.To find the determinant of a 3×3 matrix. and then write its first two columns again after the Check it out. just use the following formula. there's a nice trick for this one too. This will generate the formula for Sarrus’ Rule.

a21 + a23. • • • • • • • • The origin The point (a11. a31 + a32 + a33) .also coincides with the absolute value of a11 det a21 a31 a12 a22 a32 a13 a23 a33 Each pair of opposite faces on the parallelepiped are parallel and have the same area. a22. a21 + a22. * A parallelepiped is a three-dimensional figure formed by six parallelograms. a31 + a32) The point (a11 + a13. a21 + a22 + a23.. a31 + a33) The point (a12 + a13. a21... a33) The point (a11 + a12. Calculating Determinants 99 .1 Let's s if 1 −2 1 det 1 −2 0 1 0 0 −1 3 0 1 0 0 −1 3 has an inverse. a32 + a33) ed The point (a11 + a12 + a13. = 1 · 1 · 3 + 0 · (−1) · (−2) + 0 · 1 · 0 − 0 · 1 · (−2) − 0 · 1 · 3 − 1 · (−1) · 0 =3+0+0−0−0−0 =3 1 det 1 −2 0 1 0 0 −1 3 ≠0 So this one has an inverse t ! And the volume of the para elepiped* spa by the fo owing eight points.. a31) The point (a12. a32) The point (a13. a22 + a23. a23.

... i suppose.. 100 Chapter 4 More Matrices . ! nope So how do we calculate them? To be able to do that. i'm afraid not. The grim truth is that the formulas used to calculate determinants of dimensions four and above are very complicated.So next up are 4×4 matrices... Tee h ee More of this perhaps? Yep.

the terms in the determinant formula are formed according to certain rules. Three rules? Yep. The left side.You'll have to learn the three rules of determinants.. Calculating Determinants 101 . p F llli Take a closer look at the term indexes.. ? ru le 1 Pay special attention to the left index in each factor.

.. They seem a bit more random. and 3—like in the table to the right. 2. Actually. This is rule number two. i see it now! Permutations of 1–2 Pattern 1 Pattern 2 Permutations of 1–3 Pattern 1 Pattern 2 Pattern 3 Pattern 4 Pattern 5 Pattern 6 102 Chapter 4 More Matrices . they all go from one to the number of dimensions! exactly. Their orders are all permutations of 1. Hmm. ru le And that's rule number one! 2 now for the right indexes.Oh. they're not.

Let's start by making an agreement. Sw itc h Sw itc h We gather all this information into a table like this. We will say that the right index is in its natural order if ? e eez Squ Okay! That is. if it is odd. so don't lose concentration. we write the term as positive.ru le 3 The third rule is a bit tricky. Then we count how many switches we need for each term. Whoa. Calculating Determinants 103 . Permutations of 1–2 Corresponding term in the determinant Switches Pattern 1 Pattern 2 Permutations of 1–3 Corresponding term in the determinant and Switches Pattern 1 Pattern 2 Pattern 3 Pattern 4 Pattern 5 Pattern 6 and and and and and and and and and if the number is even. The next step is to find all the places where two terms aren't in the natural order—meaning the places where two indexes have to be switched for them to be in an increasing order. indexes have to be in an increasing order. we write it as negative.

. Try comparing our earlier determinant formulas with the columns “Corresponding term in the determinant” and “Sign. 104 Chapter 4 More Matrices . and that's the third rule. Hmm.Permutations of 1–2 Corresponding term in the determinant Switches number of switches Sign Pattern 1 Pattern 2 Permutations of 1–3 Corresponding term in the determinant and number of switches Switches Sign Pattern 1 Pattern 2 Pattern 3 Pattern 4 Pattern 5 Pattern 6 and and and and and and and and and Like this. they're the same! exactly..” Ah! Corresponding term in the determinant Sign Corresponding term in the determinant Sign Wow.

Agh! Calculating Determinants 105 . Cool! Permutations of 1–4 Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa Pa ern 1 ern 2 ern 3 ern 4 ern 5 ern 6 ern 7 ern 8 ern 9 ern 10 ern 11 ern 12 ern 13 ern 14 ern 15 ern 16 ern 17 ern 18 ern 19 ern 20 ern 21 ern 22 ern 23 ern 24 Co esponding term in the determinant Switches Num.So. say we wanted to calculate the determinant of this 4×4 matrix: These three rules can be used to find the determinant of any matrix. we could calculate the determinant if we wanted to. of switches Sign 1 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 3 4 4 4 4 4 4 2 2 3 3 4 4 1 1 3 3 4 4 1 1 2 2 4 4 1 1 2 2 3 3 3 4 2 4 2 3 3 4 1 4 1 3 2 4 1 4 1 2 2 3 1 3 1 2 4 3 4 2 3 2 4 3 4 1 3 1 4 2 4 1 2 1 3 2 3 1 2 1 a11 a22 a33 a44 a11 a22 a34 a43 a11 a23 a32 a44 a11 a23 a34 a42 a11 a24 a32 a43 a11 a24 a33 a42 a12 a21 a33 a44 a12 a21 a34 a43 a12 a23 a31 a44 a12 a23 a34 a41 a12 a24 a31 a43 a12 a24 a33 a41 a13 a21 a32 a44 a13 a21 a34 a42 a13 a22 a31 a44 a13 a22 a34 a41 a13 a24 a31 a42 a13 a24 a32 a41 a14 a21 a32 a43 a14 a21 a33 a42 a14 a22 a31 a43 a14 a22 a33 a41 a14 a23 a31 a42 a14 a23 a32 a41 2&1 2&1 2&1 3&1 3&1 3&1 3&2 3&2 3&2 2&1 2&1 2&1 2&1 2&1 2&1 2&1 2&1 2&1 3&1 3&1 3&1 3&1 3&1 3&1 3&1 3&2 3&2 3&2 3&2 3&2 3&2 4&1 4&1 4&1 4&1 4&1 4&1 4&1 4&1 4&1 4&2 4&2 4&2 4&2 4&2 4&2 4&2 4&2 4&3 4&3 4&3 4&3 4&3 4&3 4&2 3&1 3&1 4&1 4&1 4&1 4&3 4&3 4&3 3&2 3&2 3&2 4&2 4&2 4&2 4&3 4&3 4&3 0 1 1 2 2 3 1 2 2 3 3 4 2 3 3 4 4 5 3 4 4 5 5 6 + − − + + − − + + − − + + − − + + − − + + − − + Using this information.

if this is on the test.Don't worry. Ph ew i hope so. Thanks. .. Maybe i can... i think that's enough for today.. reiji. We got through a ton of new material.... You're the best! Time really flew by.. though. i'm done for. most teachers will give you problems involving only 2×2 and 3×3 matrices... 106 Chapter 4 More Matrices .

. k! Smac Biff! Calculating Determinants 107 . Pow! There shouldn't be anyone left in there at this hour..Do you need a bag for that? Jeez. Look at the time.

we avoided using it in this chapter. j)-cofactor. here’s a quick explanation. as mentioned on page 88. you first have to understand these two concepts: • • The (i. j)-minor. j)-minor is the determinant produced when we remove row i and column j from the n×n matrix A: a11 a21 a12 a22 ai2 an2 a1j a2j aij anj a1n a2n ain ann Mij = det ai1 an1 1 0 0 0 All the minors of the 3×3 matrix 1 −2 1 −1 3 are listed on the next page. Mij The (i. 108 Chapter 4 More Matrices . since most books seem to introduce the method. • • Using cofactors Using Gaussian elimination Since the cofactor method involves a lot of cumbersome calculations. However. written as Cij So first we’ll have a look at these. written as Mij The (i. To use this method.Calculating inverse Matrices Using Cofactors There are two practical ways to calculate inverse matrices.

The table below contains all cofactors of the 3×3 matrix 1 1 −2 0 1 0 0 −1 3 C11 (1. 1) = (− 1)2+1 · det = (− 1) · 0 =0 C31 (3. 1) = (− 1)3+1 · det =1·0 =0 0 0 1 −1 0 0 0 3 1 −1 0 3 C12 (1. 3) det 1 1 0 1 =1 Cij If we multiply the (i. 3) det 1 −2 1 0 =2 M21 (2. 2) = (− 1)2+2 · det =1·3 =3 C32 (3. 2) = (− 1)1+2 · det = (− 1) · 1 =−1 C22 (2. 1) det 1 −1 0 3 =3 M12 (1. 1) = (− 1)1+1 · det =1·3 =3 C21 (2. we get the (i. 3) = (− 1)2+3 · det = (− 1) · 0 =0 C33 (3. 3) = (− 1)1+3 · det =1·2 =2 C23 (2. 1) det 0 0 1 −1 =0 M32 (3. 2) det 1 −2 0 3 =3 M23 (2. j)-cofactor. The standard way to write this is Cij.M11 (1. 3) det 1 −2 0 0 =0 M31 (3. 2) = (− 1)3+2 · det = (− 1) · (− 1) =1 1 0 1 −1 1 −2 0 3 1 −1 −2 3 C13 (1. j)-minor by (−1)i+j. 2) det 1 0 1 −1 = −1 M33 (3. 3) = (− 1)3+3 · det =1·1 =1 1 1 0 1 1 −2 0 0 1 −2 1 0 Calculating inverse Matrices Using Cofactors 109 . 2) det 1 −1 −2 3 =1 M13 (1. 1) det 0 0 0 3 =0 M22 (2.

i)-cofactor1 of the original matrix is called a cofactor matrix.The n×n matrix C11 C12 C1n C21 C22 C2n Cn1 Cn2 Cnn which at place (i. This is not a typo. The sum of any row or column of the n×n matrix a11C11 a12C12 a1nC1n a21C21 a22C22 a2nC2n an1Cn1 an2Cn2 annCnn is equal to the determinant of the original n×n matrix a11 a21 an1 a12 a22 an2 a1n a2n ann Calculating Inverse Matrices The inverse of a matrix can be calculated using the following formula: a11 a21 an1 a12 a22 an2 a1n a2n ann −1 C11 = det 1 a11 a21 an1 a12 a22 an2 a1n a2n ann C12 C1n C21 C22 C2n Cn1 Cn2 Cnn 1. 110 Chapter 4 More Matrices . This is the transpose of the matrix with the cofactors in the expected positions. i )-cofactor is the correct index order. ( j. j) has the ( j.

which is why we’ll cover the Cramer’s rule method next. the value stays unchanged. the values of the determinant are multiplied by −1. determinants are usually calculated by first simplifying them with Gaussian elimination–like methods and then using these three properties. Even though Gaussian elimination is one of the best ways to solve them by hand. for example) can easily reach determinant sizes in the n = 100 range. If two rows (or columns) switch places. which with the approach used here would produce insurmountable numbers of calculations. is only one of many methods you can use to solve linear systems. as presented on page 89. Because of this.For example. A typical application (in image processing. it is always good to know about alternatives. Using Determinants 111 . which can be derived using the definition presented in the book: • • • If a row (or column) in a determinant is replaced by the sum of the row (column) and a multiple of another row (column). the inverse of the 3×3 matrix 1 1 −2 0 0 0 3 1 −1 is equal to 1 1 −2 0 0 0 3 −1 1 −1 = det 1 1 1 −2 0 0 0 3 1 −1 3 −1 2 0 3 0 0 1 1 = 3 1 3 −1 2 0 3 0 0 1 1 Using Determinants The method presented in this chapter only defines the determinant and does nothing to explain what it is used for. Solving Linear Systems with Cramer's rule Gaussian elimination. The value of an upper or lower triangular determinant is equal to the product of its main diagonal. The difference between the two methods is so extreme that determinants that would be practically impossible to calculate (even using modern computers) with the first method can be done in a jiffy with the second one.

? ? Problem Problem Use Cramer’s rule to solve the following linear system: 3x1 + 1x2 = 1 1x1 + 2x2 = 0 ! ! Solution Solution Step 1 Rewrite the system a11x1 + a12x2 + a21x1 + a22x2 + an1x1 + an2x2 + + a1nxn = b1 + a2nxn = b2 + annxn = bn If we rewrite 3x1 + 1x2 = 1 1x1 + 2x2 = 0 like so: a11 a21 an1 Step 2 we get a1n a2n ann x1 x2 xn = a12 a22 an2 b1 b2 bn 3 1 1 2 x1 x2 = 1 0 Make sure that a11 det a21 an1 a12 a22 an2 a1n a2n ann ≠0 We have 3 1 1 2 det =3·2−1·1≠0 Step 3 Replace each column with the solution vector to get the corresponding solution: Column i a11 det xi = det a21 a12 a22 b1 b2 bn a1i a2i ani a1n a2n ann a1n a2n ann • x2 = det • x1 = det det 1 0 3 1 3 1 3 1 1 2 1 2 1 0 1 2 = 3·0−1·1 5 =− 1 5 = 1·2−1·0 5 = 2 5 an1 an2 a11 a21 a12 a22 det an1 an2 112 Chapter 4 More Matrices .

5 introduction to Vectors .

Pa n t One more minute! Pa n t Bam Ossu! You can do better than that! c Sm a k Put your back into it! Ossu! Bam Let's leave it at that for today. ze Whee Pa n t Pa n t M-more! Great! .

i can do more! You can barely even stand! Th u m p Uwa!! Just one more round! i haven't gotten stronger at all yet! Heheh. fine by me! Ossu! introduction to Vectors 115 .

and i'll be great! What Are Vectors? Let's talk vectors! 5 minutes later Sorry about that! ready? vectors are actually just a special interpretation of matrices. Matrices They appear quite frequently in Linear algebra. really? 116 Chapter 5 introduction to Vectors R evive d . Just give me five minutes to digest this delicious lunch.basics Fundamentals We're going to take a look at vectors today. i'm okay. Vectors Linear Transformation eigenvalues and eigenvectors Of course! course Layout basics Do you want to meet tomorrow instead? Fundamentals Matrices Vectors Linear transformation eigenvalues and eigenvectors Fwum p reiji. so pay close attention. are you okay? no.

Haha— mine too! x2 4 Starting point O 7 x1 Our course will look like this. right? What Are Vectors? 117 . 0) and that the hole is at (7. We'll use coordinates to describe where the ball and hole are to make explaining easier. i think it'll be easier to explain using an example.. 4).What kind of interPretation? Minigolf should make an excellent metaphor. Minigolf? don't be intimidated—my putting skills are pretty rusty. That means that the starting point is at (0..

4) 3 to the right and 1 up relative to (4. 2) + (3. 0) (3. 1) = (7. 2) = (4. 1) (3. Replay first stroke x2 Second stroke x2 Third stroke x2 4 1 O 3 7 x1 4 3 O 4 7 x1 4 O 7 x1 Stroke information First stroke Ball position Ball position relative to its last position Ball movement expressed in the form (to the right. I played conservatively and put the ba in with thr strokes. up) Second stroke Point (4. 1) 3 to the right and 1 up relative to (0.Reiji Yurino Player 1 I went first. 1) + (1. 1) + (1. 3) Third stroke Point (7. 1) 118 Chapter 5 introduction to Vectors . 3) (3. 4) Point (3. 3) 1 to the right and 2 up relative to (3.

10) Second stroke Point (7. −6) = (7. 0) (10. 10) 10 to the right and 10 up relative to (0.Misa Ichinose Player 2 Ah— T hard! You gave the ba a g wa op and put the ba with two strokes. 4) −3 to the right and −6 up relative to (10. 10) (10. 10) + (−3. 4) What Are Vectors? 119 . d in REPLAY first stroke x2 10 Second stroke x2 10 4 4 O 7 10 x1 O 7 10 x1 Stroke information First stroke Ball position Ball position relative to its last position Ball movement expressed in the form (to the right. up) Point (10.

Tetsuo Ichinose Player 3 Observe my exercise in skill! Yay big bro! And your brother got a hole-in-one. 4) 7 to the right and 4 up relative to (0.. up) Point (7..of course. 4) 120 Chapter 5 introduction to Vectors . 0) (7. REPLAY first stroke x2 4 O 7 x1 Stroke information First stroke Ball position Ball position relative to its last position Ball movement expressed in the form (to the right.

4) and the 2×1 matrix to make things simpler. What Are Vectors? 121 . at least we all made it in! 1×n matrices and n×1 matrices Try to remember the minigolf example while we talk about the next few subjects. 1×2 matrix (7.Well. Vectors can be interpreted in four different ways. Let me give you a quick walk-through of all of them. I' use the 7 4 Okay.

Interpretation 1

x2

(7, 4) and

7 4

are sometimes

interpreted as a point in space.

4

The point (7, 4)

O

7

x1

Interpretation 2

x2

In other cases, (7, 4) and

7 4

are interpreted as the “a ow” from the origin to the point (7, 4).

4

O

7

x1

Interpretation 3

And in yet other cases,

x2

x2 10

(7, 4) and

7 4

can mean the sum of several a ows equal to (7, 4).

4 3 1 O 34 7 x1

4

O

7

10 x1

122

Chapter 5

introduction to Vectors

Interpretation 4 Fina

x2 4

y, (7, 4) and

7 4

can also

be interpreted as any of the a ows on my left, or a of them at the same time!

O

7

x1

Hang on a second. i was with you until that last one...

How could a of them be representations of

(7, 4) and

7 4

when they start in completely di erent places?

While they do start in different places, they're all the same in that they go “seven to the right and four up,” right?

Yeah, i guess that's true!

What Are Vectors?

123

Vectors are kind of a mysterious concept, don't you think?

Well, they may seem that way at first.

but...

Once you've got the basics down, you'll be able to apply them to all sorts of interesting problems.

For example, they're frequently used in physics to describe different types of forces.

Cool!

Putting force

Gravitational force

124

Chapter 5

introduction to Vectors

Vector Calculations

even though vectors have a few special interpretations, they're all just 1×n and n×1 matrices...

And they're calculated in the exact same way.

A

ition

• (10, 10) + (−3, −6) = (10 + (−3), 10 + (−6)) = (7, 4) •

10 10

+

−3 −6

=

10 + (−3) 10 + (−6)

=

7 4

Subtraction • (10, 10) − (3, 6) = (10 − 3, 10 − 6) = (7, 4) •

10 10

3 6

=

10 − 3 10 − 6

=

7 4

Scalar multiplication • 2(3, 1) = (2 · 3, 2 · 1) = (6, 2) • 2

3 1

=

2·3 2·1

=

6 2

Matrix Multiplication •

3 1

(1, 2) = 1 2

3·1 3·2 1·1 1·2

=

3 1

6 2
Simple!

• (3, 1)

= (3 · 1 + 1 · 2) = 5 8 · 3 + (−3) · 1 2·3+ 1·1 21 7 3 1

8 −3 3 2 1 1

=

=

=7

Vector Calculations

125

Horizontal vectors like this one are called row vectors.

And vertical vectors are called column vectors.

Makes sense.

We also call the set of n all n×1 matrices R .

Sure, why not...

When writing vectors by hand, we usually draw the leftmost line double, like this.

All 2×1 vectors

All 3×1 vectors

All n×1 vectors

lot in linear algebra, so make sure you remember it. no problem.

Rn appears a

126

Chapter 5

introduction to Vectors

c) c 0 1 Yup. Can you s how the point (0.” O x1 O x1 0 c is an arbitrary 1 real number A straight line Even the straight line x1 = 3 can be expre ed as: x2 x1 = 3 x2 3 1 0 +c 3 1 0 +c 0 c is an arbitrary 1 real number O x1 O x1 no problem.Geometric interpretations Let's have a look at how to express points. The notation might look a bit weird at first. A point Let's say that c is an arbitrary real number. Geometric interpretations 127 . lines. 0 1 c is an arbitrary real number c 0 1 c is an arbitrary real number “|” can be read as “where. but you'll get used to it. O x1 O x1 An axis Do you understand this notation? x2 x2 c the x2-axis Yeah. and spaces with vectors. c) 0 and the vector c 1 are related? x2 x2 the point (0.

slanted drawing board.. c2 are arbitrary real numbers x2 c1 1 0 + c2 0 1 c1. c2 are arbitrary real numbers Sure enough! O x1 Another plane It can also be wri c1 3 1 + c2 1 2 x2 en another way: c1. c2 are arbitrary real numbers Hmm. c2 are arbitrary real numbers 3 1 1 2 c1. so it's like a weird..A plane And the x1x2 plane R2 can be expre c1 1 0 + c2 0 1 ed as: c1. c1 + c2 O x1 128 Chapter 5 introduction to Vectors .

x3 1 0 0 0 0 1 c1.. spa ed by x1. and x3 like this: 1 0 0 0 0 1 c1 0 + c 2 1 + c 3 0 c1. c3 are arbitrary real numbers x2 Sounds familiar. It is spa ed by x1. . c2.. xn: 1 c1 0 0 + c2 0 1 0 + … + cn 0 0 1 c1... x2 . ... cn are arbitrary real numbers i understand the formula.. but this one's a little harder to visualize.A vector space The thr -dimensional space R3 is the natural next step. c3 are arbitrary real numbers x1 c1 0 + c 2 1 + c 3 0 Another vector space Now try to imagine the n-dimensional space Rn. c2.. Geometric interpretations 129 . x2. c2.

. reiji. Hey! ? eh? We'd better get back to work! 130 Chapter 5 introduction to Vectors . anyway? Um. well.. i'm beat! Let's take a break. then. no special reason really... why did you decide to join the karate club... Good idea! by the way.Whoo.

..6 More Vectors They're bases. independent! Yeah. Definitely bases. we're linearly. bases! .

. let's have a look at linear independence and bases. okay? i'll try not to. right.. Question one. The two are pretty similar. Linear independence ? Problem 1 Problem 1 ? Find the constants c1 and c2 satisfying this equation: Why don't we start off today with a little quiz? 0 0 = c1 1 0 + c2 0 1 Sure.Well then. 132 Chapter 6 More Vectors . but we don’t want to mix them up.

That's easy. c2.. c3. it is. question two. c1 = 0 c2 = 0 Correct! ? Problem 2 Problem 2 ? isn't that also c1 = 0 c2 = 0 Find the constants c1 and c2 satisfying this equation: 0 0 = c1 3 1 + c2 1 2 Well then. .. . ? Problem 3 ? Problem 3 Find the constants c1. and c4 satisfying this equation: 0 0 = c1 1 0 + c2 0 1 + c3 3 1 + c4 1 2 Last one.

are other possible answers. .. c1 = 0 c2 = 0 c1 = 0 c2 = 0 c3 = 0 c4 = 0 Again? ? c3 = 0 c4 = 0 isn't wrong.... but.not quite. O 1 x1 −2 O 1 x1 −3 −4 0 0 =1 1 0 +2 0 1 +0 3 1 −1 1 2 0 0 =1 1 0 −3 0 1 −1 3 1 +2 1 2 Try to keep this in mind while we move on to the main problem. c1 = 1 x2 2 c2 = 2 c3 = 0 c4 = −1 and c1 = 1 c2 = −3 c3 = −1 c4 = 2 x2 You're right.

.... .. .. a12 a22 .... = c1 a11 a21 + c2 a12 a22 + … + cn a1n a2n amn Weee! Linear dependence Linear independence We say that its vectors Their vectors a11 a21 .. and a1n a2n a11 a21 . am1 am2 amn am1 am2 are linearly independent. ... . amn Ah... ..c1 = 0 . . Linear independence .. .... .. and a1n a2n 135 .. are called linearly dependent. And linear dependence is similarly sometimes called one-dimensional dependence. 0 . a12 a22 ....... Linear independence is sometimes called one-dimensional independence.. am1 am2 no matter what we do! We can never return to the origin. As long as there is only one unique solution c2 = 0 cn = 0 As for problems like the third example. where there are solutions other than c1 = 0 c2 = 0 cn = 0 All work together we can Get back to the origin! if we to problems such as the first or second examples: 0 0 .

Here are some examples. 1 . and 0 x2 1 O 1 x3 1 x1 0 give us the equation 0 1 0 0 0 c1 = 0 0 1 0 = c1 0 + c2 1 + c3 0 which has the unique solution c2 = 0 c3 = 0 The vectors are therefore linearly independent. k at Example example1 1 1 The vectors 0 0 0 0 1 0 . 136 Chapter 6 More Vectors . Let's l linear independence first.

Linear independence 137 .example22 Example 1 0 0 0 The vectors 0 and 1 x2 1 O x3 1 x1 0 give us the equation 0 1 0 = c1 0 + c2 1 0 0 c1 = 0 c2 = 0 0 This one too? which has the unique solution These vectors are therefore also linearly independent.

1 . 138 Chapter 6 More Vectors . and 1 x2 1 O x3 1 3 x1 0 give us the equation 0 1 0 3 = c1 0 + c2 1 + c3 1 0 0 0 0 c1 = 0 c3 = 0 c1 = 3 c3 = −1 which has several solutions. for example c2 = 0 and c2 = 1 This means that the vectors are linearly dependent. example1 1 Example 1 0 0 0 3 0 The vectors 0 .And now we’ l k at linear dependence.

. 0 . 1 . c2 = 0 c3 = 0 c4 = 0 1 0 0 0 1 0 0 0 1 and c1 = a 1 c2 = a 2 c3 = a 3 c4 = −1 a1 a2 am The vectors . and a2 0 = c1 0 + c2 1 + c3 0 + c4 a2 The vectors are linearly dependent because there are several solutions to the system— c1 = 0 for example. and are similarly linearly dependent because there are several solutions to the equation 0 0 0 1 0 0 0 1 0 c1 c2 Among them is cm = 0 cm+1 = 0 =0 =0 but also cm = am cm+1 = −1 0 0 1 c1 c2 a1 + cm+1 a2 am = a1 = a2 = c1 + c2 + … + cm Linear independence 139 .example2 2 Example 1 0 0 as well as the equation 0 1 0 0 0 0 1 0 0 a1 a3 0 1 a1 a3 Suppose we have the vectors 0 . .

.bases Here are three more problems. First one. Mhmm. it kinda looks like the other problems. Correct! O 7 x1 140 Chapter 6 More Vectors . ? Problem 4 Problem 4 ? Find the constants c1 and c2 satisfying this equation: 7 4 = c1 1 0 + c2 0 1 x2 7 4 4 =7 1 0 +4 0 1 c1 = 7 c2 = 4 should work..

Let's see. ? Problem 5 ? Problem 5 Find the constants c1 and c2 satisfying this equation: 7 4 = c1 3 1 + c2 1 2 x2 7 4 4 =2 3 1 +1 1 2 c1 = 2 c2 = 1 right? Correct again! O 7 x1 You're really good at this! Well those were pretty easy.Here's the second one. bases 141 .....

c2. and c4 satisfying this equation: 7 4 = c1 1 0 + c2 0 1 + c3 3 1 + c4 1 2 Sharp answer! c1 = 7 There's c1 = 0 and c1 = 5 and of course c2 = 4 c3 = 0 c4 = 0 c2 = 0 c3 = 2 c4 = 1 c2 = −5 c3 = −1 c4 = 5 . Ah.. 4 O 7 x1 142 Chapter 6 More Vectors . c3.Last one. it has lots of possible solutions. x2 7 4 =5 1 0 −5 0 1 −1 3 1 +5 1 2 That's enough.. doesn't it? ? Problem 6 6 ? H m! Find the constants c1.

Linear dependence and independence are closely related to the concept of a basis. Have a look at the following equation: y1 y2 ym = c1 a11 a21 am1 + c2 a12 a22 am2 + … + cn a1n a2n amn There's only one solution to c1 and c 2 Point For us! where the left side of the equation is an m arbitrary vector in R and the right side is a number of n vectors of the same dimension. as well as their coefficients. Does that mean that the solution 1 0 . 3 1 . bases 143 . Point and the solution 3 1 . Okay. a1n a2n amn Point too! n make up a basis for R . 1 2 oo. but the solution What do we dooo? 1 0 . 0 1 for problem 4 There are so many different ways! Which ones should we use. then our vectors a12 . 1 2 exactly! for problem 6 isn't? Here are some examples of what is and what is not a basis..…. 0 1 . = cn = 0 a11 a21 am1 If there's only one solution to the equation. a22 am2 . i don't know! for problem 5 are bases... basis For us c1 = c2 = ..

a basis is a minimal set of vectors n ded to expre an arbitrary vector in Rm. 1 . 0 0 1 The set 2 1 3 . 0 . Another important feature of bases is that they're a linearly independent.A these vector sets make up bases for their graphs. 0 0 0 0 1 The set x2 2 3 0 1 0 . 1 1 2 1 O 1 x1 O 1 3 x1 x2 The set 1 0 0 . 2 0 −5 −1 −5 1 −1 1 x3 O 1 O x1 x3 1 3 x1 In other words. 144 Chapter 6 More Vectors . x2 The set x2 1 .

c2. erent ways erent choices for c1. owing set do not x2 The set 2 1 1 . c3. 3 .” bases 145 . have a l k at the fo owing equation: y1 1 0 3 1 = c1 + c2 + c3 + c4 y2 0 1 1 2 where y1 y2 can be formed in many di (using di y1 2 y2 is an arbitrary vector in R . Because of this. the set does not form “a minimal set of vectors n ded to expre an arbitrary vector in Rm. 1 0 1 1 2 O 1 3 x1 To understand why they don't form a basis. and c4).The vectors of the fo form a basis. 0 .

Neither of the two vector sets below is able 0 to describe the vector 0 . then there's no way that they could describe “an arbitrary vector in R3. 1 . 146 Chapter 6 More Vectors . the set 1 0 0 0 . 0 0 0 1 forms a basis. even though they're both linearly independent. they're not bases. 1 0 0 does not.” Because of this. while the set 1 0 0 . 2 0 0 0 0 1 O x3 1 O x1 x3 1 1 x1 Just because a set of vectors is linearly independent doesn't mean that it forms a basis. x2 The set 1 0 . For instance. 1 . 1 0 0 0 x2 2 The set 1 0 1 . and if they can’t 1 describe that vector.

.Since bases and linear independence are confusingly similar.…. . forms a basis if there’s only one solution to the equation where the left side is an arbitrary vector in Rm. is linearly independent if there’s only one solution to the equation = c1 + c2 + … + cn where the left side is the zero vector of Rm. a basis bases 147 .…. Bases a11 a21 am1 a12 a22 am2 y1 y2 ym a1n a2n amn = c1 a11 a21 am1 y1 y2 ym is a minimal set of vectors needed to express an arbitrary vector in Rm. And once again. Linear Independence a11 a21 am1 c1 = 0 c2 = 0 cn = 0 0 0 0 a11 a21 am1 a12 a22 am2 a1n a2n amn a12 a22 am2 a1n a2n amn We say that a set of vectors . + c2 a12 a22 am2 + … + cn a1n a2n amn A set of vectors . I thought I'd talk a bit about the di erences betw n the two.

wait a sec! no big deal! 148 Chapter 6 More Vectors .. They are.. linearly independent! Yep. bases are about finding clear-cut paths to any m vector in a given space R ? We're They're bases.So. exactly! not a lot of people are able to grasp the difference between the two that fast! i must say i'm impressed! That's all for tod— Ah. While linear independence is about finding a clear-cut path back to the origin.

i've been thinking... wow. 2 dimensions R2 Set 3 dimensions R3 Set You know.Dimension it's kind of obvious that a basis is made up of two vectors 2 when in R and three 3 vectors when in R . i didn't think you'd notice.. but why is it that the basis of an m-dimensional space consists of n vectors and not m? Oh.. more precise definition of bases. we'll have to take a look at another. i'm up for it! There's also a more precise definition of vectors. which can be hard to understand. Dimension 149 . To answer that.

since you asked and all. 150 Chapter 6 More Vectors .. Subspaces So it's another word for subset? it's kinda like this. Let's have a look.You sure? i— i think so. but first we have to tackle another new concept: subspaces. not quite. Let me try again. Subspace So let's talk about them. O-okay. it's actually not that hard—just a little abstract.. no.

then c ∈W  The sum of two arbitrary elements in W is still an element in W. If both of these conditions hold. then + ∈W Th is is th e definition. (Closed under scalar multiplication.What Is a Subspace? Let c be an arbitrary real number and W be a nonempty subset of Rm satisfying these two conditions:  An element in W multiplied by c is still an element in W.. then W is a subspace of Rm. (Closed under addition. Umm. Dimension 151 .. This picture illustrates the relationship.) a1i a2i ami a1i a2i ami If ∈ W.) a1i a2i ami a1j a2j amj a1i a2i ami a1j a2j amj If ∈ W and ∈ W.

cα1 α1  c 0 = 0 0 0 α ∈ 0 0 α is an arbitrary real number α α is an arbitrary real number  α1+α2 α1 α2 0 + 0 = 0 0 0 0 ∈ 0 0 It seems like they do! This means it actually is a subspace. the x-axis. Two-dimensional subspaces are similarly planes through the origin. so you might have to read it a few times before it starts to sink in. Other subspaces can also be visualized. 152 Chapter 6 More Vectors . arbitrary real number x2 α 0 0 O x3 α x1 If it really is a subspace.It's pre y abstract. more concrete way to l k at onedimensional subspaces is as lines through the origin. I made some examples of spaces that are subspaces— and of some that are not. Have a l k! This Is a Subspace Let’s have a look at the subspace in R3 defined by the set α 0 0 α is an . Another. it should satisfy the two conditions we talked about before. but not as easily. in other words.

a real values—it's just not enough to test with a few chosen numerical examples. it'll make more sense after solving a few problems. The vector set is a subspace only if both conditions hold for a kinds of vectors.. and therefore it is not a subspace! I'd imagine you might think that “Both  and  hold if we use α1 = α2 = 0. Dimension 153 . so it should be a subspace!” It's true that the conditions hold for those values. but since the conditions have to hold for arbitrary real values—that is. Let’s use our conditions to see why: α1  c α12 = 0 cα1 cα1 0 2 cα1 ≠ (cα1) 0 α1+α2 α12+α22 0 2 α ∈ α2 0 α is an arbitrary real number α α is an arbitrary real number  α1 α2 α12 + α22 = 0 0 α1+α2 ≠ (α1+α2)2 0 ∈ α2 0 The set doesn’t seem to satisfy either of the two conditions.. If this sti is hard! doesn't make sense. don’t give up! This i think i get it.This Is Not a subspace α The set α2 0 α is an arbitrary real number is not a subspace of R3.

be spanned by using the two vectors 3 1 and 1 2 x2 like so: c1 3 1 + c2 1 2 c1 and c2 are arbitrary numbers 2 1 O 1 3 x1 154 Chapter 6 More Vectors .. . a12 a22 am2 a11 a21 am1 . a1n a2n amn a12 a22 am2 a1n a2n amn span the following subspace in Rm: c1 + c2 + … + cn c1.The fo owing subspaces are ca spans and are a bit special. c2. Example example1 1 The x1x2-plane is a subspace of R2 and can. for example. .. ed linear What Is a Linear Span? We say that a set of m-dimensional vectors a11 a21 am1 . and cn are arbitrary numbers This set forms a subspace and is called the linear span of the n original vectors.

which you could probably te from l king at the example on page 152.example 2 Example 2 The x1x2-plane could also be a subspace of R3. as you might have gue ed from example 1. creating this set: 0 1 0 c1 0 + c 2 1 0 0 c1 and c2 are arbitrary numbers x2 x1 x -pl ane 2 1 O x3 1 x1 Rm is also a subspace of itself. Remember. A subspaces contain the zero factor. and we could span it using the 1 vectors 0 0 0 and 1 . they must pa through the origin! 0 0 0 Dimension 155 .

and a1n a2n amn . Hmm. i’m a little lost.. we say that the set forms a basis to the subspace W. . a12 a22 am2 .. “The dimension of the subspace W ” is usually written as dim W. 156 Chapter 6 More Vectors . .basis and Dimension Here are the definitions of basis and dimension. and cn are arbitrary numbers a11 a21 am1 a12 a22 am2 a1n a2n amn . The dimension of the subspace W is equal to the number of vectors in any basis for W.…. Sorry for the wait. . W= c1 + c2 + … + cn When this equality holds.. c2. What Are Basis and dimension? Suppose that W is a subspace of Rm and that it is spanned by the linearly independent vectors a11 a21 am1 This could also be written as follows: a11 a21 am1 a12 a22 am2 a1n a2n amn c1.

2 c1 1 + c2 2 The fact that this equality holds means that the vector set is a basis of the subspace W. Dimension 157 . So suppose that W is a subspace of R3 and is spanned by the linearly independent vectors 3 1 0 and 1 2 . Since the base contains two vectors. dim W = 2. 0 x2 2 1 O x1 x -pl ane 2 1 3 x 1 i see! x3 We have this: 3 W= 0 1 0 c1 and c2 are arbitrary numbers 3 0 1 0 1 . for simplicity’s sake.This example might clear things up a li le. Example example Let’s call the x1x2-plane W.

What do you think? Were you able to follow? three dimensions if the subspace's basis has two vectors. right? You got it! That's enough for today. then the dimension of the subspace has to be two. looks like we're going the same way.. Let's talk about linear transformations next time. so come prepared! Of course! Oh. Fun as always! Thanks for all the help. 158 Chapter 6 More Vectors .. Sure! it's like this. it's also an important subject.

i'm tired of being such a wimp.. i want to get stronger... Dimension 159 . Well.... i didn't really answer your question before. That way i can. ? i joined the karate club because...… Umm.

a lot more homemade lunches. then! To survive my brother's reign of terror. Wha—? .i guess you'll be needing.. i'll make you my super special stamina-lunch extravaganza every week from now on! Thank you. Don't worry.. that is. Hehe. Misa... don't worry 'bout it! 160 Chapter 6 More Vectors .

4) 0O 0 1 In this kind of system. I’ll try explaining the difference between the two using the image below. it’s much easier to use only the trivial basis: 7 0 1 =7 +4 1 0 0 4 1 0 0 1 0 .…. 4) 7 4 O x2 =2 3 1 +1 1 2 u2 7 x1 point (2.Coordinates Coordinates in linear algebra are a bit different from the coordinates explained in high school. 1) Coordinates 161 1 u1 . . x2 4 point (7. the relationship between the origin and the point in the top right is interpreted as follows: 7 4 =7 1 0 +4 0 1 O x2 4 7 x1 point (7. point O point When working with coordinates and coordinate systems at the high school level.

for example.It is important to understand that the trivial basis is only one of many bases when we move into the realm of linear algebra—and that using other bases pro7 x1 O duces other relationships between the origin and a given point. and u2 = 1 2 7 4 x2 u2 point (2. 1) 3 1 1 2 =2 +1 1 2 u1 O x1x This alternative way of thinking about coordinates is very useful in factor analysis. 1) in a system using the nontrivial basis consisting of 3 1 the two vectors u1 = . 162 Chapter 6 More Vectors . The image below illustrates the point (2.

7 Linear Transformations .

huh? i guess i'll be sitting it out.. You're in.. A match. Are you telling me what to do? Oh no! Of course not! S l a m 164 Chapter 7 .A practice match with nanhou University? That’s right! We go head-to-head in two weeks. sensei? isn't it a bit early for. Yurino! What?! For real? Um.

... Dismissed! What am i going to do. A match. er Shudd i have to do my best. Linear Transformations 165 . Stop thinking like that! He’s giving me this opportunity..i'll be looking forward to seeing what you've learned so far! Ossu! Of course! Understood? Great.

. and f a function from X to Y. Let xi and xj be two arbitrary elements. 166 Chapter 7 Linear Transformations .What is a Linear Transformation? it seems we've finally arrived at linear transformations! Course layout Fundamentals Prep Matrices Vectors Linear transformations eigenvalues and eigenvectors Let's start with the definition.. Linear Transformations We touched on this a bit in Chapter 2. Yeah. We say that f is a linear transformation from X to Y if it satisfies the following two conditions:  f(xi) + f(xj) and f(xi + xj) are equal  cf(xi) and f(cxi) are equal but this definition is actually incomplete. c an arbitrary real number. Sounds good.

.Li n ea r s f o r m at i o n s i think you’re ready for the real one now! eep! Let x1i x2i xni and x1j x2j xnj Linear Transformations be two arbitrary elements from Rn. So.  cf and f c are equal. and f a function from Rn to Rm.. we're dealing with vectors instead of numbers? exactly! What is a Linear Transformation? 167 . A linear transformation from Rn to Rm is sometimes called a linear map or linear operation. We say that f is a linear transformation from Rn to Rm if it satisfies the following two conditions: x1i x2i xni x1i x2i xni x1j x2j xnj x1i x2i xni x1i + x1j x2i + x2j xni + xnj f +f and f are equal. c an arbitrary real number.

Then it shouldn't be a surprise to hear that f can be written as an m×n matrix..And if f is a linear transformation from Rn to Rm.. 168 Chapter 7 Linear Transformations . Um...it shouldn’t? Have a look at the following equations.

What is a Linear Transformation? 169 .x1i  We’ll verify the first rule first: f x2i xni +f x1j x2j xnj = f x1i + x1j x2i + x2j xni + xnj We just replace f with a matrix. then simplify: a11 a21 am1 a12 a22 am2 a1n a2n amn x1i x2i xni + a11 a21 am1 a12 a22 am2 a1n a2n amn x1j x2j xnj a11x1i + a12x2i + … + a1nxni = a21x1i + a22x2i + … + a2nxni am1x1i + am2x2i + … + amnxni + a11x1j + a12x2j + … + a1nxnj a21x1j + a22x2j + … + a2nxnj am1x1j + am2x2j + … + amnxnj a11(x1i + x1j) + a12(x2i + x2j) + … + a1n(xni + xnj) = a21(x1i + x1j) + a22(x2i + x2j) + … + a2n(xni + xnj) am1(x1i + x1j) + am2(x2i + x2j) + … + amn(xni + xnj) a11 = a21 am1 a12 a22 am2 a1n a2n amn x1i + x1j x2i + x2j xni + xnj Uh-huh.

x1i  Now for the second rule: cf x2i xni = f c x1i x2i xni Again. just replace f with a matrix and simplify: a11 c a21 am1 a12 a22 am2 a1n a2n amn x1i x2i xni a11x1i + a12x2i + … + a1nxni = c a21x1i + a22x2i + … + a2nxni am1x1i + am2x2i + … + amnxni a11(cx1i) + a12(cx2i) + … + a1n(cxni) = a21(cx1i) + a22(cx2i) + … + a2n(cxni) am1(cx1i) + am2(cx2i) + … + amn(cxni) a11 = a21 am1 a11 = a21 am1 a12 a22 am2 a12 a22 am2 a1n a2n amn a1n a2n amn c cx1i cx2i cxni x1i x2i xni Oh. i see! 170 Chapter 7 Linear Transformations .

. x2j x2i O x1j x1i x1 O a11x1i + a12x2i x1 a11x1j + a12x2j then a .... as f. R2 a If you first. O x1i + x1j x1 O a11(x1i + x1j) + a12(x2i + x2j) x1 You get the same final result! What is a Linear Transformation? 171 ..  We’ll show that the first rule holds: a11 a12 x1i a11 a12 x1j + = a21 a22 x2i a21 a22 x2j R2 x2 a11 a12 a21 a22 x1i + x1j x2i + x2j R2 x2 a21x1j + a22x2j a21x1i + a22x2i If you multiply first...We can demonstrate the same thing visua We’ use the 2×2 matrix a11 a12 a21 a22 y. R2 x2 x2 a21(x1i + x1j) + a22(x2i + x2j) x2i + x2j then multiply..

 And the second rule, too: c

a11 a12 x1i a21 a22 x2i

=

a11 a12 a21 a22

c

x1i x2i

R2 x2 If you multiply by the matrix first... x2i O x1i x1 O x2

R2

a21x1i + a22x2i

a11x1i + a12x2i x1

If you multiply by c first...

then multiply by c...

R2 x2 x2

R2

a21(cx1i) + a22(cx2i) then multiply by the matrix... cx1i x1 O x1

cx2i

O

a11(cx1i) + a12(cx2i)

You get the same final result!

Awesome!

172

Chapter 7

Linear Transformations

So when f is a linear transformation from Rn to Rm, we can also say that f is equivalent to the m×n matrix that defines the linear transformation from Rn to Rm.

a11 a21 am1

a12 a22 am2

a1n a2n amn
now i get it!

Why We Study Linear Transformations

So...What are linear transformations good for, exactly?

Well, it's not really a question of importance...

Well...

They seem pretty important. i guess we'll be using them a lot from now on?

So why do we have to study them?

That's exactly what i wanted to talk about next.

Why We Study Linear Transformations

173

Consider the linear transformation from Rn to Rm defined by the following m×n matrix:

a11 a21 am1 y1
If

a12 a22 am2

a1n a2n amn

x1
is the image of

y2 ym

x2 xn

under this linear transformation,

then the fo

owing equation is true:

image?

Yep. Here’s a definition.

images Suppose xi is an element from X.
We talked a bit about this before, didn't we?

The element in Y corresponding to x i under f is called “xi ’s image under f.”

Fl ip
!
174 Chapter 7 Linear Transformations

Yeah, in Chapter 2.

but that definition is a bit vague. Take a look at this.

Okay.

Doesn't it kind of look like a common onedimensional equation y = ax to you?
maybe if i squint...

What if i put it like this?

i guess that makes sense.

Multiplying an n-dimensional space by an m×n matrix...

turns it m-dimensional!

Why We Study Linear Transformations

175

We study linear transformations in an effort to better understand the concept of image, using more visual means than simple formulae.

Ta - d a

!

Huh?

i have to learn this stuff because of...that?

Ooh, but ”that” is a lot more significant than you might think!

Take this linear transformation from three to two dimensions, for example.

You could write it as this linear system of equations instead, if you wanted to.

but you have to agree that this doesn't really convey the feeling of ”transforming a threedimensional space into a two-dimensional one,” right?
176 Chapter 7

. Why We Study Linear Transformations 177 .. This! i wonder what these linear systems are supposed to be? Multiplying a three-dimensional space.. a x ....is the same as. Linear transformations are definitely one of the harder-tounderstand parts of linear algebra. i remember having trouble with them when i started studying the subject.. by tri ma 3 2× Ah.. Turns it twodimensional! So that's what it is! i think i'm starting to get it.

Computer graphics. translation.Special Transformations i wouldn't want you thinking that linear transformations lack practical uses. Aww ! cute ! C li c k Let's use one of my drawings. Let (x1. and 3-D projection. let's have a look at some of the transformations that let us do things like scaling. for example. The top of the dorsal fin will do! The point 178 Chapter 7 Linear Transformations . x2) be some point on the drawing. really? Yeah. As we're already on the subject. rotation. rely heavily on linear algebra and linear transformations in particular. though.

. sure. The point The point and So that means that applying the set of rules y1 = αx1 y2 = βx2 Multiply all x1 values by α Multiply all x2 values by β onto an arbitrary image is basically the same thing as passing the image through a linear transformation in R2 equal to the following matrix! α 0 0 β Oh. the point the point Special Transformations 179 .. right? Yeah. it’s a one-toone onto mapping! Could be rewritten like this.Scaling Let’s say we decide to Multiply all x1 values by α Multiply all x2 values by β y1 = αx1 y2 = βx2 This gives rise to the interesting relationship Uh-huh.

180 Chapter 7 Linear Transformations . that is + . x2 0 2 by θ degrees gets us x1cosθ x1sinθ = + −x2sinθ x2cosθ the point (x1cosθ − x2sinθ. x2cosθ) the point (0. x1sinθ + x2cosθ) the vector −x2sinθ x2cosθ O the vector x1cosθ x1sinθ x1cosθ − x2sinθ x1sinθ + x2cosθ * θ is the Greek letter theta. x1sinθ) θ O the point (x1.. x2) θ O x1 0 x1 • Rotating x .rotation i hope you’re up on your trig. 0) • Rotating 0 x2 by θ degrees gets us −x2sinθ x2cosθ the point (−x2sinθ. You know it! • Rotating x1 0 by θ* degrees gets us x1cosθ x1sinθ the point (x1cosθ..

. Rotating an arbitrary image by θ degrees consequently means we’re using a linear transformation in R2 equal to this matrix: cosθ −sinθ sinθ cosθ Another one-toone onto mapping! the point the point Special Transformations 181 .So if we wanted to rotate the entire picture by θ degrees.Due to this relationship.. the point the point .. we'd get.. Aha.

182 Chapter 7 Linear Transformations . but okay. if we wanted to.Translation the point The point If we instead decide to Translate all x1 values by b1 Translate all x2 values by b2 y1 = x1 + b1 y2 = x2 + b2 we get another interesting relationship: And this can also be rewritten like so: That’s true. we could also rewrite it like this: Seems silly.

.. wait a minute! Why are you dragging another dimension into the discussion all of a sudden? And what was the point of that weird rewrite? “ y = ax !” Special Transformations 183 . Hey.So applying the set of rules Translate all x1 values by b1 Translate all x2 values by b2 onto an arbitrary image is basically the same thing as passing the image through a linear transformation in R3 equal to the following matrix: 1 0 0 0 b1 1 b2 0 1 the point the point Yet another one-toone onto mapping.

..even rotations and scaling operations. not too different.. y = y 2 y1 a11 a12 a aa 21 22 x x 2 x1 instead of with y = y 2 y1 a11 a12 a aa 21 22 x + b x b 2 2 x1 b1 The first formula is more practical than the second.. with Errr .. A computer stores all transformations as 3×3 matrices. especially when dealing with computer graphics. i guess.. conventional Linear transformations Linear Transformations used by computer graphics systems Scaling y1 y2 = α 0 0 β x1 x2 y1 1 y1 1 x1 x2 α 0 cosθ 0 0 β 0 0 0 1 x1 x2 1 0 0 1 x1 x2 1 ing y2 = 0 Rotation y1 y2 = cosθ −sinθ sinθ cosθ −sinθ cosθ 0 x1 x2 1 y2 = sinθ Translation y1 y2 = 1 0 b1 * 0 x1 + b2 1 x2 y1 1 1 0 0 b1 1 b2 0 1 y2 = 0 * Note: This one isn’t actua y a linear transformation. 184 Chapter 7 Linear Transformations . You can verify this by se b1 and b2 to 1 and checking that both linear transformation conditions fail.We'd like to express translations in the same way as rotations and scale operations..

Special Transformations 185 . Don’t worry too much about the details.th e po int 3-D Projection next we’ll very briefly talk about a 3-D projection technique called perspective projection. So i’ll cheat a little bit and skip right to the end! The linear transformation we use for perspective projection is in R4 and can be written as the following matrix: 1 x3 − s 3 −s3 0 0 0 0 −s3 0 0 s1 s2 0 1 0 0 0 −s3 Coool. -th e pl an Oh. an onto mapping! e The math is a bit more complex than what we’ve seen so far. th e po i nt th e po i nt Perspective projection provides us with a way to project three-dimensional objects onto a near plane by tracing our way from each point on the object toward a common observation point and noting where these lines intersect with the near plane.

actually.. we’ll cover all the important topics. Hehe. i think. i'm going to the gym after this..but that's enough for today. 186 Chapter 7 Linear Transformations Yeah. Be am You shouldn't worry either. Final lesson? So soon? so much to learn. Oh. We'll be talking about eigenvectors and eigenvalues in our next and final lesson.. Thanks.. you heard? Hm? p zi zz zz About the match. you know.. Don't worry. i hope i don't lose too badly. .And that's what transformations are all about! Yeah. why would i worry? You're such a good teacher. my brother told me. Heh..

i'll do my best! i know you can do it..Don’t say that! You’ve got to stay positive! Ththanks. Special Transformations 187 ..

. and the other advanced topics we’re going to cover in the remainder of this chapter... + xn a1n a 2n amn a1n a 2n amn x1 x1 x2 xn 1 0 0 a1n a 2n amn + x2 0 1 0 + .Some Preliminary Tips Before we dive into kernel. Also note that the function f referred to throughout this chapter is the linear transformation from Rn to Rm corresponding to the following m×n matrix: a11 a 21 am1 a12 a 22 am2 a1n a 2n amn 188 Chapter 7 Linear Transformations . + xn 0 0 1 As you can see. the product of the matrix M and the vector x can be viewed as a linear combination of the columns of M with the entries of x as the weights. there’s a little mathematical trick that you may find handy while working some of these problems out. The equation y1 y2 ym = a11 a 21 am1 a12 a 22 am2 a1n a 2n amn x1 x2 xn can be rewritten like this: y1 y2 ym = a11 a 21 am1 a11 = a 21 am1 a11 = x1 a 21 am1 + x2 a12 a 22 am2 a12 a 22 am2 a12 a 22 am2 + .. rank.

and the Dimension Theorem for Linear Transformations The set of vectors whose images are the zero vector. see “Basis and Dimension” on page 156. The dimension theorem for linear transformations further explores this observation by defining a relationship between the two: dim Ker f + dim Im f = n Note that the n above is equal to the first vector space’s dimension (dim Rn). The image of f is equal to the set of vectors that is made up of all of the possible output values of f. image.Kernel.* Rn Ker f 0 0 0 Rm Im f 0 0 0 * If you need a refresher on the concept of dimension. and the Dimension Theorem for Linear Transformations 189 . image. that is x1 x2 xn 0 0 0 = a11 a 21 am1 a12 a 22 am2 a1n a 2n amn x1 x2 xn is called the kernel of the linear transformation f and is written Ker f. as you can see in the following relation: y1 y2 ym y1 y2 ym = a11 a 21 am1 a12 a 22 am2 a1n a 2n amn x1 x2 xn (This is a more formal definition of image than what we saw in Chapter 2.) An important observation is that Ker f is a subspace of Rn and Im f is a subspace of Rm. The image of f (written Im f ) is also important in this context. but the concept is the same. Kernel.

example 1 3 Suppose that f is a linear transformation from R2 to R2 equal to the matrix 1 Then: x1 x2 y1 y2 0 0 y1 y2 3 1 1 2 3 1 1 2 x1 x2 x1 x2 x1 x2 y1 y2 0 0 3 1 3 1 1 2 1 2 0 0 1 2 . Ker f = = = = x1 + x2 = Im f = = = y1 y2 = x1 + x2 = R2 n And: dim Im f example 2 =2 =2 dim Ker f = 0 3 Suppose that f is a linear transformation from R2 to R2 equal to the matrix 1 Then: x1 x2 0 0 3 6 1 2 x1 x2 x1 x2 −2 1 y1 y2 3 1 0 0 3 1 6 2 . Ker f = = = = [x1 + 2x2] = c y1 y2 y1 y2 3 6 1 2 x1 x2 c is an arbitrary number y1 y2 = [x1 + 2x2] 3 1 Im f = = = = c c is an arbitrary number n And: dim Im f =2 =1 dim Ker f = 1 190 Chapter 7 Linear Transformations .

image. 0 0 0 Ker f = x1 x2 y1 0 0 y1 1 0 0 1 0 0 = 0 x1 x2 = x1 x2 y1 0 0 y1 1 0 1 0 = 0 0 0 = x1 0 + x 2 1 0 0 1 0 Im f = y2 y3 y2 = 0 1 y3 0 0 x1 x2 = y2 y3 1 0 y2 = x1 0 + x 2 1 y3 0 0 0 0 c1 and c2 are arbitrary numbers = c1 0 + c2 1 n And: dim Im f =2 =2 dim Ker f = 0 Kernel. and the Dimension Theorem for Linear Transformations 191 .example 3 1 Suppose f is a linear transformation from R to R equal to the 3×2 matrix Then: 2 3 0 1 .

Then: x1 x2 x3 x4 n And: dim Im f =4 =2 dim Ker f = 2 192 Chapter 7 Linear Transformations . x2 + x3 + 2x4 = 0 0 0 = x1 1 0 + x2 0 1 + x3 3 1 + x4 1 2 0 0 = 1 0 0 1 3 1 1 2 0 1 3 1 1 2 .example 4 Suppose that f is a linear transformation from R4 to R2 equal to the 2×4 matrix 1 0 x1 Ker f = x2 x3 x4 x1 = x2 x3 x4 x1 = x2 x3 x4 −3 = c1 −1 1 0 + c2 −1 −2 0 1 x1 Im f = y1 y2 y1 y2 y1 y2 y1 y2 = 1 0 0 1 3 1 1 2 x2 x3 x4 = = x1 1 0 + x2 0 1 + x3 3 1 + x4 1 2 = R2 c1 and c2 are arbitrary numbers x1 + 3x3 + x4 = 0.

can be rewritten as follows: y1 y2 = = 3 6 1 2 x1 x2 = x1 = x1 + x2 + 2x2 3 1 6 2 3 1 = [x1 + 2x2] So the rank of 3 1 6 2 3 1 is 1.rank The number of linearly independent vectors among the columns of the matrix M (which is also the dimension of the Rm subspace Im f ) is called the rank of M. that is y1 y2 = 3x1 + 6x2 1x1 + 2x2 3 1 3 1 . Also note that det rank 193 . and it is written like this: rank M. example 1 The linear system of equations 3x1 + 1x2 = y1 1x1 + 2x2 = y2 3x1 + 1x2 1x1 + 2x2 . that is y1 y2 = 3x1 + 1x2 1x1 + 2x2 3 1 1 2 . so the rank of Also note that det 3 1 1 2 = 3 · 2 − 1 · 1 = 5 ≠ 0. pages 133 and 135. 6 2 = 3 · 2 − 6 · 1 = 0. example 2 The linear system of equations 3x1 + 6x2 = y1 1x1 + 2x2 = y2 3x1 + 6x2 1x1 + 2x2 . as can be seen on 3 1 1 2 is 2. can be rewritten as follows: The two vectors 3 1 and y1 y2 1 2 = = 3 1 1 2 x1 x2 = x1 + x2 are linearly independent.

y2 = 0x1 + 1x2 = 0 = x1 0 + x 2 1 1 0 The two vectors 0 and 1 are linearly independent. as we discovered 0 0 1 on page 137. so the rank of 0 0 0 1 0 is 2.example 3 1x1 + 0x2 = y1 The linear system of equations 0x1 + 1x2 = y2 0x1 + 0x2 = y3 y1 can be rewritten as: y3 1x1 + 0x2 0x1 + 0x2 1 0 0 1 0 x1 x2 . 0 194 Chapter 7 Linear Transformations . that is y1 y3 1 0 1x1 + 0x2 0x1 + 0x2 0 0 y2 = 0x1 + 1x2 . The system could also be rewritten like this: y1 1x1 + 0x2 1 0 1 0 0 0 0 x1 x2 x3 y2 = 0x1 + 1x2 = 0 y3 0x1 + 0x2 0 1 Note that det 0 0 0 1 0 0 0 = 0.

can be rewritten as follows: x1 y1 y2 = 1x1 + 0x2 + 3x3 + 1x4 0x1 + 1x2 + 1x3 + 2x4 1 0 + x2 0 1 + x3 3 1 = 1 0 1 2 0 1 3 1 1 2 x2 x3 x4 = x1 + x4 The rank of 1 0 3 1 0 1 1 2 is equal to 2. rank 195 . but no formal proof will be given in this book. that is . This is indeed so. as we’ll see on page 203. The system could also be rewritten like this: y1 y2 y3 y4 = 1x1 + 0x2 + 3x3 + 1x4 0x1 + 1x2 + 1x3 + 2x4 0 0 = 1 0 0 0 0 1 0 0 3 1 0 0 1 x1 2 x2 0 x3 0 x4 1 Note that det 0 0 0 0 1 0 0 3 1 0 0 1 2 0 0 = 0.example 4 The linear system of equations y1 y2 = 1x1 + 0x2 + 3x3 + 1x4 0x1 + 1x2 + 1x3 + 2x4 1x1 + 0x2 + 3x3 + 1x4 = y1 0x1 + 1x2 + 1x3 + 2x4 = y2 . The four examples seem to point to the fact that a11 det a 21 an1 a12 a 22 an2 a1n a 2n ann = 0 is the same as rank a11 a 21 an1 a12 a 22 an2 a1n a 2n ann ≠ n.

take a look at the following matrix: 1 2 3 4 5 4 8 6 12 It’s immediately clear that the third column of this matrix is equal to the first column times 4. This leaves two linearly independent vectors (the first two columns). you’ll have to buckle down and actually calculate the rank. The resulting matrices can then be multiplied with any arbitrary matrix in such a way that the number of linearly independent columns becomes obvious. In those cases. Now look at this matrix: 1 0 0 0 3 5 It should be obvious right from the start that these vectors form a linearly independent set. Though this might seem like “cheating” at first. which means this matrix has a rank of 2. then we’ll establish a good * way of thinking . it’s not too hard! First we’ll explain the  ? Problem . An elementary matrix is created by starting with an identity matrix and performing exactly one of the elementary row operations used for Gaussian elimination (see Chapter 4). so we know that the rank of this matrix is also 2.Calculating the Rank of a Matrix So far. we’ve only dealt with matrices where the rank was immediately apparent or where we had previously figured out how many linearly independent vectors made up the columns of that matrix. ? Problem Calculate the rank of the following 2×4 matrix: 1 0 0 1 3 1 1 2 * Way of Thinking Before we can solve this problem. 196 Chapter 7 Linear Transformations . Of course there are times when this method will fail you and you won’t be able to tell the rank of a matrix just by eyeballing it. But don’t worry. we need to learn a little bit about elementary matrices. and then finally we’ll tackle the  ! Solution . For example. these techniques can actually be very useful for calculating ranks in practice.

With this information under our belts. If we multiply the matrix to the right of A.) 0 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 a11 a 21 a 31 a41 a12 a 22 a 32 a42 a13 a 23 a 33 a43 0·a12 + 0·a22 + 0·a32 + 1·a42 0·a12 + 1·a22 + 0·a32 + 0·a42 0·a12 + 0·a22 + 1·a32 + 0·a42 1·a12 + 0·a22 + 0·a32 + 0·a42 0·a13 + 0·a23 + 0·a33 + 1·a43 0·a13 + 1·a23 + 0·a33 + 0·a43 0·a13 + 0·a23 + 1·a33 + 0·a43 1·a13 + 0·a23 + 0·a33 + 0·a43 0·a11 + 0·a21 + 0·a31 + 1·a41 = 0·a11 + 1·a21 + 0·a31 + 0·a41 0·a11 + 0·a21 + 1·a31 + 0·a41 1·a11 + 0·a21 + 0·a31 + 0·a41 a41 = a 21 a 31 a11 a42 a 22 a 32 a12 a43 a 23 a 33 a13 rank 197 . we can state the following four useful facts about an arbitrary matrix A: a11 a 21 am1 Fact 1 a12 a 22 am2 a1n a 2n amn Multiplying the elementary matrix 1 0 0 0 0 0 1 0 Column i 0 1 0 0 0 0 0 1 Row i Row j Column j to the left of an arbitrary matrix A will switch rows i and j in A. then the columns will switch places in A instead. • Example 1 (Rows 1 and 4 are switched.

0 0 1 Row i 198 Chapter 7 Linear Transformations .• Example 2 (Columns 1 and 3 are switched.) a11 a 21 a 31 a41 a12 a 22 a 32 a42 a13 a 23 a 33 a43 0 0 1 0 1 0 1 0 0 a11·0 + a12·1 + a13·0 a 21·0 + a 22·1 + a 23·0 a 31·0 + a 32·1 + a 33·0 a41·0 + a42·1 + a43·0 a11·1 + a12·0 + a13·0 a 21·1 + a 22·0 + a 23·0 a 31·1 + a 32·0 + a 33·0 a41·1 + a42·0 + a43·0 a11·0 + a12·0 + a13·1 = a 21·0 + a 22·0 + a 23·1 a 31·0 + a 32·0 + a 33·1 a41·0 + a42·0 + a43·1 a13 = a12 a11 a 23 a 22 a 21 a 33 a 32 a 31 a43 a42 a41 Fact 2 Multiplying the elementary matrix 1 0 0 0 k 0 Column i to the left of an arbitrary matrix A will multiply the ith row in A by k. Multiplying the matrix to the right side of A will multiply the ith column in A by k instead.

• Example 1 (Row 3 is multiplied by k.) 1 0 0 0 0 1 0 0 0 0 k 0 0 a11 0 a 21 0 a 31 1 a41 a12 a 22 a 32 a42 a13 a 23 a 33 a43 1·a12 + 0·a22 + 0·a32 + 0·a42 0·a12 + 1·a22 + 0·a32 + 0·a42 0·a12 + 0·a22 + k·a32 + 0·a42 0·a12 + 0·a22 + 0·a32 + 1·a42 1·a13 + 0·a23 + 0·a33 + 0·a43 0·a13 + 1·a23 + 0·a33 + 0·a43 0·a13 + 0·a23 + k·a33 + 0·a43 0·a13 + 0·a23 + 0·a33 + 1·a43 1·a11 + 0·a21 + 0·a31 + 0·a41 = 0·a11 + 1·a21 + 0·a31 + 0·a41 0·a11 + 0·a21 + k·a31 + 0·a41 0·a11 + 0·a21 + 0·a31 + 1·a41 a11 = a 21 a41 a12 a 22 a42 a13 a 23 a43 ka 31 ka 32 ka 33 • Example 2 (Column 2 is multiplied by k.) a11 a 21 a 31 a41 a12 a 22 a 32 a42 a13 a 23 a 33 a43 1 0 0 0 k 0 0 0 1 a11·0 + a12·k + a13·0 a 21·0 + a 22·k + a 23·0 a 31·0 + a 32·k + a 33·0 a41·0 + a42·k + a43·0 a11·0 + a12·0 + a13·1 a 21·0 + a 22·0 + a 23·1 a 31·0 + a 32·0 + a 33·1 a41·0 + a42·0 + a43·1 a11·1 + a12·0 + a13·0 = a 21·1 + a 22·0 + a 23·0 a 31·1 + a 32·0 + a 33·0 a41·1 + a42·0 + a43·0 a11 = a 21 a 31 a41 ka12 a13 ka 22 a 23 ka 32 a 33 ka42 a43 rank 199 .

Fact 3 Multiplying the elementary matrix 1 0 0 0 0 1 k 0 0 0 1 0 0 0 0 1 Row i Row j Column i Column j to the left of an arbitrary matrix A will add k times row i to row j in A. • Example 1 (k times row 2 is added to row 4.) 1 0 0 0 0 1 0 k 0 0 1 0 0 a11 0 a 21 0 a 31 1 a41 a12 a 22 a 32 a42 a13 a 23 a 33 a43 1·a12 + 0·a22 + 0·a32 + 0·a42 0·a12 + 1·a22 + 0·a32 + 0·a42 0·a12 + 0·a22 + 1·a32 + 0·a42 0·a12 + k·a22 + 0·a32 + 1·a42 1·a13 + 0·a23 + 0·a33 + 0·a43 0·a13 + 1·a23 + 0·a33 + 0·a43 0·a13 + 0·a23 + 1·a33 + 0·a43 0·a13 + k·a23 + 0·a33 + 1·a43 1·a11 + 0·a21 + 0·a31 + 0·a41 = 0·a11 + 1·a21 + 0·a31 + 0·a41 0·a11 + 0·a21 + 1·a31 + 0·a41 0·a11 + k·a21 + 0·a31 + 1·a41 a11 = a 21 a 31 a41+ ka 21 a12 a 22 a 32 a42+ ka 22 a13 a 23 a 33 a43+ ka 23 200 Chapter 7 Linear Transformations . Multiplying the matrix to the right side of A will add k times column j to column i instead.

since elementary matrices are invertible.) a11 a 21 a 31 a41 a12 a13 a 22 a 23 a 32 a 33 a42 a43 1 0 k 0 1 0 0 0 1 a11·0 + a12·1 + a13·0 a 21·0 + a 22·1 + a 23·0 a 31·0 + a 32·1 + a 33·0 a41·0 + a42·1 + a43·0 a11·0 + a12·0 + a13·1 a 21·0 + a 22·0 + a 23·1 a 31·0 + a 32·0 + a 33·1 a41·0 + a42·0 + a43·1 a11·1 + a12·0 + a13·k = a 21·1 + a 22·0 + a 23·k a 31·1 + a 32·0 + a 33·k a41·1 + a42·0 + a43·k a11 + ka13 = a12 a13 a 21 + ka 23 a 22 a 23 a 31 + ka 33 a 32 a 33 a41 + ka43 Fact 4 a42 a43 The following three m×n matrices all have the same rank: 1. rank 201 . The matrix: a11 a 21 am1 a12 a 22 am2 a1n a 2n amn 2. multiplying A by any elementary matrix—on either side—will not change A’s rank. The right product using an invertible n×n matrix: a11 a 21 am1 a12 a 22 am2 a1n a 2n amn c11 c21 cn1 c12 c22 cn2 c1n c2n cnn In other words.• Example 2 (k times column 3 is added to column 1. The left product using an invertible m×m matrix: b11 b21 bm1 b12 b22 bm2 b1m b2m bmm a11 a 21 am1 a12 a 22 am2 a1n a 2n amn 3.

! Solution The following table depicts calculating the rank of the 2×4 matrix: 1 0 0 1 3 1 1 2 Begin with 1 0 0 1 3 1 1 2 Add (−1 · column 2) to column 3 1 1 0 0 1 3 1 1 2 0 0 0 0 1 0 0 0 −1 1 0 0 0 0 1 = 1 0 0 1 3 0 1 2 Add (−1 · column 1) to column 4 1 1 0 0 1 3 0 1 2 0 0 0 0 1 0 0 0 0 1 0 −1 0 0 1 = 1 0 0 1 3 0 0 2 Add (−3 · column 1) to column 3 1 1 0 0 1 3 0 0 2 0 0 0 0 1 0 0 −3 0 1 0 0 0 0 1 = 1 0 0 1 0 0 0 2 Add (−2 · column 2) to column 4 1 1 0 0 1 0 0 0 0 2 0 0 0 1 0 0 0 0 1 0 0 −2 0 1 = 1 0 0 1 0 0 0 0 202 Chapter 7 Linear Transformations .

This means it has a rank of 2. this explanation is a bit vague. 1 0 0 1 3 1 1 2 and 1 0 0 1 0 0 0 0 have 1 0 One look at the simplified matrix is enough to see that only and 0 1 are linearly independent among its columns. and so does our initial matrix. we know that both the same rank. The relationship between Linear Transformations and Matrices We talked a bit about the relationship between linear transformations and matrices on page 168. We said that a linear transformation from Rn to Rm could be written as an m×n matrix: a11 a 21 am1 a12 a 22 am2 a1n a 2n amn As you probably noticed. = a11 a12 a 21 a 22 am1 am2 a1n a 2n amn x1 x2 xn is an arbitrary element in Rn and f is a function from Rn to Rm. The more exact relationship is as follows: The relationship between linear Transformations and matrices x1 If x2 xn then f is a linear transformation from Rn to Rm if and only if x1 f x2 xn for some matrix A. The relationship between Linear Transformations and Matrices 203 .Because of Fact 4.

.

8 eigenvalues and eigenvectors y 4 z x Times 2 ba ck w Time s ar d .

begin! 206 Chapter 8 .Yurino from Hanamichi University! And Jumonji from nanhou University! He looks tough. ready! St ar e i'll have to give it my all.

Pow i’ve got to show them how strong i can be! Snuh i have to win this! Y a a a a a enough! ...Sma Baff ck S e y a a a a ngh.. Sh ud de r Guh..

Good match. very much...nanhou University! Thank you... Damn... 208 Chapter 8 .

Yeah. really? Don't worry about it. though. eigenvalues and eigenvectors 209 . You’ll do better next time.i'm sorry about the match... My brother said you fought well... i know it! i'm sorry! You're completely right! Sulking won't accomplish anything.

for example. but i'll try to be as concrete as i can. it's a pretty abstract topic. Finding the p power of an n×n matrix. i’m ready for anything! basics Fundamentals Studying eigenvalues and eigenvectors comes in handy when doing physics and statistics. today's our last lesson.. 210 Chapter 8 th i appreciate it! . Okay.Anyway.. And i thought we'd work on eigenvalues and eigenvectors. Prep Matrices Vectors Linear transformations eigenvalues and eigenvectors They also make these kinds of problems much easier.

Find the image of c1 3 1 + c2 1 2 using the linear transformation determined by the 2×2 matrix 8 −3 2 1 (where c1 and c2 are real numbers)... like this? Hmm. Okay. exactly! So.. Like this? So close! Oh. first problem.the answer can be expressed using multiples of the original two vectors? What Are eigenvalues and eigenvectors? 211 ..What Are eigenvalues and eigenvectors? What do you say we start off with a few problems? Sure.

212 Chapter 8 eigenvalues and eigenvectors .... 8 2 −3 1 Like so... Oh..transforms All points on the x1 x2 plane.That’s right! So you could say that the linear transformation equal to the matrix .

. What Are eigenvalues and eigenvectors? 213 .Let's move on to another problem. 1 0 0 0 4 0 0 0 using Hmm Find the image of c1 0 + c2 1 + c3 0 1 0 2 0 0 the linear transformation determined by the 3×3 matrix 0 −1 (where c1. Like this? Correct. and c3 are real numbers). c2. So this solution can be expressed with multiples as well..

.So you could say that the linear transformation equal to the matrix . Like this...transforms every point in the x1 x2 x3 space..transforms every point in the x1 x2 x3 space. i get it! Time s 4 214 Chapter 8 eigenvalues and eigenvectors Times 2 Loo ba k ck i n g ..... 4 0 0 0 2 0 0 0 −1 .

Let's have a look at the definition... 0 −1 Eigenvalue λ = 7. Rn x1 x2 xn Rn x1 x2 xn λ So the two examples could be summarized like this? Matrix exactly! 8 2 −3 1 4 0 0 0 2 0 0 You can generally never find more than n different eigenvalues and eigenvectors for any n×n matrix.. rs to ve c Eigenvalues and eigenvectors x1 x2 xn is equal to λ If the image of a vector a11 a 21 an1 a12 a 22 an2 x1 x2 xn a1n a 2n ann through the linear transformation determined by the matrix x1 x2 xn . g s an d e i g e n keeping those examples in mind. and is said to be an eigenvector corresponding to the eigenvalue λ. 2. Oh. −1 3 1 1 2 the vector corresponding to λ = 4 the vector corresponding to λ = 2 the vector corresponding to λ = 1 1 0 0 0 1 0 0 0 1 Eigenvector the vector corresponding to λ = 2 .. The zero vector can never be an eigenvector. λ is said to be an eigenvalue to the matrix. 2 the vector corresponding to λ = 7 λ = 4.

Let's start off with the relationship.. Okay.. between the determinant and eigenvalues of a matrix.Calculating eigenvalues and eigenvectors The 2×2 matrix 8 −3 2 1 will do fine as an example. The relationship between the determinant and eigenvalues of a matrix λ is an eigenvalue of the matrix a11 a 21 an1 a12 a 22 an2 a1n a 2n ann a11 − λ if and only if det a 21 an1 a 22 − λ an2 a12 a1n a 2n ann − λ =0 216 Chapter 8 eigenvalues and eigenvectors . Let's have a look at calculating these vectors and values.

.. The values are seven and two? Correct! Calculating eigenvalues and eigenvectors 217 . Okay. So. it's pretty cool. give it a shot.This means that solving this characteristic equation gives us all eigenvalues corresponding to the underlying matrix.. Go ahead..

Let’s plug our value into the formula: 8−2 2 −3 1−2 x1 x2 = 6 2 −3 −1 x1 x2 = 6x1 − 3x2 2x1 − x2 = [2x1 − x2] 3 1 = 0 0 Done! This means that x2 = 2x1. 218 Chapter 8 eigenvalues and eigenvectors . Let’s plug our value into the formula: 8−7 2 −3 1−7 x1 x2 = 1 −3 2 −6 x1 x2 = x1 − 3x2 2x1 − 6x2 = [x1 − 3x2] 1 2 = 0 0 This means that x1 = 3x2. For example. Problem 2 Find an eigenvector corresponding to λ = 2. which leads us to our eigenvector x1 x2 = c2 2c2 = c2 1 2 where c2 is an arbitrary nonzero real number.Finding eigenvectors is also pretty easy. which leads us to our eigenvector x1 x2 = 3c1 c1 = c1 3 1 where c1 is an arbitrary nonzero real number. that is 8−λ 2 −3 1−λ x1 x2 = 0 0 Problem 1 Find an eigenvector corresponding to λ = 7. we can use our previous values in this formula: 8 −3 2 1 x1 x2 =λ x1 x2 .

refer to page 91 to see why exists. Calculating the pth power of an n x n matrix 219 . Let's choose c1 = c2 = 1.Calculating the pth Power of an nxn Matrix We've already found the eigenvalues and eigenvectors of the matrix it's finally time to tackle today's real problem! Finding the pth power of an n×n matrix. Makes sense.. Using the two calculations above. for simplicity’s sake. Let’s multiply to the right of both sides of the equation.. 8 −3 2 1 So let's just build on that example.

okay.....this it? Looking at your calculations. would you say this relationship might be true? Uhhh.Try using the formula to calculate 8 −3 2 1 2 Hmm. 220 Chapter 8 eigenvalues and eigenvectors .. Yep! Yay! is..

it actually is! This formula is very useful for calculating any power of an n×n matrix that can be written in this form.. When p = 1. The eigenvector corresponding to λ1 The eigenvector corresponding to λ2 Got it! The eigenvector corresponding to λn Oh.. nice! . and by the way. we say that the formula diagonalizes the n×n matrix a11 a 21 an1 a12 a 22 an2 a1n a 2n ann And that's it! The right side of the equation is the diagonalized form of the middle matrix on the left side.

That was the last lesson! Yeah. not at all! How could i possibly have been tired after all that wonderful food you gave me? i should be thanking you! i'll miss these sessions.. and you've been awfully tired because of your karate practice. thanks to you. though.. you know! My afternoons will be so lonely from now on. i know you're busy. 222 Chapter 8 eigenvalues and eigenvectors . thanks for helping me out. How do you feel? Did you get the gist of it? Awesome! really.

. Hmm? Yeah. sounds like fun! So when would you like to go? ......we could go out sometime.to look for math books. if you don't have anything else to do.. Sure. or something....Well... you know.

then diagonalization is possible.Multiplicity and Diagonalization We said on page 221 that any n×n matrix could be expressed in this form: The eigenvector corresponding to λ1 −1 a11 a 21 an1 a12 a 22 an2 a1n a 2n ann = x11 x21 xn1 x12 x22 xn2 x1n x2n xnn λ1 0 0 0 λ2 0 0 0 λn x11 x21 xn1 x12 x22 xn2 x1n x2n xnn The eigenvector corresponding to λ2 The eigenvector corresponding to λn This isn’t totally true. The situation becomes more complicated when we have to deal with eigenvalues that have multiplicity greater than 1. (x + 2) has 2. For example. and x has 1. The multiplicity of any polynomial root reveals how many identical copies of that same root exist in the polynomial. in the polynomial f(x) = (x − 1)4(x + 2)2x. 224 Chapter 8 eigenvalues and eigenvectors . as the concept of multiplicity1 plays a large role in whether a matrix can be diagonalized or not. We will therefore look at a few examples involving: • • Matrices with eigenvalues having multiplicity greater than 1 that can be diagonalized Matrices with eigenvalues having multiplicity greater than 1 that cannot be diagonalized 1. if all n solutions of the following equation a11 − λ det a 21 an1 a12 a1n a 2n ann − λ a 22 − λ an2 =0 are real and have multiplicity 1. the factor (x − 1) has multiplicity 4. For instance.

Multiplicity and Diagonalization 225 . = (1 − λ)(1 − λ)(3 − λ) + 0 · (−1) · (−2) + 0 · 1 · 0 − 0 · (1 − λ) · (−2) − 0 · 1 · (3 − λ) − (1 − λ) · (−1) · 0 = (1 − λ)2(3 − λ) = 0 λ = 3. Find all eigenvalues and eigenvectors of the matrix. 1 Note that the eigenvalue 1 has multiplicity 2. The eigenvalues λ of the 3×3 matrix 1 1 −2 0 1 0 0 −1 3 1−λ are the roots of the characteristic equation: det 1−λ det 1 −2 0 1−λ 0 0 −1 3−λ 1 −2 0 1−λ 0 0 −1 3−λ = 0.A Diagonalizable Matrix with an Eigenvalue Having Multiplicity 2 ? Problem Use the following matrix in both problems: 1 1 −2 0 1 0 0 −1 3 1. Express the matrix in the following form: x11 x21 x31 x12 x22 x32 x13 x23 x33 λ1 0 0 0 λ2 0 0 0 λ3 x11 x21 x31 x12 x22 x32 x13 x23 x33 −1 ! Solution 1. 2.

a.

The eigenvectors corresponding to λ = 3 Let’s insert our eigenvalue into the following formula: 1 1 −2 0 1 0 0 −1 3 x1 x1 1−λ 1 −2 0 1−λ 0 0 −1 3−λ x1 0

x2 = λ x2 , that is x3 x3

x2 = 0 x3 0

This gives us: 1−3 1 −2 0 1−3 0 0 −1 3−3 x1 x2 = x3 −2 1 −2 0 −2 0 0 −1 0 x1 −2x1 0

x2 = x1 − 2x2 − x3 = 0 x3 −2x1 0

The solutions are as follows: x1 = 0 x3 = −2x2 x1 and the eigenvector 0 0

c1 = c1 1 x2 = x3 −2c1 −2

where c1 is a real nonzero number. The eigenvectors corresponding to λ = 1 Repeating the steps above, we get 1−1 1 −2 0 1−1 0 0 −1 3−1 x1 x2 = x3 0 1 −2 0 0 0 0 −1 2 x1 0 0

b.

x2 = x1 − x 3 = 0 x3 −2x1 + 2x3 0

and see that x3 = x1 and x2 can be any real number. The eigenvector consequently becomes x1 x3 c1 c1 1 1 0 0

x2 = c2 = c1 0 + c2 1

where c1 and c2 are arbitrary real numbers that cannot both be zero.

226

Chapter 8

eigenvalues and eigenvectors

3.

We then apply the formula from page 221: The eigenvector corresponding to 3

1 1 −2

0 1 0

0 −1 = 3

0 1 −2

1 0 1

0 1 0

3 0 0

0 1 0

0 0 1

0 1 −2

1 0 1

0 1 0

−1

The linearly independent eigenvectors corresponding to 1

A Non-Diagonalizable Matrix with a Real Eigenvalue Having Multiplicity 2
? Problem

Use the following matrix in both problems: 1 −7 4
1. 2.

0 0

0 3

1 −1

Find all eigenvalues and eigenvectors of the matrix. Express the matrix in the following form: x11 x21 x31 x12 x22 x32 x13 x23 x33 λ1 0 0 0 λ2 0 0 0 λ3 x11 x21 x31 x12 x22 x32 x13 x23 x33
−1

! Solution
1.

The eigenvalues λ of the 3×3 matrix 1 −7 4 0 0 0 3

1 −1 1−λ −7 4 0 1−λ 0 0 −1 3−λ
227

are the roots of the characteristic equation: det

= 0.

Multiplicity and Diagonalization

1−λ det −7 4

0 1−λ 0

0 −1 3−λ

= (1 − λ)(1 − λ)(3 − λ) + 0 · (−1) · 4 + 0 · (−7) · 0 − 0 · (1 − λ) · 4 − 0 · (−7) · (3 − λ) − (1 − λ) · (−1) · 0 = (1 − λ)2(3 − λ) = 0 λ = 3, 1 Again, note that the eigenvalue 1 has multiplicity 2. The eigenvectors corresponding to λ = 3 Let’s insert our eigenvalue into the following formula: 1 −7 4 0 1 0 0 −1 3 x1 x1 1−λ −7 4 0 1−λ 0 0 −1 3−λ x1 0

a.

x2 = λ x2 , that is x3 x3

x2 = 0 x3 0

This gives us 1−3 −7 4 0 1−3 0 0 −1 3−3 x1 x2 = x3 −2 −7 4 0 −2 0 0 −1 0 x1 −2x1 0

x2 = −7x1 − 2x2 − x3 = 0 x3 4x1 0

The solutions are as follows: x1 = 0 x3 = −2x2 x1 and the eigenvector 0 0

c1 = c1 1 x2 = x3 −2c1 −2

where c1 is a real nonzero number.

228

Chapter 8

eigenvalues and eigenvectors

b.

The eigenvectors corresponding to λ = 1 We get 1−1 −7 4 0 1−1 0 0 −1 3−1 x1 x2 = x3 0 −7 4 0 0 0 0 −1 2 x1 0 0 x2 = −7x1 − x3 = 0 x3 4x1 + 2x3 0

and see that

x3 = −7x1 x3 = −2x1

But this could only be true if x1 = x3 = 0. So the eigenvector has to be x1 x2 x3 0 0 0 0

= c2 = c2 1

where c2 is an arbitrary real nonzero number. Since there were no eigenvectors in the form x12 x32 x13 x33

3.

c2 x22 + c3 x23

for λ =1, there are not enough linearly independent eigenvectors to express 1 −7 4 0 0 0 3

x11 x21 x31

x12 x22 x32

x13 x23 x33

λ1 0 0

0 λ2 0

0 0 λ3

x11 x21 x31

x12 x22 x32

x13 x23 x33

−1

1 −1 in the form

It is important to note that all diagonalizable n×n matrices always have n linearly independent eigenvectors. In other words, there is always a basis in Rn consisting solely of eigenvectors, called an eigenbasis.

Multiplicity and Diagonalization

229

. don't be like that! Stop it! Let me go! We just want to get to know you better.. Hey there.Looks like i got here a bit early. Misa! . been waiting long? reij— Kyaa! That voice! Aww.

.. my girlfriend in middle school. let me go! We just want to hang out. What if it happens all over again? On my first date with Yuki... tr em b le but. i already have a date.! i have to do something. And fast! Those jerks.. Yuki! .... Hey...

Cra s h Seriously? One kick? What a wuss......who are you? Kick Guh? Haha Time to go. You. stop it! i have to do something! Uh. Lemme go! You there! .Those jerks....

Messing with that guy is suicide! Unng. who is it now? Oh no! The legendary leader of the Hanamichi karate club. Can't you see she doesn't want to go with you? Y-you're. “The Hanamichi Hammer!” in the flesh..Stop.. Jeez. Let’s get outta here! Hanamichi Hammer? Moan ....but i think your boyfriend needs medical attention... T-thank you! Don't worry about it... ..

..don't think i can see you anymore. i’ll show them— i’m so much stronger now— Come on. i couldn't do anything...... i. but...Listen.. k Yan This time will be different! Help! LeT Her GO! epilogue 235 . .. i'm so sorry.. it just wasn’t enough. . i'm glad you stood up for me.. not this time. sweetie....

Just who do you think you are?! Baff reiji! Hey! Stop right there. let's go.Sm a ck Wha—? come on. all of you! Let’s get him! . look! He thinks he's a hero! Stay away from her. ! Haha.

reiji! Ooph Th u m p Stop it! Please! Let me go! enough! Stubborn...a ck Sm run! You little. huh? rab G . be careful.

.Attacking my little sister.. a ck Cr The Hanamichi Hammer! it's ichinose! Mommy! run! Sensei? He’s out cold. . you have given me little choice. are we? Tetsuo! i don't like excessive violence..but in attacking Misa..

. Um. epilogue 239 .no problem..reiji? reiji. Thank you.....Misa told me what happened. wake up. reiji! You’re okay! Whoa! Yurino. but i don't deserve your thanks.

...i couldn't help Misa.. but— reiji! He’s right..i couldn't even help myself.. even though the fight itself was unnecessary. i haven't changed at all! i'm still a weakling! Well. Misa. you may not be a black belt yet... i don't know what to say. You should be proud! but you're definitely no weakling.. Thank you.. That kind of courage is admirable. 240 epilogue . Putting Misa's safety before your own shows great courage.

by the way..Thank you for everything! Ah.... epilogue 241 .... Well. Sensei. Huh?! i. S-sure. i guess it's okay.. would you consider doing me another favor? Thanks.. Uh... What the—! I thought I was pretty clear about the rules..Misa's not a kid anymore. Heh.

Math... So. i mean. if he doesn't graduate soon. What? He could really use the help. plus and minus? Sounds like you’ll need more lunches! 242 epilogue . too.. What do you say? Sure! Of course! Great! Let's start off with plus and minus.. being in his sixth year and all. then! it'd mean a lot to me. too. Um.i'd like you to teach me.

They include: Appendix Appendix Appendix Appendix Appendix A: Workbook B: Vector Spaces C: Dot Product D: Cross Product E: Useful Properties of Determinants There's more! Updates Visit http://www. .com/linearalgebra.Online resources The Appendixes The appendixes for The Manga Guide to Linear Algebra can be found online at http://www.nostarch.com/linearalgebra for updates. errata. and other information.nostarch.

.i still don't get it! no need to get violent. bro! Sn a p ! .. You can do it.

144 i i (imaginary unit). 216–218 finding pth power of n×n matrix. 25 computer graphics systems. 210–215 relation of linear algebra to. 138–139. 55–60 complex numbers. 91. 25 imaginary unit (i). 50–61 onto and one-to-one. 140–148. 95 diagonalization. 24 elementary matrices. 88. 46–47 overview. 39. 25–26 identity matrices. 126 combinations. of vectors. 127–130 graphs. 108 geometric interpretation. 184 conventional linear transformations. linear transformations used by. 161–162 Cramer’s rule. 67 in sets. 225 diagonal matrices. linear. 221. 111–112 D dependence. 30. 185 θ (theta). 40–44 overview. 149–162 dimension theorem for linear transformations. 82–84. calculating inverse matrices using. 39. 44–45 and images. 143 determinants calculating. multiplicity and. of vectors. 219–221. 39 domain and range. 229 eigenvalues calculating. 224–229 overview. 189–192 imaginary numbers. 40–43 inverse. 156–158 binomial coefficients. 24 eigenvectors calculating. 111–112 overview. 216–218 finding pth power of n×n matrix. 196 elements in matrices. 125 axis. 48–49 linear transformations. 127 b basis. 96–105. writing as matrices. 80–81 dimensions. 88–89. 174. 60 C co-domain. 32 equations. 25–26 . 219–221. 45 cofactor matrices. 69 equivalence. 92 images and functions.index Special Characters and numbers 3-D projections of linear transformations. 224–229 overview. 70 with vectors. 108–111 column vectors. expressing with vectors. 189–192 domain. 40–43 G Gaussian elimination. 29 F functions defined. 110 cofactors. 210–215 relation of linear algebra to. 224–229 diagonalizing matrices. 184 coordinates. 44–45 e eigenbasis. 180 A addition with matrices. 35–39 f(x). 135.

70–76 determinants. 138–139. 79 matrices calculations with. in sets. 62–69 rank of. 193–203 relation of linear algebra to. 27 r range. 80–81 diagonalizable. 108–111 overview. 146–147 one-to-one functions. 46–47 P permutations. 72–76. calculating. 94 irrational numbers. 78 upper triangular. 196–203 relation of linear algebra to. 30 one-dimensional dependence. 185 planes. 48–49 inverse matrices calculating using Gaussian elimination. 179 translation. 189–192 functions and. 167 linear spans. 108–111 overview. 82–83 with matrices. 25–26 O objects. 125 overview. 193–195 rational numbers. 127 polynomial roots. 86–87 invertible matrices. 203 symmetric. 215 identity. 86–87 lower triangular. 189–192 L linear algebra. 88–94 calculating using cofactors. 46–47 onto functions. 79 triangular matrices and. 143. 79 transpose. 143. linear. 69 zero. 143. 146–147 linear map. 77 multiplicity. 154–155 linear systems. 180–181 246 index . 173–177 dimension theorem for. 166–173 rank. 168. 50–61 overview. 132–139. 224 propositions. 79 multiplication with. 82–84 inverse calculating using Gaussian elimination. 135. 95–105. 227–229 number systems. 203 rotation. 80 identity matrices and. 79 writing systems of equations as. 143 linear independence. 146–147 integers. 25 inverse functions. 55–60 perspective projection. solving with Cramer’s rule. 196–203 overview. 24 relationship with matrices. and diagonalization. 88–94 calculating using cofactors. 182–184 lower triangular matrices. 111–112 diagonal. 225–227 eigenvalues and eigenvectors. 80–81 with identity matrices. 24 relationship with linear transformations. 224–229 multiplication with diagonal matrices. 126 K kernel. 25 scaling.implication. 72–76 with vectors. 132–139. 82 overview. 25 real numbers. overview. 132–139. 79 M main diagonal diagonal matrices and. 135. 138–139. 67 symmetric matrices and. 125 n natural order. 103 non-diagonalizable matrices. 27–28 independence. 185 applications of. 44–45 rank of matrices. 128 points. 9–20 linear dependence. 111–112 linear transformations 3-D projections of. 25 Rn. 143 one-dimensional independence. 167 linear operation.

179. 101 theta (θ). 180–181. 78 triangular matrices. 72 with vectors. 132–139 overview. linear. 126 rules of determinants. 125–126 dimensions of. 140–148 calculating. 75 overview. 184 row vectors. 78 symmetric matrices. 40–43 for imaginary units. 98 scalar multiplication with matrices. 129 Z zero matrices. 125 symbols for equivalence. 28 of sets. 184 set theory sets.rotating linear transformations. 25–26 for inverse functions. 29 for functions. 127–130 linear independence. 77 index 247 . 67 straight lines. 125 scaling linear transformations. 33 for transpose matrices. 39 S Sarrus’ rule. 149–162 geometric interpretation of. 69 T target set. 79 systems of equations. 30–31 set symbols. 101 functions as. 127 subscripts. 33–34 subspaces. 39 f(x). 32 subsets. 71 with vectors. 66 subsets. 33–34 square matrices multiplying. writing as matrices. 180 3-D projections of linear transformations. See linear transformations translating linear transformations. 32 for subsets. 49 for propositions. 79 U upper triangular matrices. 39 term indexes. 79 V vectors basis. 24 vector space. 185 transformations. 150–155 subtraction with matrices. 116–124 relation of linear algebra to. 182–184 transpose matrices.

notes .

notes .

notes .

Ltd. Homepage: http://www. He received a master’s degree from Kyushu Institute of Design (known as Kyushu University today). of Tokyo. scenario writer.About the Author Shin Takahashi was born 1972 in Niigata. Japan. Once each title is translated into English. he is now an author specializing in technical literature. we rewrite and edit the translation as necessary and have an expert review each volume. and expert scientist or mathematician. Each title in the best-selling Manga Guide series is the product of the combined work of a manga illustrator. one of Japan’s oldest and most respected scientific and technical book publishers. .jp/ Production Team for the Japanese edition scenario: artist: dtp: re_akino Iroha Inoue Emi Oda How This book Was Made The Manga Guide series is a co-publication of No Starch Press and Ohmsha. Having previously worked both as an analyst and as a seminar leader. The result is the English version you hold in your hands.takahashishin.

More Manga Guides Find more Manga Guides at your favorite bookstore. and learn more about the series at http://www.nostarch.com/manga. .

nostarch. dimension. and linear span Practical apPlications of linear algebra in fields like computer graphics.com/manga . and engineEring Real math. real romance.95 T H E F I N E ST I N G E E K E N T E RTA I N M E N T ™ ($25. calculating determinants. RecomMended. and finding eigenvectors and eigenvalues.nostarch. and bases Using GausSian elimination to calculate inverse matrices Subspaces. and real action come together like never before in The Manga Guide to Linear Algebra. you’lL learn about: Basic vector and matrix operations such as adDition. independence. and multiplication Linear dependence. FolLow along in The Manga Guide to Linear Algebra as Reiji takes Misa from the absolute basics of this tricky subject through mind-bending operations like performing linear transformations. the girl of his dreams.95 CDN) SHELVE IN: MATHEMATICS/ALGEBRA w w w. understandable. Reiji transforms abstract concepts into something concrete. $24. and even fun. subtraction. As you folLow Misa through her linear algebra crash course. Luckily. cryptography.” — Otaku USA Magazine Real Math! Romance! Karate! Reiji wants two things in life: a black belt in karate and Misa.” — Choice Magazine “Stimulus for the next generation of scientists. Misa’s big brother is the captain of the university karate club and is ready to strike a deal: Reiji can join the club if he tutors Misa in linear algebra.PRaise for the manga guide series “Highly RecomMended.com Find more Manga Guides at wWw.” — Scientific computing “A great fit of FOrm and subject. With memorable examples like miniature golf games and karate tournaments.

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->