ECE 174: Introduction to Linear and Nonlinear Optimization with Applications

Fall 2021

Homework # 2
Due: October 30, 11:59 pm, via Gradescope

Collaboration Policy: This homework set allows limited collaboration. You are expected
to try to solve the problems on your own. You may discuss a problem with other students to
clarify any doubts, but you must fully understand the solution that you turn in and write it up
entirely on your own. Blindly copying another student’s result will be considered a violation
of academic integrity. *

1. Problem 1: For any matrix A P R mˆn , prove that rank(A) = rank(A T A) = rank(AA T ).

2. Problem 2: This problem will teach you the idea of a left inverse, right inverse, and the
inverse.

(a) A matrix A P R mˆn is said to have a left inverse if there exists B P R nˆm such that

BA = Inˆn

Answer the following questions and provide justification for your answers.
i. If A has a left inverse, what should be the rank of A?
ii. Can a matrix A have more than one left inverse, i.e., can there exist B1 such that
BA = B1 A = Inˆn ? If yes, then specify the condition under which this is true.
(b) Similarly, if there exists a matrix C P R nˆm such that AC = Imˆm then it is called a
right inverse of A. Answer the following questions and provide justification for your
answers.

i. If A has a right inverse, what should be the rank of A?

ii. Can a matrix A have more than one right inverse, i.e., can there exist C1 such
that AC = AC1 = Imˆm ? If yes, then specify the condition under which this is
true.
(c) A matrix is said to be invertible if it has both left and right inverse. Answer the
following questions and provide justification for your answers.
i. What should be the relation between m and n if A is invertible?
ii. If a matrix has both left and right inverses, what should be the rank(A)?
iii. In this case, will the left and right inverse be the same matrix? Justify.
* For more information on Academic Integrity Policies at UCSD, please visit http://academicintegrity.ucsd.
edu/excel-integrity/define-cheating/index.html
i rankA nullity i.dimlNATADidmwwi caas.me
integers

RAHNCAk iiniii.in
NAI isallx thatAx 0
ATAXIAT0 o forall x
rank A tdmWA n
rank 三点
N ATAl is all xthat AAxo ǎh吢是⽠
ǐǜǜij
i rankATA
let t.ranhn rank.AM
ATAko as weknow
TATAx 0 RankA
rauklA
A
ADTAxl i.rank.AT AM
A rank.AT
N
品刚 _rankAAT
⼀
ATA ENCA
i.NL A NCAA i V9咏 A rank AT A rank AN

__

me

is rank A n D there are morethan one left

⽐ he BA Inxnf.BE Rnxm inverse for matrix A
if pm A ER MM
for A ERMM we can regard as
RIA Em an A Fn FM
andfor Bt 中以⼼ there are morethan oneleft
if inverse
RB Eman
than for B Bi TSF
i RIBAEmin RA RB了
we can get BA BA

RCI En
EMM
Inx butIc.mn In FmsFn
B BDA BA BA
RIBA nx i.R.CI
if Aalsohaveone riqhnwec
n Em
we have Act In
i RanklA En Soas13121
B BFIB.BDIn B B.AE
in rank A n 近
3 BDAJC oni.in⼝
⼈when Bt Bi s no right here c
mtni.fr multiply砦terse
 ODD rank A i yes there are morethan one night
issame as n inverse for matrix A
he have AC⼆Imxm for
a Rnxm A ER MM
if mm we can regard as
for At RMM Ai Fu FM
RIA n M there are morethan one left
if inverse
andfor Ct 中NM than for C.CI TUF
RIC E na m get CA CA
Imx i.RU
we can
A Emin RA RB ImxwiFmsFM
E M
but I Epmxm K D A A CAT mxn
RCI En if Aalsohaveone left mere B
t.R.CI RICA wehaveBA

i RanklA Em
I i.nz CCFIC
SoasC121
CDInBQ C.tt
213 CDAJ
him em onynci i.mn multiple
When Cta s left mera B
right
no

mfi.fi werse.
m n

i if it has a right inverse C we have men

if Ahas a left inverse B we have nam
we get men if A is invertible
ii by Drank A in an
rank Ak mn
bybismnhy.ms n

is.i mini.A is a square matrix so that have only

one nurse is left me see
right here
73 __

X.Yis leftinverse of Z
⼆ Xz 冮run 任 江mn
wehave ⼩⼗⽇ 1
北⼗ 2 ⼩凡⼗BYE
BY
⼩ImntpImn
⼆ ⼩ Imn
⼆ T.mn
⼩⼼BYis alsoleftinverseofZ forany Scalars X FB that⼩切 1

___
a let A⼆ ⻔ 志 的but none
not leftinvertible
ㄨ 了 灴⼦
沿 107问讯了but臨 1 ⻔汇⾦品
AX 订 AY汇订了成 浙 we let ㄨ
引 ㄚ汇订as
but TY At o AE 凹 A 2 Y
if A's left_invertible A
bH Xt AY
wehave left inverseBthatBAIn Y
wehasAKAY
BAEBAY
瑟邓
i if ABleftinvertible Hhtfrom
AX AY we get
EY
n __

taint
i.fordD PHkditdzttdsttt
PHFatcz.tt Gt2
DE0698 tdn.int 8
Cztxstt M 1 Cntn 2 动⼼ 3 0
t.cztzcstt.r.tlu DCntn 2
a i.ddtdztt dn.int 666 灬
dz IS
diana 䰘活 点
峾阎
ixnm. o
5
__

a wiiǜiüà
01000000

㦠
j

1 o o oo
鬬 i
o10 aijlljiz o.ch others

this is A
14 ㄨ 81mati bfmilXM
X YXz.tixxt_YX4fiix.JP5
X1 ⼈ X了 X4 ㄨ 5

i
o 1 000
0t t o o i

o o i i o
o o o l o
o o o ⼠⽓

this is A 9 5 matrix A zn

i 愳等
DX aij.fi it
PG 0 all others

zines
 92 ㄨ 3 93亿
f ME Aㄨ a xX

let A ⼆
⼀刹
识器影
蕊
灬
沿圝 㓹 㵞 纵 a ⼼

肛 品尝您 f is a linearfunction
of X ME

圝 癗 䦴 器毕
品 1
劐
了
贯
㘚 飇
吠䎙
ooooooo

㗻
共 oooo 㔼䨊
㶕 ti 癌
ooooooo

器器器
A
3. Problem 3:

(a) Textbook problem 11.1.

(b) Textbook problem 11.3.

4. Problem 4: Textbook problem 6.13.

5. Problem 5: Textbook Problem 7.4.

6. Problem 6:

(a) Textbook problem 8.3.

(b) Textbook problem 8.4.