-
Notifications
You must be signed in to change notification settings - Fork 118
/
pset3.lyx
413 lines (333 loc) · 7.45 KB
/
pset3.lyx
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
#LyX 2.3 created this file. For more info see http://www.lyx.org/
\lyxformat 544
\begin_document
\begin_header
\save_transient_properties true
\origin unavailable
\textclass article
\begin_preamble
\renewcommand{\vec}[1]{\mathbf{#1}}
\renewcommand{\labelenumi}{(\alph{enumi})}
\renewcommand{\labelenumii}{(\roman{enumii})}
\end_preamble
\use_default_options false
\maintain_unincluded_children false
\language english
\language_package default
\inputencoding auto
\fontencoding global
\font_roman "times" "default"
\font_sans "default" "default"
\font_typewriter "default" "default"
\font_math "auto" "auto"
\font_default_family default
\use_non_tex_fonts false
\font_sc false
\font_osf false
\font_sf_scale 100 100
\font_tt_scale 100 100
\use_microtype false
\use_dash_ligatures true
\graphics default
\default_output_format default
\output_sync 0
\bibtex_command default
\index_command default
\paperfontsize default
\spacing single
\use_hyperref false
\papersize default
\use_geometry true
\use_package amsmath 1
\use_package amssymb 1
\use_package cancel 1
\use_package esint 0
\use_package mathdots 0
\use_package mathtools 1
\use_package mhchem 1
\use_package stackrel 1
\use_package stmaryrd 1
\use_package undertilde 1
\cite_engine basic
\cite_engine_type default
\biblio_style plain
\use_bibtopic false
\use_indices false
\paperorientation portrait
\suppress_date false
\justification true
\use_refstyle 0
\use_minted 0
\index Index
\shortcut idx
\color #008000
\end_index
\topmargin 1in
\secnumdepth 3
\tocdepth 3
\paragraph_separation indent
\paragraph_indentation default
\is_math_indent 0
\math_numbering_side default
\quotes_style english
\dynamic_quotes 0
\papercolumns 1
\papersides 2
\paperpagestyle default
\tracking_changes false
\output_changes false
\html_math_output 0
\html_css_as_file 0
\html_be_strict false
\end_header
\begin_body
\begin_layout Section*
18.335 Problem Set 3
\end_layout
\begin_layout Standard
Due Friday, 15 March 2019.
\end_layout
\begin_layout Subsection*
Problem 1: QR and orthogonal bases
\end_layout
\begin_layout Enumerate
Trefethen, problem 10.4.
\end_layout
\begin_layout Enumerate
Prove that
\begin_inset Formula $A=QR$
\end_inset
and
\begin_inset Formula $B=RQ$
\end_inset
have the same eigenvalues, assuming
\begin_inset Formula $A$
\end_inset
is a square matrix.
Then do a little experiment: Construct a random
\begin_inset Formula $5\times5$
\end_inset
real-symmetric matrix in Julia via
\family typewriter
X=rand(5,5); A = X' + X
\family default
.
Use
\family typewriter
QR = qr(A)
\family default
(do
\family typewriter
using LinearAlgebra
\family default
first) to compute the QR factorization of
\begin_inset Formula $A$
\end_inset
, and then compute
\family typewriter
B = QR.R * QR.Q
\family default
.
Then find the QR factorization
\begin_inset Formula $B=Q'R'$
\end_inset
, and compute
\begin_inset Formula $R'Q'$
\end_inset
...repeat this process until the matrix converges (maybe writing a loop and/or
a function).
From what it converges to, suggest a procedure to compute the eigenvalues
and eigenvectors of a real-symmetric matrix (no need to prove that it converges
in general—we will discuss this in class).
\end_layout
\begin_layout Enumerate
Trefethen, problem 28.2,
\end_layout
\begin_layout Subsection*
Problem 2: Schur fine
\end_layout
\begin_layout Standard
In class, we will show that any square
\begin_inset Formula $m\times m$
\end_inset
matrix
\begin_inset Formula $A$
\end_inset
can be factorized as
\begin_inset Formula $A=QTQ^{*}$
\end_inset
(the
\emph on
Schur factorization
\emph default
), where
\begin_inset Formula $Q$
\end_inset
is unitary and
\begin_inset Formula $T$
\end_inset
is an upper-triangular matrix (with the same eigenvalues as
\begin_inset Formula $A$
\end_inset
, since the two matrices are similar).
\end_layout
\begin_layout Enumerate
\begin_inset Formula $A$
\end_inset
is called ``normal'' if
\begin_inset Formula $AA^{*}=A^{*}A$
\end_inset
.
Show that this implies
\begin_inset Formula $TT^{*}=T^{*}T$
\end_inset
.
From this, show that
\begin_inset Formula $T$
\end_inset
must be diagonal.
Hence, any normal matrix (e.g.
unitary or Hermitian matrices) must be unitarily diagonalizable.
Hint: consider the diagonal entries of
\begin_inset Formula $TT^{*}$
\end_inset
and
\begin_inset Formula $T^{*}T$
\end_inset
, starting from the (1,1) entries and proceeding diagonally downwards by
induction.
\end_layout
\begin_layout Enumerate
Given the Schur factorization of an arbitary
\begin_inset Formula $A$
\end_inset
(not necessarily normal), describe an algorithm to find the eigenvalues
and eigenvectors of
\begin_inset Formula $A$
\end_inset
, assuming for simplicity that all the eigenvalues are distinct.
The flop count should be asymptotically
\begin_inset Formula $Km^{3}+O(m^{2})$
\end_inset
; give the constant
\begin_inset Formula $K$
\end_inset
.
\end_layout
\begin_layout Subsection*
Problem 3: Caches and backsubstitution
\end_layout
\begin_layout Standard
In this problem, you will consider the impact of caches (again in the ideal-cach
e model from class) on the problem of
\emph on
backsubstitution
\emph default
: solving
\begin_inset Formula $Rx=b$
\end_inset
for
\begin_inset Formula $x$
\end_inset
, where
\begin_inset Formula $R$
\end_inset
is an
\begin_inset Formula $m\times m$
\end_inset
upper-triangular matrix (such as might be obtained by Gaussian elimination).
The simple algorithm you probably learned in previous linear-algebra classes
(and reviewed in the book, lecture 17) is (processing the rows from bottom
to top):
\end_layout
\begin_layout LyX-Code
\begin_inset Formula $x_{m}=b_{m}/r_{mm}$
\end_inset
\end_layout
\begin_layout LyX-Code
\series bold
for
\series default
\begin_inset Formula $j=m-1$
\end_inset
down to
\begin_inset Formula $1$
\end_inset
\end_layout
\begin_layout LyX-Code
\begin_inset Formula $x_{j}=(b_{j}-\sum_{k=j+1}^{m}r_{jk}x_{k})/r_{jj}$
\end_inset
\end_layout
\begin_layout Standard
Suppose that
\begin_inset Formula $X$
\end_inset
and
\begin_inset Formula $B$
\end_inset
are
\begin_inset Formula $m\times n$
\end_inset
matrices, and we want to solve
\begin_inset Formula $RX=B$
\end_inset
for
\begin_inset Formula $X$
\end_inset
—this is equivalent to solving
\begin_inset Formula $Rx=b$
\end_inset
for
\begin_inset Formula $n$
\end_inset
different right-hand sides
\begin_inset Formula $b$
\end_inset
(the
\begin_inset Formula $n$
\end_inset
columns of
\begin_inset Formula $B$
\end_inset
).
One way to solve the
\begin_inset Formula $RX=B$
\end_inset
for
\begin_inset Formula $X$
\end_inset
is to apply the standard backsubstitution algorithm, above, to each of
the
\begin_inset Formula $n$
\end_inset
columns in sequence.
\end_layout
\begin_layout Enumerate
Give the asymptotic cache complexity
\begin_inset Formula $Q(m,n;Z)$
\end_inset
(in asymptotic
\begin_inset Formula $\Theta$
\end_inset
notation, ignoring constant factors) of this algorithm for solving
\begin_inset Formula $RX=B$
\end_inset
.
\end_layout
\begin_layout Enumerate
Suppose
\begin_inset Formula $m=n$
\end_inset
.
Propose an algorithm for solving
\begin_inset Formula $RX=B$
\end_inset
that achieves a better asymptotic cache complexity (by cache-aware/blocking
or cache-oblivious algorithms, your choice).
Can you gain the factor of
\begin_inset Formula $1/\sqrt{Z}$
\end_inset
savings that we showed is possible for square-matrix multiplication?
\end_layout
\end_body
\end_document