Scribd
Upload
22 views2 pages

Course Offering FA18

Exercise

Uploaded by

Hilbert
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
22 views2 pages

Course Offering FA18

Exercise

Uploaded by

Hilbert
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

BM20A6200 Inverse problems and normed spaces, 2018

Exercise 1

Deadline for submitting solutions via Moodle is 11.9.2018 at 21:00.


Format for submission:
• To produce pdf-file from Matlab-solutions, see exercise_example.m and exerci-
se_example.pdf that is produced from the first by command publish(’exercise_example’,’pdf’).
• Please include all functions that you define by yourself into on main m-file, named
exercise1.m. Please, use similar format that is used in exercise_example.m.
• Possible "pen and paper", solutions you can for example scan and then combine
with Matlab-pdf file. Or if you are familiar with Latex-you might be able to write
your solutions inside Matlab file (publish command can do some Latex).
• Submit one all inclusive pdf-file and one all inclusive Matlab m-file via Moodle.
1. Let consider space X = P2 ([0, 1]), i.e. space of polynomials up to second order that
are defined on interval [0, 1]. Study if mapping k·k : X → [0, ∞[ satisfy conditions
for being a norm.
(a) kpk := |a0 | + |a1 | + |a2 |
(b) kpk := |a0 | + |a1 |2 + |a2 |
R1 2
(c) kpk := 0 x |p(x)|dx

In all cases p(x) = a0 + a1 x + a2 x2 . Assume in your studies that zero element of


space X is constant function z(x) := 0.
2. (a) Let V = { f : [0, 1] → R | f is first, second or third order polynomial}. Check
that V is not a vector space.
(b) Let W = { f : [0, 1] → R | f is first, second or third order polynomial, f (0) =
0}. Check that W a vector space.
(c) Define some spanning set for W .
(d) Can the functions φ1 (x) = x + x2 , φ2 (x) = x2 + 3x3 and φ3 (x) = 2x − 6x3 span
the space W ?
3. Do this task with Matlab.
(a) Take samples from functions φ0 (t) = 1, φ1 (t) = t and φ2 (t) = t 2 at points
t = 0, 0.1, 0.2, . . . , 1 and form this way vectors v0 , v1 and v2 in R11 . You can
consider vi vectors as discrete approximations of φi (t) functions.
(b) Let assume now that V = span{v0 , v1 , v2 } and y ∈ R11 is discrete approxima-
tion of function f (t) = cos(πt) (in similar fashion as vi are for φi . Find coef-
ficients a0 , a1 and a2 such that ky − ∑2i=0 ai vi k22 is minimized. Notice that this
is just traditional LSQ-optimization task. Plot y and ∑2i=0 ai vi into same figu-
re. Plot f (t) and ∑2i=0 ai φi (t) into another figure (use 100 t-values in this later
plot).
q
4. Let define norm kvk := v21 + 7v22 + v23 for vectors v = [v1 , v2 , v3 ]T ∈ R3 . It is not
hard to give analytical proof that this mapping is a norm. First two conditions for
being a norm are quite trivial, and checking triangle inequality would not be hard
either. But now you should check this "numerically":
Generate randomly 5000 vectors ui ∈ R3 and 5000 vectors wi ∈ R3 and check check
how many times triangle inequality
kui − wi k ≤ kui k + kwi k is not true. Make here randomization so that each element
of each vectors ui and wi will be taken as sample from normal distribution N(0, 1).

You might also like