Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
72 views

MAT 3310 (2009-2010, Second Term) Suggested Solution For Assignment 2

This document provides the suggested solution to an assignment in a mathematics course. It includes 5 problems solved with calculus concepts like Lagrange multipliers, projections, and variations. The key steps are finding the stationary points that minimize a function subject to constraints for problem 1, finding the point on a plane closest to a given point using vectors for problem 2, computing projections onto lines and showing identities involving projections for problem 3, relating potential energy, strain energy and work for problem 4, and solving a differential equation to minimize a functional for problem 5.

Uploaded by

julianli0220
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
72 views

MAT 3310 (2009-2010, Second Term) Suggested Solution For Assignment 2

This document provides the suggested solution to an assignment in a mathematics course. It includes 5 problems solved with calculus concepts like Lagrange multipliers, projections, and variations. The key steps are finding the stationary points that minimize a function subject to constraints for problem 1, finding the point on a plane closest to a given point using vectors for problem 2, computing projections onto lines and showing identities involving projections for problem 3, relating potential energy, strain energy and work for problem 4, and solving a differential equation to minimize a functional for problem 5.

Uploaded by

julianli0220
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

MAT 3310 (2009-2010, Second Term)

Suggested Solution for Assignment 2

1. We formulate the problem


Min (P ) = (4y1 + 4y2 )
s.t.
y12 + 4y22 = 1,

(1)

y1 0,

(2)

y2 0.

(3)

Define the Lagrange multiplier,


L = (4y1 + 4y2 ) + (y12 + 4y22 1)
The stationary point should satisfy
L

y1 = 4 + 2y1 = 0
L
y = 4 + 2y1 = 0

L1
2
2
= y1 + 4y2 1 = 0
Solve the system, y1 =

2 5
5 ,

y2 =

5
10 ,

5.

The maximum perimeter is P = 4y1 + 4y2 = 4

2 5
5

(4)

+4

5
10

= 2 5.

2. For point A (y1 , y2 , y3 ) on the plane, if the distance from the point P (1, 1, 1) to A is

minimal, then the vector AP = (1 y1 , 1 y2 , 1 y3 ) must be multiple of the normal

vector ~r = (1, 2, 3), so we have


(
(1 y1 ) : (1 y2 ) : (1 y3 ) = 1 : 2 : 3
(5)
y1 + 2y2 + 3y3 = 18
Let 1 y1 = t, 1 y2 = 2t, 1 y3 = 3t, then y1 = 1 t, y2 = 1 2t, y3 = 1 3t, and
25
substitute it into y1 + 2y2 + 3y3 = 18, we get t = 76 , y1 = 13
, y2 = 19
7 , y3 = 7 ,
q
7
~ k = ( 6 )2 + ( 12 )2 + ( 18 )2 = 6 14 .
the minimal distance is kAP
7

Please email to the tutors if there are mistakes: xmdeng@math.cuhk.edu.hk.

3. (a) Let A = (1, 2, 3)T , Ax and ~y are projections of vector ~a = (1, 1, 1)T on the line
y1 = y22 = y33 and its vertical line respectively, then
~a = Ax + ~y

(6)

(Ax) ~y = 0

(7)

Since x is a nonzero scalar, AT ~y = 0 by equation (7), and we can reformulate it into


the following matrix form:


I A
~y
~a
=
AT 0
x
0
solve the system, we get ~y = ( 47 , 17 , 27 )T , x = 73 .
the minimum distance from point (1, 1, 1) to the line is k~y k =

21
7 .

(b) The projection matrix P onto the line is :

1 2 3
1
2 4 6
P = A(AT A)1 AT =
14
3 6 9
(c) By the definition of P , P T = A(AT A)1 AT = P ;
and P 2 = P P = A(AT A)1 AT A(AT A)1 AT = A(AT A)1 (AT A)(AT A)1 AT =
A(AT A)1 AT = P .
For any vector ~b,
kP bk2 = (P b)T (P b) = bT P T P b = bT P P b = bT P b
2

k(I P )bk = kb P bk = (b P b) (b P b)
= bT b 2bT P b + (P b)T (P b)
2

(8)
(9)
(10)

= kbk 2b P b + kP bk

(11)

= kbk2 2kP bk2 + kP bk2

(12)

= kbk kP bk

(13)

The identity kP bk2 + k(I P )bk2 = kbk2 is thus proved.


4. (a) At the equilibrium state where K~u = f~, Pmin is achieved.
1
Pmin = (K~u, ~u) (f~, ~u)
2
1
= (K~u, ~u) (K~u, ~u)
2
1
= (K~u, ~u)
2
The strain energy U ,

1
U = (K~u, ~u) = Pmin
2
2

(14)
(15)
(16)

By w = CA~u, K = AT CA,
1
Qmin = (C 1 w, w),
2
1
= (A~u, CA~u)
2
1
= (AT CA~u, ~u)
2
1
= (K~u, ~u) = U
2

(17)
(18)
(19)
(20)

(b) At equilibrium,

1
1
U = (K~u, ~u) = (f~, ~u)
2
2
U is NOT equal to the external potential energy (f~, ~u)!
If they are equal, then U = (f~, ~u) = 0, which means U = 21 (K~u, ~u) = 0 by (a).
Since K is positive symmetric matrix, ~u = ~0 and f~ = K~u = ~0, which means the bar
is not stretched, and there are no external force, it is impossible when the f~ is the
gravitational forces on the masses.

5. (a)
"

#
d(u + tv) 2
dx
dx
0
2 #
2
Z "
1 1 d2 u
du dv
dv
d2 v
du
dx
=
+ 2t

+ t2
+t 2 +
2
2 0 dx
dx
dx
dx dx
dx

Z
Z
t2 1 dv 2
t 1 d2 v
du dv
= P (u) +
dx +
+
2

dx
2 0
dx
2 0 dx2
dx dx

1
P (u + tv) =
2

d2 (u + tv)
+
dx2

(21)
(22)
(23)

For fixed u and v, we can treat P (u + tv) as a quadratic function on t. The first
variation P
u is achieved at
P (u + tv) P (u)
1
P (u) v = lim
=
t0
t
2

(b) Functions satisfying

1 2
d v

du dv
+2

dx
dx2
dx dx

u00 (x) = 3
u(0) = 1

are u(x) = 1.5x2 + c1 x + 1. c1 is to be determined.


We know u(x) that can minimize P (u) satisfies:

Z 1 2
du dv
d v
0
+2

dx = 0
2P (u) v =
dx2
dx dx
0
Z 1
(v 00 + 2u0 v 0 )dx = 0
0

(24)

(25)

(26)
(27)

Integration by part,
Z
Z

(v 00 2uv 00 )dx + 2uv 0 |10 = 0

(28)

(1 2u)v 00 dx + 2u(1)v 0 (1) 2u(0)v 0 (0) = 0

(29)

0
1

for any candidate (u + tv), from (u + tv)00 (x) = 3, and u00 (x) = 3, we know v 00 (x) = 0,
then v 0 (x) C, C is some constant. The above equation now become:
0 + 2C(u(1) u(0)) = 0

(30)

u(1) = u(0) = 1

(31)

Substitute x = 1 into u(1) = 1.5 + c1 + 1 = 1, c1 = 1.5,


and u(x) = 1.5x2 1.5x + 1.

You might also like