Download presentation
Presentation is loading. Please wait.
1
ENGG2013 Unit 17 Diagonalization Eigenvector and eigenvalue Mar, 2011.
2
EXAMPLE 1 kshumENGG20132
3
Q6 in midterm u(t): unemployment rate in the t-th month. e(t)= 1-u(t) The unemployment rate in the next month is given by a matrix multiplication Equilibrium: Solve kshumENGG20133 Unemployment rate at equilibrium = 0.2
4
Equilibrium kshumENGG20134 Unstable Stable
5
If stable, how fast does it converge to the equilibrium point? kshumENGG20135 0.2 Fast convergenceSlow convergence
6
Question Suppose that the initial unemployment rate at the first month is x(1), (for example x(1)=0.25), and suppose that the unemployment evolves by matrix multiplication Find an analytic expression for x(t), for all t. kshumENGG20136
7
EXAMPLE 2 kshumENGG20137
8
How to count? Count the number of binary strings of length n with no consecutive ones. kshumENGG20138
9
SOLVING RECURRENCE RELATION kshumENGG20139
10
Fibonacci numbers F 1 = 1 F 2 = 1 For n > 2, F n = F n-1 +F n-2. The Fibonacci numbers are – 1,1,3,5,8,13,21,34,55,89,144 kshumENGG201310 http://en.wikipedia.org/wiki/Fibonacci_number
11
A matrix formulation Define a vector Initial vector Find the recurrence relation in matrix form kshumENGG201311
12
A general question Given initial condition and for t 2 Find v(t) for all t. kshumENGG201312
13
Matrix power Need to raise a matrix to a very high power kshumENGG201313
14
A trivial special case Diagonal matrix The solution is easy to find Raising a diagonal matrix to the power t is easy. kshumENGG201314
15
Decoupled equations When the equation is diagonal, we have two separate equation, each in one variable kshumENGG201315
16
DIAGONALIZATION kshumENGG201316
17
Problem reduction A square matrix M is called diagonalizable if we can find an invertible matrix, say P, such that the product P –1 M P is a diagonal matrix. A diagonalizable matrix can be raised to a high power easily. – Suppose that P –1 M P = D, D diagonal. – M = P D P –1. – M n = (P D P –1 ) (P D P –1 ) (P D P –1 ) … (P D P –1 ) = P D n P –1. kshumENGG201317
18
Example of diagonalizable matrix Let A is diagonalizable because we can find a matrix such that kshumENGG201318
19
Now we know how fast it converges to 0.2 The matrix can be diagonalized kshumENGG201319
20
Convergence to equilibrium The trajectory of the unemployment rate – the initial point is set to 0.1 kshumENGG201320
21
EIGENVECTOR AND EIGENVALUE kshumENGG201321
22
How to diagonalize? How to determine whether a matrix M is diagonalizable? How to find a matrix P which diagonalizes a matrix M? kshumENGG201322
23
From diagonalization to eigenvector By definition a matrix M is diagonalizable if P –1 M P = D for some invertible matrix P, and diagonal matrix D. or equivalently, kshumENGG201323
24
The columns of P are special Suppose that kshumENGG201324
25
Definition Given a square matrix A, a non-zero vector v is called an eigenvector of A, if we an find a real number (which may be zero), such that This number is called an eigenvalue of A, corresponding to the eigenvector v. kshumENGG201325 Matrix-vector productScalar product of a vector
26
Important notes If v is an eigenvector of A with eigenvalue, then any non-zero scalar multiple of v also satisfies the definition of eigenvector. kshumENGG201326 k 0
27
Geometric meaning A linear transformation L(x,y) given by: L(x,y) = (x+2y, 3x-4y) If the input is x=1, y=2 for example, the output is x = 5, y = -5. kshum27 x x + 2y y 3x – 4y
28
Invariant direction An Eigenvector points at a direction which is invariant under the linear transformation induced by the matrix. The eigenvalue is interpreted as the magnification factor. L(x,y) = (x+2y, 3x-4y) If input is (2,1), output is magnified by a factor of 2, i.e., the eigenvalue is 2. kshum28
29
Another invariant direction L(x,y) = (x+2y, 3x-4y) If input is (-1/3,1), output is (5/3,-5). The length is increased by a factor of 5, and the direction is reversed. The corresponding eigenvalue is -5. kshum29
30
Eigenvalue and eigenvector of First eigenvalue = 2, with eigenvector where k is any nonzero real number. Second eigenvalue = -5, with eigenvector where k is any nonzero real number. kshumENGG201330
31
Summary Motivation: want to solve recurrence relations. Formulation using matrix multiplication Need to raise a matrix to an arbitrary power Raising a matrix to some power can be easily done if the matrix is diagonalizable. Diagonalization can be done by eigenvalue and eigenvector. kshumENGG201331
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.