AMS210 Homework 10

5.3:10,12,20; 5.6:5,6,11,12; DUE: Wednesday Nov. 22, Beginning of Class

Solutions to Problems:


In the solutions below,"&" replaces "lambda", the Greek symbol conventionally
used to denote an eigenvalue.


5.3:10) The e-values for A = [ 2   3 ] are determined from
			     [ 4   1 ]
	 det(A-&I) = 0  
  ==>    (2-&)(1-&) - 12 = 0
  ==>	 &^2 - 3& - 10 = 0
  ==>    (&-5)(&+2) = 0
  ==>    & = 5, -2.

Since the e-values are distinct, we already know that A is diagonalizable.
For & = 5, to find an e-vector solve the system (A-5I)x = 0:

	[ -3   3   0 ]     [ 1  -1   0 ]  so an e-vector is v_1 =  [ 1 ]
	[  4  -4   0 ]  ~  [ 0   0   0 ] 		           [ 1 ]

For & = -2, solve 

	[ 4   3   0 ]	   [ 4   3   0 ]   so x_2 = -4/3 x_1;  pick x_1 = 3
	[ 4   3   0 ]	   [ 0   0   0 ]   to give e-vector  v_2 = [ 3 ]
								   [-4 ]
A is diagonalizable.  A = P D (P^-1) , where P is the 2x2 matrix having 
columns equal to the e-vectors:

  P = [ v_1  v_2 ] = [ 1   3 ]  and  D = [ 5   0 ]
      		     [ 1  -4 ]	         [ 0  -2 ]


5.3:12)  The e-values are given to be & = 2 (multiplicity 2) and & = 8.
To find a basis for eigenspace for & = 2, we find a basis for the solution 
set of (A-2I)x = 0:
	[  2   2   2   0 ]         [  2   2   2   0 ]	    [ 1   1   1   0 ]
	[  2   2   2   0 ]    ~    [  0   0   0   0 ]   ~   [ 0   0   0   0 ]
	[  2   2   2   0 ]         [  0   0   0   0 ]	    [ 0   0   0   0 ]
							  so x_1 = -x_2 - x_3.
There are 2 free variables so the dimension of the e-space is 2, matching the 
multiplicity of the e-value, so we can conclude now that A is diagonalizable 
(note the dimension of the eigenspace for & = 8 must be 1).   
A basis for the eigenspace for & = 2 is  { [-1 ] , [-1 ] }
					   [ 1 ]   [ 0 ]
					   [ 0 ]   [ 1 ]

Now solving (A-8I)x = 0 we find an e-vector spanning the 1-dimensional
eigenspace for & = 8:

	[ -4   2   2   0 ]         [  2  -1  -1   0 ]	    [ 2  -1  -1   0 ]
	[  2  -4   2   0 ]    ~    [  0  -3   3   0 ]   ~   [ 0   1  -1   0 ]
	[  2   2  -4   0 ]         [  0   3  -3   0 ]	    [ 0   0   0   0 ]

	[  2   0  -2   0 ]	   [  1   0  -1   0 ]	  		 [ 1 ]
   ~	[  0   1  -1   0 ]    ~    [  0   1  -1   0 ]   x_2 = x_3 = x_1  [ 1 ]
	[  0   0   0   0 ]	   [  0   0   0   0 ]  so an e-vector is [ 1 ]

Hence A = P D (P^-1), where

   P =  [ -1  -1   1 ]  and  D = [ 2   0   0 ]
	[  1   0   1 ]		 [ 0   2   0 ]
	[  0   1   1 ]		 [ 0   0   8 ]


5.3:20)  Notice A is lower triangular matrix so det(A-&I) is the product of the
diagonal entries of A-&I:

	[  4   0   0   0 ]              [ 4-&   0   0    0 ]
   A =	[  0   4   0   0 ]   A-&I  =    [  0   4-&  0    0 ]    
	[  0   0   2   0 ]      	[  0    0  2-&   0 ]  
	[  1   0   0   2 ]              [  1    0   0   2-&]

det(A-&I) = (4-&)(4-&)(2-&)(2-&) = 0  ==>  & = 4 with multiplicity 2
					   & = 2 with multiplicity 2
Find a basis for solutions of (A-2I)x = 0:

	[  2   0   0   0   0 ]            [  1   0   0    0   0 ]
	[  0   2   0   0   0 ]      ~     [  0   1   0    0   0 ]    
	[  0   0   0   0   0 ]            [  0   0   0    0   0 ] 
	[  1   0   0   0   0 ]            [  0   0   0    0   0 ]

Thus x_3, x_4 free,  and x_1 = 0 = x_2.
Thus a basis for the e-space for & = 2  is { [ 0 ] , [ 0 ] }
				             [ 0 ]   [ 0 ]
				             [ 1 ]   [ 0 ]
				             [ 0 ]   [ 1 ]
Similarly,  
(A-4I)x = 0

	[  0   0   0   0   0 ]            [  1   0   0   -2   0 ]
	[  0   0   0   0   0 ]      ~     [  0   0   1    0   0 ]    
	[  0   0  -2   0   0 ]            [  0   0   0    0   0 ] 
	[  1   0   0  -2   0 ]            [  0   0   0    0   0 ]


 implies        x_4, x_2 free
		x_3  =  0
		x_1  = 2 x_4.
Therefore, any vector in the eigenspace corresponding to & = 4 is of the form 
x_2 (0,1,0,0)  +  x_4 (2,0,0,1). 
 So a basis for the e-space for & = 4  is { [ 0 ] , [ 2 ] }
				            [ 1 ]   [ 0 ]
				            [ 0 ]   [ 0 ]
				            [ 0 ]   [ 1 ]
A is diagonalizable with A P = P D , where a choice is (note order dependent):

   P =  [  0   2   0   0 ]  and  D = [ 4   0   0   0 ]
	[  1   0   0   0 ]	     [ 0   4   0   0 ]
	[  0   0   1   0 ]	     [ 0   0   2   0 ]
	[  0   1   0   1 ]	     [ 0   0   0   2 ]



5.6:5)  det(A-&I) = det [  .4-&   .3  ]  =  (.4-&)(1.2-&) + .0975
			[ -.325  1.2-&]      =  &^2 - 1.6& + .48 + .0975
						= &^2 - 1.6& + .5775
is the characteristic polynomial.
	To find eigenvalues for A we set the characteristic polynomial to 0 and
solve for & (use quadratic formula):
		& = (1/2) [1.6 +- sqrt([1.6]^2 - 4[.5775])] 
		  = (1/2) [1.6 +- sqrt(.25)]
		  = (1/2) [1.6 +- .5].
Thus &_1 = 1.05  and &_2 = .55 are the eigenvalues.	
Solve (A - 1.05I)x = 0 to find a basis for the eigenspace for &_1 = 1.05:

	[ (.4 - 1.05)      .3     ]	[ -.65   .3   0 ]    [ -.65   .3   0 ]
	[    -.325   (1.2 - 1.05) ]  ~  [-.325  .15   0 ] ~  [   0     0   0 ]

Thus  .65x_1 = .3x_2  ==>  (13/6)x_1 = x_2,   so
an eigenvector is v_1  =  [ x_1 ]  =  [  6 ].
		          [ x_2 ]     [ 13 ]

The dynamical system x_k+1 = A x_k has solution
	x_k = A^k x_0
	    = c_1 (1.05)^k [  6 ] + c_2 (.55)^k v_2,   where [ c_1 ] is the 
			   [ 13 ]                            [ c_2 ]
coordinates of x_0 relative to the eigenvector basis { v_1, v_2 }.
If c_1 is not zero, then in the long term (for k large (.55)^k is negligible)

	x_k = c_1 (1.05)^k [  6 ]   approximately.
			   [ 13 ] 
So in long term both populations eventually grow at almost 5% rate (per month),
and the eventual ratio of owls to flying squirrels is 6 to 13 (thousand).


5.6:6)  det(A-&I) = det [  .4-&   .3  ]  =  (.4-&)(1.2-&) + .15
			[  -.5  1.2-& ]      =  &^2 - 1.6& + .48 + .15
						= &^2 - 1.6& + .63
						 = (& - .9)(& - .7)
Thus &_1 = .9  and &_2 = .7 are the eigenvalues.	
Since both eigenvalues are less than 1, the populations will eventually perish
since the dynamical system x_k+1 = A x_k has solution
	x_k = A x_0
	    = c_1 (.9)^k v_1 + c_2 (.7)^k v_2   converges to 0 as k ->infinity.

   To find a value of p in the matrix  A = [ .4   .3 ] 
					   [ -p  1.2 ]
such that the populations remain constant, we need to 
determine what p yields an e-value of &_1 = 1  (and with &_2 <= 1).

Method 1 (for finding p):
   If & = 1 is e-value, then there is a (non-zero) e-vector solution of
Ax = x  ==>  (A-I)x = 0  ==>  [ -.6  .3 ] has linearly DEPENDENT rows
	    has a free	      [ - p  .2 ]     ==> p =.4. 
	     variable

Method 2 (alternative):
   The characteristic polynomial is 
	det(A-&I) = &^2 - 1.6& + .48 + .3p
		  = &^2 - 1.6& + .3(1.6 + p)
   The roots are (using quadratic formula)
	& = (1/2) [ 1.6 +- sqrt{(2.56) - (1.2)(1.6+p)} ]
   Thus we want sqrt(discriminant) to equal .4 (then &_1 = 1, &_2 = .6);
that is we want discriminant to equal .16.
   Set .16 = 2.56 - (1.2)(1.6+p) and solve for p.
	(1.2)(1.6+p) = 2.4
  ==>   1.2p + 1.92 = 2.4  ==>  p = .48/1.2 = .4.

   An eigenvector for & = 1 is found by solving (A-I)x = 0 to get v_1 = [ 1 ]. 
									[ 2 ]
   For large k, x_k = c_1 (1)^k [ 1 ]  =  c_1 [ 1 ]  assuming c_1 non-zero.
				[ 2 ]	      [ 2 ]
Thus the eventual relative population sizes are 1 owl per 2(thousand)squirrels.


5.6:11)
     det(A-&I) = det [ .4-&   .5  ]  =  (.4-&)(1.3-&) + .2 = &^2 - 1.7& + .72
		     [ -.4   1.3-&]	
  		                                          = (&-.9)(&-.8) 

Thus the e-values are &_1 = .9 and &_2 = .8.
Since both of these are < 1, x_k --> 0 vector as k goes to infinity; the origin
is an attractor.
To find an e-vector for &_2 = .8 solve the system (A-.8I)x = 0 to get 
   v   = [ 5 ]   (or any non-zero multiple), which is direction of greatest
	 [ 4 ]			attraction since .8 smallest.



5.6:12)
     det(A-&I) = det [ .5-&   .6  ]  =  (.5-&)(1.4-&) + .18 = &^2 - 1.7& + .88
		     [ -.3   1.4-&]	
  		                                          = (&-1.1)(&-.8) 

Thus the e-values are &_1 = 1.1 and &_2 = .8.
Since  1.1 > 1 > .8, the origin is a saddle.
The direction of greatest attraction corresponds to any e-vector for 
&_2 = .8 (the e-value less than 1 in magnitude), such as   [ 2 ].
			                     		   [ 1 ]

The direction of greatest repulsion corresponds to any e-vector for 
&_1 = 1.1, such as (line through origin and)    [ 1 ].
			 			[ 1 ]