Principles of Mathematics in Operations Research

(Rick Simeone) #1
14.4 Problems 201

Remark 14.4.6 In order to solve the linear difference system

y(k + l) = Ay(k) + f(k); y(0) = y 0 ,

we will take the Z transform of the components of y(k), then we have

(zI-A)rj(z) = zy 0 + (t>{z),

where n(z) = [771(2),-•• ,r]n(z)}T is the vector of Z transforms of the com-
ponents of y. If z is not an eigenvalue of A, then the coefficient matrix is
nonsingular. Thus, for sufficiently large \z\, the unique solution is

V(z) = z(zl - A)-
l
Vo + (zl - i4)" V(«),

where we have
Z{Ak}=z(zI-A)-^1.
In order to find a particular solution, we solve (zl — A)^1 p(z) = 4>(z) forp(z)
and find its inverse Z transform.

Example 14.4.7 Let us take our previous example problem:

yi(k + 1) = y 2 (k) + 1, 2/i(0) = 3,

y 2 (k + l) = yi(k) + l, 2/2(0) = 1.

V(z) =
z 1
(z-l)(z + l) [lz_

Z{Aky 0 }
z-\





2
2

1 zl
(*-l)(z + l) [if
z 1
-1

C^IFJ

n(z) =


z
(z-ir

z
+ (z-iy

=> y(k) = k^2  1
L 4 J


  • k


' 1'
4
3
.4.

"0"
1

. 2 _






z+1

z
(z-1)

+ (-l)fc









(*+l)
19
8
13

Problems


14.1. Solve y"(t) - y(t) = e2t; y(0) = 2, j,'(0) = 0.


14.2. Solve y(k + 1) = y(k) + 2ek; y(0) = 1.

Free download pdf