21.04.2015 Views

The Computable Differential Equation Lecture ... - Bruce E. Shapiro

The Computable Differential Equation Lecture ... - Bruce E. Shapiro

The Computable Differential Equation Lecture ... - Bruce E. Shapiro

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

172 CHAPTER 9. DIFFERENTIAL ALGEBRAIC EQUATIONS<br />

we conclude that<br />

v = (I − ND + · · · + N k−1 D k−1 )h (9.60)<br />

= h − Nh ′ + N 2 h ′′ − · · · + N k−1 h ( k − 1) (9.61)<br />

In other words, equation 9.55 is not a differential equation at all, despite the presence<br />

of the derivative, but is an algebraic equation for v. This is sometimes referred<br />

to as a hidden constraint of the DAE.<br />

Two key points emerge from the preceding discussion. First, we see that DAE’s<br />

may have hidden constraints, namely, part of the solution may be fully determined<br />

independently of any initial conditions. Second, two many initial conditions can<br />

over-determine the system. If too many initial conditions are provided, that is,<br />

more than are necessary to determine u, then there may not be any solution at all.<br />

<strong>The</strong>orem 9.3. Let Ay ′ + By = f be a solvable linear constant coefficient DAE with<br />

index k ≥ 1. <strong>The</strong>n (A + λB) −1 has a pole of order k at λ = 0, and (A + λB) −1 A<br />

has pole of order k − 1 at λ = 0.<br />

Proof. We will give the proof for the first statement. <strong>The</strong> second is left as an exercise.<br />

Hence<br />

I = Q −1 Q (9.62)<br />

= Q −1 (A + λB) −1 (A + λB)Q (9.63)<br />

= Q −1 (A + λB) −1 P −1 P (A + λB)Q (9.64)<br />

= Q −1 (A + λB) −1 P −1 (P AQ + λP BQ) (9.65)<br />

Q −1 (A + λB) −1 P −1 = (P AQ + λP BQ) −1 (9.66)<br />

(A + λB) −1 = Q(P AQ + λP BQ) −1 P (9.67)<br />

( ) −1 I + λC 0<br />

= Q<br />

P (9.68)<br />

0 N + λI<br />

( )<br />

(I + λC)<br />

−1<br />

0<br />

= Q<br />

0 (N + λI) −1 P (9.69)<br />

Since (this can be verified by multiplying out the product, using the fact that N k =<br />

0)<br />

I = 1 [I − 1 λ λ N + 1 ]<br />

λ 2 N 2 + · · · + (−1)k−1<br />

λ k−1 N k−1 [N + λI] (9.70)<br />

we know that<br />

(N + λI) −1 = 1 λ<br />

∑k−1<br />

i=0<br />

−1 i<br />

λ i N i (9.71)<br />

Hence A + λB has elements that are proportional to 1/λ k . This gives a pole of oder<br />

k when λ → 0.<br />

Math 582B, Spring 2007<br />

California State University Northridge<br />

c○2007, B.E.<strong>Shapiro</strong><br />

Last revised: May 23, 2007

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!