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

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

CHAPTER 5. RUNGE-KUTTA METHODS 113<br />

=⇒<br />

Proof. When applied to the test equation y ′ = λy the general Runge-Kutta method<br />

has<br />

ν∑<br />

K i = λy n−1 + z a ji K i , = 1, 2, . . . , ν (5.212)<br />

i=1<br />

y n = y n−1 + h ∑ b i K i (5.213)<br />

where z = hλ. Let us define the following vector in R ν :<br />

⎛ ⎞<br />

K 1<br />

⎜ ⎟<br />

ξ = ⎝ . ⎠ (5.214)<br />

K ν<br />

<strong>The</strong>n we can rewrite equation 5.212 as<br />

ξ = λ1y n−1 + zAξ ⇐= (5.215)<br />

Rearranging and solving for ξ, ⇐=<br />

ξ − zAξ = λ1y n−1 (5.216)<br />

(I − zA)ξ = λ1y n−1 (5.217)<br />

ξ = λ(I − zA) −1 1y n−1 (5.218)<br />

hence ⇐=<br />

ν∑<br />

y n = y n−1 + h b j K j (5.219)<br />

which is equivalent to the desired result.<br />

j=1<br />

<strong>The</strong>orem 5.3. <strong>The</strong> stability function satisfies<br />

= y n−1 + hb T ξ (5.220)<br />

= ( 1 + zb T (I − zA) −1 1 ) y n−1 (5.221)<br />

R(z) = det(I − zA + z1bT )<br />

det(I − zA)<br />

(5.222)<br />

Proof. For the test problem y ′ = λy, ⇐=<br />

(I − zA)ξ = λ1y n−1 (5.223)<br />

−hb T ξ + y n = y n−1 (5.224)<br />

hence<br />

( ) ( ) ( )<br />

I − zA 0 ξ<br />

λ1<br />

−hb T = y<br />

1 y n−1<br />

n 1<br />

(5.225)<br />

<strong>The</strong> result follows from Cramer’s Rule (solve for y n ).<br />

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

Last revised: May 23, 2007<br />

Math 582B, Spring 2007<br />

California State University Northridge

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

Saved successfully!

Ooh no, something went wrong!