This preview shows page 1. Sign up to view the full content.
Unformatted text preview: y(x) 1 eRx (1 ; ) eRx(1 ; ) c + ;10
) e (1 +
1. (If = e
determine c as ; and the solution as 2R x>0 then the above relation is exact at x = 1.) We would then
c R2 ;1
y(x) 2eR2 ;1
Rx x > 0: With R large, there is the possibility of having catastrophic growth in the solution even
when is small. In order to demonstrate this, we solved this problem using the MATLAB
Runge-Kutta procedure ode45. While an explicit Runge-Kutta code is not the most
e cient solution procedure for this problem, e ciency was not our main concern. The
maximum errors in the solution component y1 with R = 1 10 are reported in Table 7.1.1.
The error has grown by six decades for a ten-fold increase in R. Indeed, the procedure
failed to nd a solution of the BVP with R = 100. R ke1k =ky1k
1 2:445 10 8
10 2:940 10 2
Table 7.1.1: Maximum errors in y1 for Example 7.1.1
1 1 ;
; Let's pursue this di culty with a bit more formality. The exact solution of the linear
4 BVP (7.1.1) can be written as (Problem 1...
View Full Document
This document was uploaded on 03/16/2014 for the course CSCI 6820 at Rensselaer Polytechnic Institute.
- Spring '14