syl

Announcements , Course Information , Outline , Homework , Matlab Tutorial , Internet Resources

Announcements:

11/16:Homework 2 is HERE.

10/26:第二題h1,h2的取法應該是

1.先取五百個不同的H(可有規律)
2.在(0,H)間亂數取出h1, h2.
這種取樣法得到的結果比較有意義, 程式範例如下:

for i=1:500
H=10/i
h1=H*rand(1)
h2=H*rand(1)
h=max(h1,h2)
d2u_h=....
d2u_exac=....
err=abs(d2u_h-d2u_exac)
.....
end

1/16: I upload some sample codes for linear and nonlinear scalar transport equation. You can download them and try to see the simulation

10/20: To call cgs (Conjugate Gradients Squared Method),use "u=cgs(A,b',1.e-9,N-1)", where A is the (N-1) x (N-1) matrix and 1.e-9 is the tolerance.

10/5: To measure the CPU time, use "cputime".

10/4: Homework 1 is HERE.

10/4: You can use "polyfit" to fit the data and find the slope (order).

For example,

T=log10(n)
% n is the number
of the intervals
[c,d]=polyfit(T,log10(err_lvl),1)
% err_lvl is the
array of errors
set(axes,'fontsize',18)
hold on
plot(T,log10(err_lvl),'s');
err_fit=c(1)*T+c(2);
plot(T,err_fit);
xlabel('-log_{10} N');
ylabel('log_{10} err');
hold off

top

Elementary Numerical Partial Differential Equations:
Finite Difference Methods

Time and Place: Tue. 10:10-12:00; Wed. 17:10-18:00, Math Building
Lecturer: Min-Hung Chen
Office: Math Building Rm. 404; Ph. 2757575 Ext 65127
e-mail: mhchen@math.ncku.edu.tw
Web: www.math.ncku.edu.tw/~mhchen
Prerequisite: Calculus, Linear Algebra.
Text: Finite Difference Methods for Differential Equations by Randall J. LeVeque
References:

Time Dependent Problems and Difference Methods by Bertil Gustafsson, Heinz-Otto Kreiss and Joseph Oliger
Finite Difference Schemes and Partial Differential Equations by John Strikwerda

Grades will be based upon: Homework and in-class participation
top