why is there an error about index out?

I think my code is fine and check for several times. But there is still a problem about index. Can someone help me? I am a beginner. The code is as below:

clear all; close all;

ns=20;%step number

dx=-100;

dz=0.05;

Q=200;

b=50;

g=9.81;
n=0.025;

S=0.0005;

x_0=0;

y_0=4.62;

z_0=0;
A_0=y_0*b;

P_0=2*y_0+b;

R_0=A_0/P_0;

V_0=Q/A_0;

H1_0=y_0+z_0+V_0^2/2/g;

Sf_0=V_0^2*n^2/R_0^(4/3);

k=1:ns+1;

x=dx*(k-1);

z=dz*(k-1);

Sf=zeros(1,length(k));Sf(1)=Sf_0;

H1=zeros(1,length(k));H1(1)=H1_0;

yy=zeros(1,length(k));yy(1)=y_0;

for i=1:ns;

for j=1:100;

   y=y_0-j*0.01;

   A=y*b;

   P=2*y+b;

    R=A/P;

     V=Q/A;

    H1(i+1)=z(i+1)+y+V^2/2/g;

     Sf1=V^2*n^2/R^(4/3);

    Sf(i+1)=Sf1;

    Sf_avg=(Sf(i)+Sf(i+1))/2;

    avg=Sf_avg*dx;

    H2=H1(i)-avg;

while abs(H2-H1(i+1))<0.005

    yy(i+1)=y;

    i=i+1;

    end

   end

end

[Walter Roberson]

Which line does the error occur on?

[Dongyu]

on the line: while abs(H2-H1(i+1))<0.005

[Jan]

If you omit the clear all, you can set a breakpoint in the failing line or let Matlab stop there automatically:

dbstop if error

Then you can investigate the reasons of the problem much easier.

Killing the breakpoints by clear all is a bad idea, because the debugger is the friend of the programmer.

Answer 1

[Roger Stafford]

While you are increasing i in the while-loop there is the possibility of reaching an i such that i+1 is beyond the end of array y in the line "yy(i+1)=y;" or beyond the end of H1 in the condition "abs(H2-H1(i+1))<0.005". This would produce an error message about improper indexing.

You should be aware that changing i within the for-loop which indexes with i is not good programming practice. Mathworks' documentation states "Avoid assigning a value to the index variable within the body of a loop. The for statement overrides any changes made to the index within the loop."

[Dongyu]

Thank you! That should be the problem. Can you tell me way to revise it? What can I do if I don't change i in the loop?

[Roger Stafford]

I suggest you use another index, ii, in place of i which starts out at the current i value for indexing within the while-loop, and place the additional condition that ii cannot reenter the while-loop with a value greater than ns. That would guarantee that your particular indexing difficulty could not occur. I have used the "short circuit" && to avoid trouble with H1(ii+1) in that while condition.

I cannot promise that other difficulties will not become apparent when the indexing problem is corrected. I am uncertain as to what your code is meant to accomplish.

    .......
    H2=H1(i)-avg;
    ii = i; % Change i over to the index ii
    while ii<=ns && abs(H2-H1(ii+1))<0.005 % First check that ii<=ns
      yy(ii+1)=y;
      ii=ii+1;
    end
    .....

[Dongyu]

Thank you all the same! But when I do this, the loop doesn't work well any more. However, since the data I need is yy, I could simply save yy in a mat file in the loop. But the solution to this mistake still confuses me a lot.

Answer 2

[Dongyu]

I found out that if I put i<ns-1 as a condition, while i<ns-1 && abs(H2-H1(i+1))<0.005 it will be lack of the last term, but if I construct a new loop, only to calculate the last term, the program works well. It's kinda stupid though.