Loop problem

% I have got this simple program runnning a loop to satisfy the condition 'nvvni = 0' stop the program

PT= 50;
c = [0.20,0.10,0.10;0.20,0.20,0.20;0 0 0];
P= [190 72.2 51.6; 20.44 15.57 4.956;0 0 0];
K = P/PT;
A = c(1,1)*(K(1,1)-1)+ c(1,2)*(K(1,2)-1)+c(1,3)*(K(1,3)-1)+c(2,1)*(K(2,1)-1)+c(2,2)*(K(2,2)-1)+c(2,3)*(K(2,3)-1);
B = (c(1,1)*(K(1,1)-1)/K(1,1))+(c(1,2)*(K(1,2)-1)/K(1,2))+(c(1,3)*(K(1,3)-1)/K(1,3))+(c(2,1)*(K(2,1)-1)/K(2,1))+(c(2,2)*(K(2,2)-1)/K(2,2))+(c(2,3)*(K(2,3)-1)/K(2,3));
%disp(B)
for i=1:10
n=0;
 nvni = (A )/(A-B);
n=n+1;
nvvni = (c(1,1)*(K(1,1)))/(nvni*(K(1,1)-1)+1)+(c(1,2)*(K(1,2)))/(nvni*(K(1,2)-1)+1)+(c(1,3)*(K(1,3)))/(nvni*(K(1,3)-1)+1)+(c(2,1)*(K(2,1)))/(nvni*(K(2,1)-1)+1)+(c(2,2)*(K(2,2)))/(nvni*(K(2,2)-1)+1)+(c(2,3)*(K(2,3)))/(nvni*(K(2,3)-1)+1);

      if nvvni == 0;
          disp(nvni);
          break
      else
          nvvvi = (c(1,1)*(K(1,1)^2)/(nvni*(K(1,1) + 1)^2))+(c(1,2)*(K(1,2)^2)/(nvni*(K(1,2) + 1)^2))+(c(1,3)*(K(1,3)^2)/(nvni*(K(1,3) + 1)^2))+(c(2,1)*(K(2,1)^2)/(nvni*(K(2,1) + 1)^2))+(c(2,2)*(K(2,2)^2)/(nvni*(K(2,2) + 1)^2))+(c(2,3)*(K(2,3)^2)/(nvni*(K(2,3) + 1)^2));
          nvvni = nvni - (nvvni/nvvvi);
          nvni =nvvni;
      end
  end

[Jan]

Please format your code properly. I've done this for you this time.

What is your question?

Answer 1

[Jan]

It is very unlikely that the result of such a complicated formula is exactly 0.0 due to rounding errors. Better check for a certain limit:

if abs(nvvni) < 1e-14

The size of the limit depends on the physical meaning of the formula and cannot be defined uniquely. An analysis of the sensitivity is recommended in addition: How sensitive is nvvni to tiny variations of the inputs, e.g. when using c+eps.

[Walter Roberson]

http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F