floating point precision

I am getting three kind of values for a variable (say vmax) after running my code. First value is 1. Second value is 1.0000 Third value is <1 I applied if else statements here as follows :

if vmax>=1

    figure(1);
    z=imread('C:\Program Files (x86)\MATLAB71\work\success.jpg');
    imshow(z);

else

    figure(2);
    z=imread('C:\Program Files (x86)\MATLAB71\work\fail.jpg');
    imshow(z);

end

As 1.0000 is slightly greater than 1 and I need same output for vmax=1 and vmax = 1.0000

But for few inputs which are giving vmax = 1.0000 , the if else works properly while for the other inputs which are giving vmax = 1.0000 , the if else executes "else" part. I am not getting why is this happening? For the same value of vmax, sometimes if else works the way it should while other times it works maliciously. Please help me.

Answer 1

[James Tursa]

E.g.,

>> X = 1
X =
     1
>> Y = X + eps(X)
Y =
    1.0000
>> Z = X - eps(X)
Z =
    1.0000
>> num2strexact(X)
ans =
1
>> num2strexact(Y)
ans =
1.0000000000000002220446049250313080847263336181640625
>> num2strexact(Z)
ans =
0.9999999999999997779553950749686919152736663818359375

X, Y, and Z all print as 1 or 1.0000, but they are in fact different. You can find num2strexact here:

href=""<http://www.mathworks.com/matlabcentral/fileexchange/22239-num2strexact-exact-version-of-num2str</a>>

Answer 2

[Sean de Wolski]

I'm not toally clear on your question. If you're looking to see if something is close to 1,

isclose21 = abs(x-1)<1^-10; %is absolute difference between x and 1 less than 1e-10?

Answer 3

[Walter Roberson]

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