why coding this function returns un exact numerical results?

4 views (last 30 days)
Cantor Set
Cantor Set on 22 Jun 2021
Commented: Walter Roberson on 23 Jun 2021
Ackley's Function (a=20,b=0.2) d=2.
function [ y ] = objfun14 (phen )
d=size(phen,2);
v=cos(2*pi*phen); x=(1/d)* sum(v,2);
t=-20*exp(-0.2*sqrt((1/d)*sum(phen.^2,2))) - exp( x ) +20+exp(1);
y=sum(t,2);
end
%when phen=[0 0] the output should be t=0 however excuting the code gives t=8.8818e-16
  1 Comment
Scott MacKenzie
Scott MacKenzie on 22 Jun 2021
t=8.8818e-16 is pretty close to zero. You are dealing with floating point arithmetic, that's why the result is not exactly what you expect. See Floating-Point Numbers for further details.

Sign in to comment.

Answers (1)

Kishan Dhakan
Kishan Dhakan on 23 Jun 2021
Try using 'format long', which shows 15 decimal places. If it still doesn't work, then your value is probably smaller than what many call 'machine epsilon' and hence is getting rounded to zero.
  1 Comment
Walter Roberson
Walter Roberson on 23 Jun 2021
The other way around, the function is returning -4*eps when the user expects 0

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by