Your fitness function is
function y= simple_fitcat1(x)
y =(300*x(1)-(pi*x(2)*x(3)^2*0.25))*1000*0.05 + ((7.85*10^-6*x(2)*pi*x(3)^2*1000)*10);
and your lower bounds ensure that all variables are positive. So it is obvious that minimizing simple_fitcat1 entails making x(2) and x(3) as large as possible.
ga is just solving the problem that you gave it. Try using fmincon, you'll get the same answer, but more quickly and reliably.
Alan Weiss
MATLAB mathematical toolbox documentation
