Max number iterations for tolerance

Question

Jami Turnquist 2020-3-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/510184-max-number-iterations-for-tolerance

回答： Matt J 2020-3-11

x0 = [3,4];
function [F,gradient] = MyFunc(x)
F = 10*x(1).^2 + x(2)^2;
gradient = [20*x(1), 2*x(2)];
options = optimset('TolX', 0.01);
x = fminsearch(@MyFunc, x0, options)
end

I'm tryng to find the max number of iterations needed to find a minimum for the above function, using a tolerance of 0.01. For some reaspon this code will not yield results but it will run, just not show anything. Can anyone figure out the error?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Steven Lord 2020-3-10

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/510184-max-number-iterations-for-tolerance#answer_419529

Don't call fminsearch with MyFunc as your objective function from inside MyFunc itself. Move the lines where you call optimset and fminsearch outside MyFunc.

Assuming that MyFunc was nested inside its parent function (so it can "see" the definition of x0) and that parent function called MyFunc with an input x with at least two elements, this would:

call MyFunc which would call fminsearch which would
call MyFunc which would call fminsearch which would
call MyFunc which would call fminsearch ...

Eventually MATLAB would reach the recursion limit and throw an error or it would crash. If you've set the recursion limit too high, it potentially could crash the machine.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Jami Turnquist 2020-3-10

Are the x values I obtain the number of iterations it takes?

请先登录，再进行评论。

Answer 2

Matt J 2020-3-11

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/510184-max-number-iterations-for-tolerance#answer_419549

在 MATLAB Online 中打开

Call fminsearch with more output arguments to get information about how many iterations it took.

x0 = [3,4];
options = optimset('TolX', 1e-6);
[x,fval,exitflag, stats] = fminsearch(@MyFunc, x0, options)
function [F,gradient] = MyFunc(x)
F = 10*x(1).^2 + x(2)^2;
gradient = [20*x(1), 2*x(2)];
end

x =
   1.0e-06 *
   -0.2376    0.1417
fval =
   5.8448e-13
exitflag =
     1
stats = 
  struct with fields:
    iterations: 64
     funcCount: 124
     algorithm: 'Nelder-Mead simplex direct search'
       message: 'Optimization terminated:↵ the current x satisfies the termination criteria using OPTIONS.TolX of 1.000000e-06 ↵ and F(X) satisfies the convergence criteria using OPTIONS.TolFun of 1.000000e-04 ↵'