Finding minimum point from any function file input.

Asked by Jia Qing Soo

Jia Qing Soo (view profile)

on 9 Oct 2013
Latest activity Edited by Matt J

Matt J (view profile)

on 9 Oct 2013
Given ANY function file input, which we can assume to be continuous between [1,2], what is the code to estimate the x-value where the graph achieves its minimum point?

Tags

Answer by Matt Kindig

Matt Kindig (view profile)

on 9 Oct 2013
Edited by Matt Kindig

Matt Kindig (view profile)

on 9 Oct 2013

I would check out the documentation for fminsearch() and fminbnd(), or if you have the Optimization Toolbox, fminunc():
doc fminsearch
doc fminbnd
doc fminunc
For example, if your function is f(x)=sin(x), you can call it like this:
sol = fminbnd(@sin, 1, 2)

Jia Qing Soo

Jia Qing Soo (view profile)

on 9 Oct 2013
Is there another way of finding the minimum point?
For example, writing a code such as to input an x-value (from x = 1 to 2 with an interval of 0.1) to derive the corresponding f(x) and then getting matlab to figure out which f(x) is the minimum?
Matt Kindig

Matt Kindig (view profile)

on 9 Oct 2013
Are the f(x) functions pre-defined, or can they be literally anything? If they can be anything, this problem is impossible to solve, as there are an infinite number of f(x) functions possible.
What exactly is your goal here?

Answer by Matt J

on 9 Oct 2013
Edited by Matt J

Matt J (view profile)

on 9 Oct 2013

If f(x) is vectorized, you would just do
[minval,minloc] = min(f(1:0.1:2))
If you can't rely on it being vectorized, you would have to loop
x=1:0.1:2;
m=inf(size(t));
for i=1:length(m)
m(i)=f(x(i));
end
[minval,minloc]=min(m);