lokal minimum and maximum

I want to find the lokal minimum, and mark it in the plot. I succed in the first go, but in the second, not so good. For somereason it doesnt agree with me. Here is the code i type - I hope anyone can help me.

function opgave31

disp('opgave a')

title('a') x=(-5:5);

f=funktion(x);

plot(x,f)

hold on

y_ny=zeros(1,length(x));

plot(x,y_ny,'r')

x1=fzero(@funktion,-4); x2=fzero(@funktion,1); x3=fzero(@funktion,2);

plot(x1,y_ny,'b*',x2,y_ny,'b*',x3,y_ny,'b*')

close all

disp('opgave b')

title('b')

s=funktion2(x);

plot(x,s)

min=fminsearch(@funktion2,1,5)

max=fminsearch(@(x) -funktion2(x),-5,5)

x5=fzero(@funktion2,min); x6=fzero(@funktion2,max);

close all

plot(x,s,'r-',min,y_ny,'b*')

function y=funktion(x)

y=x.^3+x.^2-10*x+8

function t=funktion2(x)

t=x.^3-6*x.^2-x+30;

Answer 1

[Azzi Abdelmalek]

You can use findpeaks(y) to find maximums, and findpeaks(-y) to find minimums

Answer 2

[Mischa Kim]

Hello Rasmus, how about doing it symbolically?

 syms x
 t        = x^3 - 6*x^2 - x + 30;
 rootsdt  = double(solve(diff(t)==0));
 dt2      = diff(t,2);

 if subs(dt2,rootsdt(2))>0
    x = rootsdt(2);
 else
    x = rootsdt(1);
 end

 ezplot(t)
 hold on
 plot(x, subs(t),'rs')

Answer 3

[Image Analyst]

If you don't have the Signal Processing Toolbox, which is where findpeaks() is located, then, if you have the Image Processing Toolbox, you can use imdilate() to find local maximum and imerode() to find local minimums. They're different than findpeaks. findpeaks() finds peaks, it does not find local maxima. So if you have a peak, like a Gaussian, it will find the very tip top apex of the hump. Of course there are local maximum all along the curve, even going down the sides of the hump, right? The imdilate() function is a local max in a window that slides along your signal so it will get the local max at every location, not just at a peak.