Why do I get this function warning (Function behaves unexpectedly on array inputs)

I want to plot 4 functions on the same graph. The plots works fine but I get this warning. How can I fix it please?

figure(1)
subplot(2,1,1)
fplot(@(x) camdisp(x),[0 360])
title('S plot')
ylabel('S in inches')
xlabel('\theta in degrees')
subplot(2,1,2)
fplot(@(x) camvel(x),[0 360])
title('V plot')
ylabel('V in inches/sec')
xlabel('\theta in degrees')

function s = camdisp(t)

w = 2*(pi/5);

if t>=0 && t<80
    s = ((10.*1.25)*((t./80).^3))+((-15.*1.25).*((t./80).^4))+((6.*1.25).*((t./80).^5));

elseif t>=80 && t<170
    s = 1.25;

elseif t>=170 && t<250
    s = 1.25+((-10.*1.25).*(((t-170)./80).^3))+((15.*1.25)*(((t-170)./80).^4))+((-6.*1.25).*(((t-170)./80).^5));

elseif t>=250 && t<=360
    s = 0;
end
end

Then I get this message

Warning: Function behaves unexpectedly on array inputs. To improve performance, properly vectorize your function
to return an output with the same size and shape as the input arguments. 
> In matlab.graphics.function.FunctionLine>getFunction
  In matlab.graphics.function.FunctionLine/updateFunction
  In matlab.graphics.function.FunctionLine/set.Function_I
  In matlab.graphics.function.FunctionLine/set.Function
  In matlab.graphics.function.FunctionLine
  In fplot>singleFplot (line 232)
  In fplot>@(f)singleFplot(cax,{f},limits,extraOpts,args) (line 191)
  In fplot>vectorizeFplot (line 191)
  In fplot (line 161)
  In Project (line 11)

[Adam]

That's unusually poor English for a warning in Matlab! I read it through 3 or 4 times and it still made no sense in English, nevermind interpreting it for code!

[Faris Alfaris]

Honestly I've spent few hours trying to figure out where I made a mistake but I couldn't find anything. I hope someone can explain. Thanks

[Stephen Munkachy]

Looks like you missed a period in front of a multiplication symbol in the first if statement. It's right after the first closed parenthesis ')'.

Answer 1

[Torsten]

function s = camdisp(t)

w = 2*(pi/5);

for i=1:numel(t)
if t(i)>=0 && t(i)<80
    s(i) = ((10.*1.25)*((t(i)./80).^3))+((-15.*1.25).*((t(i)./80).^4))+((6.*1.25).*((t(i)./80).^5));

elseif t(i)>=80 && t(i)<170
    s(i) = 1.25;

elseif t(i)>=170 && t(i)<250
    s(i) = 1.25+((-10.*1.25).*(((t(i)-170)./80).^3))+((15.*1.25)*(((t(i)-170)./80).^4))+((-6.*1.25).*(((t(i)-170)./80).^5));

elseif t(i)>=250 && t(i)<=360
    s(i) = 0;
end
end
end

Best wishes

Torsten.

[Faris Alfaris]

I tried to do the same thing to the rest of the functions but it messed everything up. I only know basic matlab so I'm sorry. Can you please explain or help with that. Thanks

Second function

% velocity

function v = camvel(t)

w = 2*(pi/5);
    if t>=0 && t<80

      v = (w./((2.*pi.*80)./360)).* (((10.*1.25*3).*((t./80).^2))+((-15.*1.25*4).*((t./80).^3))+((6.*1.25.*5).*((t./80).^4)));

elseif t>=80 && t<170
    v = 0;

elseif t>=170 && t<250
    v = (w./((2.*pi.*80)./360)).* (((-10.*1.25.*3).*(((t-170)./80).^2))+((15.*1.25.*4).*(((t-170)./80).^3))+((-6.*1.25.*5).*(((t-170)./80).^4)));

elseif t>=250 && t<=360
    v = 0;
    end

end

Third function

% accelecartion

function a = camacc(t)

w = 2*(pi/5);

if t>=0 && t<80
    a = ((w/((2*pi*80)/360))^2)* (((10*1.25*3*2)*((t/80)))+((-15*1.25*4*3)*((t/80)^2))+((6*1.25*5*4)*((t/80)^3)));

elseif t>=80 && t<170
    a = 0;

elseif t>=170 && t<250
    a = ((w/((2*pi*80)/360))^2)* (((-10*1.25*3*2)*(((t-170)/80)))+((15*1.25*4*3)*(((t-170)/80)^2))+((-6*1.25*5*4)*(((t-170)/80)^3)));

elseif t>=250 && t<=360
    a = 0;

end
end

Fourth function

% jerk

function j = camjerk(t)

w = 2*(pi/5);

if t>=0 && t<=80
    j = ((w/((2*pi*80)/360))^3)* (((10*1.25*3*2*1))+((-15*1.25*4*3*2)*((t/80)))+((6*1.25*5*4*3)*((t/80)^2)));

elseif t>=80 && t<=170
    j = 0;

elseif t>=170 && t<=250
    j = ((w/((2*pi*80)/360))^3)* (((-10*1.25*3*2*1))+((15*1.25*4*3*2)*(((t-170)/80)))+((-6*1.25*5*4*3)*(((t-170)/80)^2)));

elseif t>=250 && t<=360
    j = 0;
end
end

[Guillaume]

Note that w is never used in camdisp. As I explained in my answer, and Torsten implied without any explanation, you just need to loop over the input array using your original code. See the template I've given.

[Faris Alfaris]

Thanks a lot

Answer 2

[Guillaume]

fplot is trying to call your function with a vector input, e.g.

camdisp(0:360)

Your function errors in that case, hence fplot is forced to loop over each element of the vector and called your function for each element. That degrades performance.

You can rewrite your function so that it works for vector inputs. One such way would be to simply loop over the input vector

function s = camdisp(tarray)
   s = zeros(size(tarray));
   for i = 1:numel(tarray)
      t = tarray;
      if ...
        s(i) = ...  %your normal code goes here
      elseif ...
        s(i) = ...
      ...
      end
   end
end

That would give no performance benefit since you just moved the for loop out of fplot to put it in your function. At least, it would shut up fplot.

Another way would be of course to vectorise the function. A vectorised version could be:

function s = camdisp(t)
  s = zeros(size(t));
  group = discretize(t, [0 80 170 250 360]);
  s(group == 1) = ((10.*1.25)*((t(group == 1)./80).^3))+((-15.*1.25).*((t(group == 1)./80).^4))+((6.*1.25).*((t(group == 1)./80).^5));
  s(group == 2) = 1.25;
  s(group == 3) =  1.25+((-10.*1.25).*(((t(group == 3)-170)./80).^3))+((15.*1.25)*(((t(group == 3)-170)./80).^4))+((-6.*1.25).*(((t(group == 3)-170)./80).^5));
  s(group == 4) = 0;  %that line not even needed since s is already 0 in that case
end

[Guillaume]

Something I forgot to say:

fplot(@(x) camdisp(x), [0 360]);

can be written more simply as:

fplot(@camdisp, [0 360]);

In addition to the two different implementations, you could also change your fplot calls to:

fplot(@(x) arrayfun(@camdisp, x), [0 360]);

which would get rid of the warning and would not require you to change any of your function code. The downside is that it'll be even slower. arrayfun is another way to implement a loop.