Undefined function 'minus' for input arguments of type 'function_handle'. and no output

function [t,u]=fixeddiffusion(h,dt,t0,theta,u0,uN,uinitial,x0,xN,uexact)
N=(xN-x0)/h;
x=zeros(N+1,1);
x(1)=x0;
x(N+1)=xN;
uvec=zeros(N+1,1);
uvec(1)=u0;
uvec(N+1)=uN;
A=zeros(N-1,N-1);
a=-2/(h^2);    
for i=1:N-1
        A(i,i)=a;
end    
b=(1/(h^2));    
for i=1:N-2
        j=i+1;
        A(i,j)=b;
end    
for i=2:N-1
        k=i-1;
        A(i,k)=b;
end
b=zeros(N-1,1);
for i=1:N-1
b(i)=uinitial(i+1);
end
uu=A\b;
for i=1:N-1
    uu(i)=uvec(i+1);
end 
u=zeros(N+1,1);
u(1)=u0;
for i=2:N
    u(i)=u(i-1)+dt*((1-theta)*uvec(i-1)+theta*uvec(i));
end
t=zeros(N+1);
for i=1:N+1
t(i)=t0+(i-1)*dt;
end
ureal=zeros(N+1,1);
for i=1:N+1
ureal(i)=uexact(x(i),t(i));
end
figure(1)
plot(t,u,'.',t,ureal,'--');
legend('Approx:','True:');
error=zeros(N+1,1);
error=abs(u-ureal);
norm=zeros(N+1,1);
norm(i)=symsum((h*dt(abs(error(i)^2)))^(1/2),error(i),1,i);
end

When I run the following:

uinitial=@(x)sin(pi*x);
u0=0;
uN=0;
h=0.1;
dt=0.0005;
theta=0;
x0=0;
xN=1;
uexact=@(x,t)exp((-pi^2)*t)*sin(pi*x);
[t,u]=fixeddiffusion(h,dt,theta,u0,uN,uinitial,x0,xN,uexact)

It comes up with the error message. It also says "Conversion to double from function_handle is not possible."

[Rik]

Apart from not copying the full error message (which would have contained the trace of offending lines, there are more issues.

• You used error as a variable name, wich will now shadow the builtin function, so now you can't trigger errors with a description.
• You did not auto-indent your code, which reduces readability.
• You did not use very descriptive variable names. Short names are quick to type, but provide no clues about their meaning.
• You did not write any comments explaining your code, or a function header explaining the goal of the function and syntax options.
• You did not do basic input checking (nargin and nargout are very helpful if you want a very fast check if all your inputs and outputs are as you expect).

Answer 1

[tmarske]

It would be helpful if you indicated where exactly the error occurred the Matlab error message should give you a line number. A long wall of text like this is not easy to parse.

However I believe the error here is because you missed the 't0' argument in your call to fixeddiffusion():

[t,u]=fixeddiffusion(h,dt,theta,u0,uN,uinitial,x0,xN,uexact)

cf the function definition:

function [t,u]=fixeddiffusion(h,dt,t0,theta,u0,uN,uinitial,x0,xN,uexact)

[Phoebe Tyson]

Here is the full error message after including t0:

Array indices must be positive integers or logical values.

Error in fixeddiffusion (line 50)

norm(i)=symsum((h*dt(abs(error(i)^2)))^(1/2),error(i),1,i);

Error in CW1q2 (line 12)

[t,u]=fixeddiffusion(h,dt,t0,theta,u0,uN,uinitial,x0,xN,uexact)

[Rik]

You still haven't fixed the issues I pointed out in my comment. Also, what do you think this error means? Where in this line of code are you indexing an array? What is the value of that index?