Blank figures appearing for plots

I've create an adaptive time-stepping function and I'm trying to plot the graphs but it is just creating blank figures.
Here is my function:
function [tvec,yvec,dtvec]=adaptive_euler(f,ytrue,y0,t0,dt0,dtmin,dtmax,tmax,alpha,beta,gamma,TOL)
y=zeros(10^5,1);
y(1)=y0;
dt=zeros(10^5,1);
t=zeros(10^5,1);
dt(1)=dt0;
yt=y0;
yt1=y0;
t(1)=t0;
n=1;
while (t(n)<tmax && dt(n)<dtmax && dt(n)>dtmin)
yt=yt+dt(n)*f(t(n),yt);
yt1=yt1+(dt(n)/2)*f(t(n),yt1);
yt2=yt1+(dt(n)/2)*f(t(n),yt1);
Et2=(yt2-yt)/(dt(n));
c=TOL/norm(Et2);
if norm(Et2) < TOL
y(n)=yt2;
t(n+1)=t(n)+dt(n);
n=n+1;
end
dt(n) = min(max(alpha*c*dt(n-1),gamma*dt(n-1)),beta*dt(n-1));
end
tvec=zeros(n,1);
yvec=zeros(n,1);
dtvec=zeros(n,1);
for i=1:n
tvec(i)=t(i);
yvec(i)=y(i);
dtvec(i)=dt(i);
end
yreal=zeros(n,1);
error=zeros(n,1);
for i=1:n
yreal(i)=ytrue(tvec(i));
error(i)=abs(yvec(i)-yreal(i));
end
T1=table(n,error(n));
figure(1)
plot(tvec,yvec,'.',tvec,yreal,'--');
legend('Approx:','True:');
title('Graph of the Adaptive Eulers Method approximation of dy/dx=-exp(2*t)*y^2 against real values');
figure(2)
plot(tvec,dtvec)
title('Graph of the change in dt over t');
end
And here is the script:
f=@(t,y)-exp(2*t)*y^2;
t0=0;
y0=2;
dt0=0001;
dtmin=10^-10;
dtmax=0.5;
tmax=10;
alpha=0.84;
beta=4;
gamma=0.1;
TOL=0.1;
ytrue=@(t)2*exp(-2*t);
[tvec,yvec,dtvec]=adaptive_euler(f,ytrue,y0,t0,dt0,dtmin,dtmax,tmax,alpha,beta,gamma,TOL)
Thank you!

4 Comments

Simplest option - stick a breakpoint on the plotting line and look at what is in tvec and yvec or dtvec or whichever plot is causing the problem. It may be obvious then what the problem is.
As an aside, don't use a for loop like this:
tvec=zeros(n,1);
yvec=zeros(n,1);
dtvec=zeros(n,1);
for i=1:n
tvec(i)=t(i);
yvec(i)=y(i);
dtvec(i)=dt(i);
end
It is a much slower and more convoluted version of simply
tvec = t;
yvec = y;
dtvec = dt;
And then once you've simplified it to that one would have to ask why do it at all? Why not just use t, y and dt instead of the new vectors that are an exact copy?
Because I set up the original vectors as very large vectors of zeros as I wouldn't know how many entries would be filled, but when I plot, I don't want the zeros, i just want the entries that are filled.
But you set them up as vectors of length n and then you copy from 1 to n from another vector into these.
Because in the original vectors, they are much longer and everything after n is a zero, n is the number of iterations my while loop runs for, while the length of the original vectors is much larger

Sign in to comment.

Answers (1)

You never enter the while loop
while (t(n)<tmax && dt(n)<dtmax && dt(n)>dtmin)
because dt(1) is 1 and dtmax is 0.5. Pass in a different dtmax.

2 Comments

Ah, it was supposed to be 0.001 instead of 1. now there seems to be an issue with
dt(n) = min(max(alpha*c*dt(n-1),gamma*dt(n-1)),beta*dt(n-1));
Saying "Array indices must be positive integers or logical values."
1 is also not less than 0.001, so that won't enter the loop either. So what did you pass in?

Sign in to comment.

Categories

Find more on Graphics Objects in Help Center and File Exchange

Asked:

on 10 Mar 2020

Commented:

on 10 Mar 2020

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!