Matlab Question: variables that depend on previous variables

[planar]

Hi,

I was wondering how you can get this to work in MatLab:

input is n

for i from 1 to n
x0 = 1
xi = x(i-1) + 2

So that for i = 1, x1 = x0 +2, or x1 = 1 + 2

Is there anyway to express this in MatLab?

Thanks!

 

[shoehorn]

A quick and dirty way to do it is as follows.

n = 10;
x = [1 zeros(1, n)];

for i = 2:n+1
   x(i) = x(i-1) + 2;
end

This is, however, bad practice since it takes n loops to achieve what you want; it's not the sort of thing you'd use in serious code since you should generally avoid for() loops in Matlab wherever possible.

A much better solution would probably involve some use of the filter command. Check the docs for more info.

 

[Mark44]

This is, however, bad practice since it takes n loops to achieve what you want; it's not the sort of thing you'd use in serious code since you should generally avoid for() loops in Matlab wherever possible.

A much better solution would probably involve some use of the filter command. Check the docs for more info.

I don't buy this at all. As fast as CPUs are these days, even interpreted languages (which I believe Matlab to be) are able to execute thousands of iterations in very little time. If there's any bottleneck, it's likely Matlab is waiting on I/O.

 

[matonski]

I don't buy this at all. As fast as CPUs are these days, even interpreted languages (which I believe Matlab to be) are able to execute thousands of iterations in very little time. If there's any bottleneck, it's likely Matlab is waiting on I/O.

I use Matlab everyday at my job and avoiding loops can make a huge difference. I also find it quite fun to try and vectorize various algorithms. Granted, for scripts that only take a second to run, its all a wash.

 

[matonski]

n = 10;
x = 1:2:(2*n+1);

 

[shoehorn]

I don't buy this at all. As fast as CPUs are these days, even interpreted languages (which I believe Matlab to be) are able to execute thousands of iterations in very little time. If there's any bottleneck, it's likely Matlab is waiting on I/O.

Anyone who has used Matlab for any serious work knows that the cardinal rule is to avoid loops whenever possible. This is not debatable in any sense.

Indeed, as matonski said, avoiding loops and vectorising your code can see huge gains in performance, particularly on multiple cores.

n = 10;
x = 1:2:(2*n+1);

Indeed. That's much better than a loop.