Prev: guide de matlab
Next: complicated situation :'(
From: Shi on 17 Jun 2010 18:16
I am solving a GA optimization problem, but I do not know how to write a code to stop the loop. Here is the discription of the problem:
There is a randomly generated population G of 100 chromosomes of 10 genes(binary code)
after optimization, all gene are 1, then optimization stop; my part code are
if mean(mean(G))==1
break
end
or if no optimum is found, but the result doesnt change in next 50 generations,, also stop and give the step number.
How to code this condition? how express the above two condition together?
Will you please help me for this? thanks a lot.
From: sscnekro on 17 Jun 2010 18:55
Hey Shi, have you considered a while loop? Fancy sth like...

% until the gene is all ones ..
while mean(mean(G)) <1
% do something
end

> or if no optimum is found, but the result doesnt change in next 50 generations,, also stop and give the step number.

For this you would need to create a sort of variable to "measure the change between iterations" and a sort of "counter" that would count iterations without change... You could then plug a condition in the while loop.

while counter < 50
% do something
end

> how express the above two condition together?

In general, you can condition the while loop upon one or more statements, using "and" "or" operators, but do not confuse those two in your specific context... If I am not wrong, here you 'd use

while mean(mean(G)) <1 or counter < 50
% do stuff
end

> give the step number

I'm not really sure what you mean by that, but maybe you want to know the number of "generations" of the GA evolution. For that create another counter. To show it, the command disp() could be helfpul.

By the way, when speaking about counters, I mean something like
mystep = mystep + 1

Hope have not made typos in this improvised advise. Best of luck!
From: Shi on 17 Jun 2010 21:37
"sscnekro " <stiahni.mail(a)zoznam.sk> wrote in message <hve948$cut$1(a)fred.mathworks.com>...
> Hey Shi, have you considered a while loop? Fancy sth like...
>
> % until the gene is all ones ..
> while mean(mean(G)) <1
> % do something
> end
>
> > or if no optimum is found, but the result doesnt change in next 50 generations,, also stop and give the step number.
>
> For this you would need to create a sort of variable to "measure the change between iterations" and a sort of "counter" that would count iterations without change... You could then plug a condition in the while loop.
>
> while counter < 50
> % do something
> end
>
> > how express the above two condition together?
>
> In general, you can condition the while loop upon one or more statements, using "and" "or" operators, but do not confuse those two in your specific context... If I am not wrong, here you 'd use
>
> while mean(mean(G)) <1 or counter < 50
> % do stuff
> end
>
> > give the step number
>
> I'm not really sure what you mean by that, but maybe you want to know the number of "generations" of the GA evolution. For that create another counter. To show it, the command disp() could be helfpul.
>
> By the way, when speaking about counters, I mean something like
> mystep = mystep + 1
>
> Hope have not made typos in this improvised advise. Best of luck!


the result are ten values(generations of optimization), how calculate the mean value of them? thanks
 | 
Pages: 1
Prev: guide de matlab
Next: complicated situation :'(