Issue 7 in ecspy: More elaborated observer for analisys

[ec...@googlecode.com]

Status: New
Owner: ----
Labels: Type-Defect Priority-Medium

New issue 7 by benhurna...@gmail.com: More elaborated observer for analisys
http://code.google.com/p/ecspy/issues/detail?id=7

Create a image-saver observer with all the caracteristics of the problem
in a graphic.
In the image saved by this observer, contains the current population,
mutation rate, cross-over rate, number of elites, name of the problem, best
individual, average individual, x axis=population and y axis=fitness,
number of images to be saved,et cetera.
And input values for these info, in the constructor of the observer.

Ex.: ga.observer = observers.image_observer('problem name',3)
#3 would be the number of images.

[ec...@googlecode.com]

Updates:
Owner: aaron.lee.garrett
Labels: -Type-Defect Type-Enhancement

Comment #1 on issue 7 by aaron.lee.garrett: More elaborated observer for
analisys
http://code.google.com/p/ecspy/issues/detail?id=7

On what schedule would the images be saved? For instance, if you say you
want 3 images, should it be the first 3? The last 3? The observer will be
called at the end of each generation. Unless you're terminating based on
number of generations or number of function evaluations, it will not be
possible to do something like "first, middle, last" for your 3 images
(because "middle" would be impossible to know until everything was
finished).

If you can give me some more information about your particular situation as
to why you'd need this kind of observer, we might be able to find a way to
get what you need.

Aaron

[ec...@googlecode.com]

Comment #2 on issue 7 by benhurna...@gmail.com: More elaborated observer
for analisys
http://code.google.com/p/ecspy/issues/detail?id=7

Thinking here, no need for the "number of images", but 1 image with all
request details and a file name with a problem's name+timestamp.png.
What do you think?

[ec...@googlecode.com]

Comment #3 on issue 7 by aaron.lee.garrett: More elaborated observer for
analisys
http://code.google.com/p/ecspy/issues/detail?id=7

Have you looked at the generation_plot function of the analysis module?
Here's the documentation:
http://mcis.jsu.edu/faculty/agarrett/ecspy/reference.html#module-ecspy.analysis
And here's an example that uses it:
http://mcis.jsu.edu/faculty/agarrett/ecspy/examples.html#genetic-algorithm

It generates a plot very similar to Pyevolve's summary plot. You have to
use a file_observer for your EC, and you have to pass the statistics file
generated to the analysis function. From that file, it will create a plot.

Is that similar to what you want?

[ec...@googlecode.com]

Comment #4 on issue 7 by benhurna...@gmail.com: More elaborated observer
for analisys
http://code.google.com/p/ecspy/issues/detail?id=7

In fact yes. Many thanks.
Att.

[ec...@googlecode.com]

Comment #5 on issue 7 by jellefer...@gmail.com: More elaborated observer
for analisys
http://code.google.com/p/ecspy/issues/detail?id=7

Hello Ben,

I think your suggestion is perhaps a little project specific. But surely
Aaron's pointers help you out?

[ec...@googlecode.com]

Comment #6 on issue 7 by benhurna...@gmail.com: More elaborated observer
for analisys
http://code.google.com/p/ecspy/issues/detail?id=7

certainly helped me out. Thank you all.

[ec...@googlecode.com]

Updates:
Status: Done

Comment #7 on issue 7 by aaron.lee.garrett: More elaborated observer for
analisys
http://code.google.com/p/ecspy/issues/detail?id=7

(No comment was entered for this change.)