any good g.a.'s?
castironpi at gmail.com
castironpi at gmail.com
Fri Mar 28 11:00:15 CET 2008
On Mar 25, 11:03 am, Tim Chase <python.l... at tim.thechases.com> wrote:
> > Any good genetic algorithms involving you-split, i-pick?
> I've always heard it as "you divide, I decide"...
> That said, I'm not sure how that applies in a GA world. It's
> been a while since I've done any coding with GAs, but I don't
> recall any facets related to the You Divide, I Decide problem.
> It sounds like a simple optimization of "equality", which would
> be a normal use of a GA. This would apply for both the division
> and the decision, depending on which side the GA is (the "you" or
> the "I") or two diff. GAs can be on either side of the process.
Tim, if you have any interest in the stuff:
You Divide I Decide YDID only works where resources are highly
divisible: Except if, do livelock or deadlock occur in real time? If
I get in line for going to bars, I enter a decreased-expressivity (-
outlet) state, open up, close, and come sit down. Now, what do I
generally yield? A life's story. I think I 'send' a request for
details in the form of a repetition, except it takes into account a
lot of computation, spef. my understading. I also think I look for a
'random' generator to call 'send' on, but that could just be
prioritized by strength of pole, but here's the important part, -at- -
the- -time-. Taking priority graph/queue to be a function of the
past, you can classify into spatial pole and cumulative pole; just the
sense mechanata 'say where': 'we favor X' progressively over time.
Back at the fork in the road, I was favoring a turn, not a path. I
can make a dual, and for the sake of discussion, entertain I do. Are
you a musician? Is Heaven a noninterest market? Electricity is a
uniform current, highly potent (can sterilize+neutralize), highly
uniform (verb regulize). I don't like "discovering poles" "once on
the road (just to where)"; I think they're a net quality of life
decrease; I don't like poles; they make one answer, heavy questions.
Computers at least, can't come and answer; hypothetically,
analytically, their pulls can't either. (Computers and their pulls,
doh.) However, if I create pole by my actions, they're only part to
blame (yes, instead of entirely either or not).
Except but, how do I get home? I don't get random homes, it's the
same one. Neither does my home pull random people (is it a pole to).
It's weird: you can bury latent poles, though clearly time isn't the
fuse-length. Contingently exclusive (upon?). There's a layer of
discouragement for everything and around everyone else's; surprising
me at home is repulsive, lightly. Eating is attractive, more and less
over time, sometimes crucial (oddly enough); it creates a junction by
outweighing discouragement. (My kitchen is extremely unattractive but
not repulsive at all, any farther than the newcome stir, which will
come with me after that, though the process of dissipation is a little
mysterious-- LaPlace?) There's no good framework for composing (draw
analogy to, analogee) the process of emancipation (w.c.) and
dissipation, so I lean to church, even though it's chalk-full of non-
mathematical (non-vicarous) remarks: "We come, and the vicar leaves.
He's just of no interest." Interest might be mutual too, but only in
magnitude. I don't like attraction wars. Can I get a tongue
regressed (cf. kitchen sink regression)? Are we in Economics?
There's money in it. I have to go speak French. Je dois aller.
If you receive a sign, do you just design it? What if you have a more
clever way to spell a word, qua composite sign? What if you spell it
the old way? Consign, design, ensign, and resign. The English tongue
is really 'pulling forward' quite fast.
Time is generational in facet. (Is a phoneme a particle? Phones are
I think time means 'generation', which you can do across space. Ears
receive 1-D parametric impressions, Eyes receive 2-D. I don't know of
much work in deparametrization, but method iteration ( f_i=
f( f_(i-1) ) ), representing space as time, creates generations of
something. Problem is, iterations are chosen by somebody, what
function to call next in time, so there is a lot of money in choosing
functions to call in order, es. if it's refining parameters to other
functions. Feedback on computers is lightning-fast, bringing billion-
baud guess-and-checking to hands.
If you were a tuple, what tuple would you be? Encode, hash, and
echo? Do you GOTO home or GOSUB home? What if you limited call stack
depth to like five?
More information about the Python-list