The genetic algorithm and its possible applications - page 9

New comment
 
Igor Makanu:

Adjusted OnTester() to give the worst result if not many trades are made by EA, your advice seems to work

Thank you!

No, it does not quite work this way - agents have become idle

But it works fine if you amplify the best result. I am amplifying results by number of deals:

#define  EA_MIN_TRADEES  250
double OnTester()
{
.....
   return(AccountInfoDouble(ACCOUNT_BALANCE) * ((2.0 * (double)TesterStatistics(STAT_TRADES)) / EA_MIN_TRADEES));
}
 
Реter Konow:

I don't know. Tell me. I'm new at this.

https://www.google.com.ua/search?client=opera&q=генетическое+programming+presentation&sourceid=opera&ie=UTF-8&oe=UTF-8

 
Реter Konow:

I don't know. Tell me. I'm new at this.

http://is.ifmo.ru/genalg/

Кафедра «Технологии программирования»: Генетические алгоритмы
  • is.ifmo.ru
Главная / Генетические алгоритмы (версия для печати)
 
Andrey Dik:

http://is.ifmo.ru/genalg/

Thank you. I've read it. Interested in the CAD system.

The rest of the material seemed (as I said before), research and experiments of scientists. The "smart ant" task is found everywhere.

The concept of genetic programming is simple: we take a template of some system consisting of parameters (chromosome) and optimize its function using "clever" search of values. This is clear to everyone. But further on, nothing is clear. What is the future of this approach? Does it exist at all...

Clever search veiled by terminology in a 'biological' wrapper with a 'flavour' of Evolution misleads researchers and they harbour hopes of becoming God. And in nature, things are MUCH more complicated than that.

Suppose we repurpose GA into an algorithm for assembling indexed parameters into random systems.

1. What is the target function of the random system? I.e. - what is it for?

2. If a random system is launched to an artificial environment (created by us), will it find its target function there? (without which the system cannot be optimized).

3. The artificial environment of fabricated "raw" systems with an undefined function will look like a meaningless rubbish dump generated by a GA modified to assemble an incomprehensible thing.


All in all, there is a serious lack of meaning in this... BUT - this is just my opinion. I could be wrong and later admit my mistake.

 

Perhaps the search for the meaning of random parameter assembly should begin at the outset.

Firstly, what do we mean by "system" - chromosome? These are completely different concepts. A chromosome is a tiny blueprint of an individual, whereas an individual is a living being containing trillions of chromosomes.

A chromosome contains the concept of an individual in compressed form, while an individual is an ecosystem-adapted critter occupying a place in the food chain. Individuals were only partly created 'by chance'. Onlya small fraction appeared randomly, and as a whole, the individual is an exceptionally well-thought-out system in every respect.

If we are going to 'sculpt' chromosomes randomly - then we are not thinking about the individual and the ecosystem as a whole. This is not how God created us.)) On the face of it we have a very elaborate habitat with living beings sharpened for it, and if we start looking for randomness in the universe we will find it, but as a percentage of the universe it will be negligible.

So, our 'divine' engineering has to start with building concepts, and then, repurpose the GA and run the 'conveyor of life' with all the target functions in place.

 
Реter Konow:

Thank you. I've read it. I was interested in the CAD system.

The rest of the material seemed (as I said before) to be nothing more than research and experiments of scientists. The "smart ant" task is found everywhere.

The concept of genetic programming is simple: we take a template of some system consisting of a set of parameters (chromosome) and optimize its function using "clever" search of values. This is clear to everyone. But further, nothing is clear. What is the future of this approach? Does it have any future at all...

Intelligent search of values of parameters of the system, veiled by terminology in a kind of "biological" wrapper with a "flavour" of Evolution, misleads researchers and they have empty hopes of becoming God.) In nature, after all, everything is MUCH more complicated.

Suppose we repurpose GA into an algorithm for assembling indexed parameters into random systems.

1. What is the target function of a random system? I.e. - what is it for?

2. If a random system is launched to an artificial environment (created by us), will it find its target function there? (without which the system cannot be optimized).

3. The artificial environment of fabricated "raw" systems with an undefined function will look like a meaningless rubbish dump generated by a GA modified to assemble an incomprehensible thing.


All in all, there's a serious lack of meaning in all this so far... BUT - this is just my opinion. I could be wrong later and admit my mistake.

Konow reg:

I guess the search for the meaning of random parameter assembly should start from the beginning.

Firstly, what do we mean by "system" - chromosome? These are completely different concepts. A chromosome is a tiny blueprint of an individual, whereas an individual is a living being containing trillions of chromosomes.

A chromosome contains the concept of an individual in a compressed form, while an individual is an ecosystem-adapted critter with a place in the food chain. Individuals were only partially created 'by chance'. Only a small fraction of an individual appeared randomly. Overall, the individual is an exceptionally well-thought-out system in every respect.

If we are going to 'sculpt' chromosomes randomly - then we are not thinking about the final individual - its place in the ecosystem and the ecosystem as a whole. This is not how God created us.)) There is a very elaborate habitat on the face of it with the creatures sharpened for it. If we start looking for random in the universe, we will find it, but as a percentage of the universe it will be negligible.

So, our "divine" engineering needs to start exactly with building the concepts of individuals and the ecosystem, and then, repurpose the GA and run the "conveyor of life" with all the target functions in place.

Peter, sorry, but you are a walking hoard of delusions.

Remember the amino acids I mentioned earlier.

You understand one thing - any optimization, absolutely any optimization, is based on random search. No matter how you try to apply meaningful optimization - it only seems to you that the search is meaningful, logical. in fact, the search is random. And no matter what you call AO, Genetic, Ant, Particle Swarm, Magical, Divine, it will be random.

God, the Universe, or the Supreme Mind did the following: poured in, roughly, electrons, protons. mions or whatever, elementary particles, shook it up carefully, and out of all this came the universe as you see it, with planets, stars, galaxies, clusters. and sometimes - with living things. These elemental particles make up everything, from a small, living ant, to a huge, black, silent, dead and merciless black hole.

Think about it, everything is made up of elementary particles! From those elementary particles you can mould anything, from a stool to a quantum computer, from an ant to a human being! It's not the genes and chromosomes that matter, it's the elementary particles.

Genes folded into chromosomes are only a certain level of abstraction. If you go higher, you receive an individual (a living or a non-living organism). If you go lower, you receive dna, amino acids, oxygen, hydrogen, carbon, and nitrogen. And then you get elemental particles. Move backwards, and you get lithium from hydrogen, and so on, all the way to uranium and beyond! Choose your direction and abstractions and create!

 
Andrey Dik:

Peter, I'm sorry, but you are a walking warehouse of delusion.

Remember the amino acids I mentioned earlier.

You understand one thing - any optimization, absolutely any optimization, is based on random search. No matter how you try to apply meaningful optimization - it only seems to you that the search is meaningful, logical. in fact, the search is random. And no matter what you call AO, Genetic, Ant, Particle Swarm, Magical, Divine, it will be random.

God, the Universe, or a Higher Intelligence did the following: poured in, roughly, electrons, protons. mions or whatever, elementary particles, shook it up thoroughly, and out of all this came the universe as you see it, with planets, stars, galaxies, clusters. and sometimes - with living things. These elemental particles make up everything, from a small, living ant, to a huge, black, silent, dead and merciless black hole.

Think about it, everything is made up of elementary particles! From those elementary particles you can mould anything, from a stool to a quantum computer, from an ant to a human being! It's not the genes and chromosomes that matter, it's the elementary particles.

Genes folded into chromosomes are only a certain level of abstraction. If you go higher, you receive an individual (a living or a non-living organism). If you go lower, you receive dna, amino acids, oxygen, hydrogen, carbon, and nitrogen. And then you get elemental particles. Move backwards, and you get lithium from hydrogen, and so on, all the way to uranium and beyond! Choose a direction and abstractions and create!

To create, you need an idea. A conception of what's being created. You argue that it is not needed. That all you have to do is shake the "jar" of particles and everything will build itself - the universe and living things... Well, shake it, then...

I would work on concepts first. Of course, there is much more to "shake out" quantitatively than to think about qualitatively... Well, create a suitable GA and let it bubble with entities, and we'll observe.

By the way, I created a thread called "algorithmic centrifuge". I expressed similar views there. I was thinking of creating a "parameter mixer".
 
Aleksey Mavrin:

I've seen these methods from guys who work on so-called "artificial life" projects. Interesting stuff. At first glance, it looks like mush, but genetics makes "intelligent" genes out of mush through millions of passes.

I, too, build strategies out of mush with genetics - sometimes I don't even understand how they work
 
Реter Konow:
To create, you need an idea. A concept of what's being created. You're saying you don't need one. That all you have to do is shake the particle jar a little harder and everything will build itself - the universe and the living things... Well, shake it, then...

I would work on concepts first. Of course, there is much more to "shake things up" quantitatively than there is to think about qualitatively... Well, create a suitable GA and let it bubble with entities and we'll observe.

By the way, I created a thread called "algorithmic centrifuge". I expressed similar views there. I was thinking of creating a "parameter mixer".

If I know that flights into space are possible, should I demonstrate it to you to prove that I am right? ))))

Yes, it is the random search that works wonders. The main thing is the optimization criterion, the fitness function. You have 1+E100 electrons, set ff - density of matter above collapse threshold and you get a black hole with mass of 1+E100 electrons, set ff - black, small, with tendrils and sour *ass, and you get a swarm of ants with total mass of 1+E100 electrons. The main thing is to give it a good shake and set the phf you want.

AZAT KHALITOV:
I too make strategies out of mush with genetics - sometimes I don't even understand how they work

Here, who needs to do it.

 
Andrey Dik:

If I know that flying into space is possible, do I have to demonstrate it to you to prove my point? ))))

Yes, it is the random search that works wonders. The main thing is the optimization criterion, the fitness function. You have 1+E100 electrons, set ff - density of matter above collapse threshold and you get a black hole with mass of 1+E100 electrons, set ff - black, small, with tendrils and sour *ass, and you get a swarm of ants with total mass of 1+E100 electrons. The main thing is to give it a good shake and set the phf you want.

Here, he who needs it does it.

What is the complexity of the systems created by "shaking"? This method is similar to the Rorschach test, where blotches give birth to paintings. By the way, the real approach of contemporary artists is to splash paint on canvases, and the "connoisseurs" themselves will find something to admire.)

Any serious results? Working systems have been created?
123456789101112131415
New comment
Reason: