Algorithm Optimisation Championship. - page 124
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Everything has already been stolen before us. There are gene algorithms and neural networks and stat algorithms and much more. Take it, configure it for your task, and use it.
I'm not a fan of reinventing bicycles). Only if there is a great need). It is better and more interesting to do something really useful or at least just interesting.
But to look from the outside might be interesting.
D You've already said it and cited beautiful pictures of finding the maximum by mathematical software. I asked you a question, you didn't answer. I will repeat
And when the function is unknown to the software, can this software determine the extremums? How many operations and transformations must be performed to obtain the result in the EA per mkl? How long will it take? As far as I understand, the participants will send to the black box a set of parameter values as an array doudle[x1, x2, xn].
They get the value of the function from the "black box" into their algorithm, send the next set and so on until they find the extremum. We need to make as few calls as possible in the MT environment. I think it is a useful thing in the household.
D You already said that and gave nice pictures of the mathematical software finding the maximum. I asked a question, YOU didn't answer. I will repeat it
And when the function is unknown to the software, can this software determine the extremum? How many operations and transformations must be performed to obtain the result in the EA per mkl? How long will it take? As far as I understand, the participants will send to the black box a set of parameter values as an array doudle[x1, x2, xn].
They get the value of the function from the "black box" into their algorithm, send the next set and so on until they find the extremum. We need to make as few calls as possible in the MT environment. I think it is a useful thing in the household.
When the function is unknown, of course it can. The more data, the more accurate, of course. I don't know about MCL. How long will it take? - Like in the movie "The 5th Element" - "time doesn't matter.")
Sorry, but I don't understand - why is this thing needed in the household and in the MT environment?
If you want to program something, go to the "Ticket to the Future" thread, read it. I didn't start it, but I happen to keep going. Maybe you will help me with it, if you are interested. It's also a problem with no clear-cut conditions.)
When the function is unknown, of course it can. The more data, the more accurate, of course. I don't know about the MCL. How long will it take? - Like in the movie "The 5th Element" - "time doesn't matter.")
I'm sorry, but I don't understand why this thing is needed in a household and in a MT environment?
Time is what matters. Imagine that in a break between ticks you urgently need to optimize something.
There were examples in the thread. But it is very difficult to find them among a thousand posts. I can only give you a link to what I wrote in the context of trading.
Time is of the essence. Imagine, between ticks, you need to optimise something urgently.
There were examples in the thread. But it is very difficult to find them among a thousand posts. I can only give you a link to what I wrote about trading.
Imho, optimisation problems should be solved specifically for a particular object, rather than in the abstract, like for all occasions.
As far as I understand, the regular GA of MT's optimizer doesn't give a damn what objects are in front of it and what cockroaches the author put into his Expert Advisor.
That's what confuses me).
And the notion of optimality (optimization) is not always a search for the maximum or minimum.
That's what confuses me).
But the notion of optimality (optimization), is by no means always a search for a maximum or minimum.
I agree. But if we discuss further, I'm afraid we will go on a nth circle in the branch.
I wish you success in the "Ticket to the Future" project. Perhaps if successful, there will be no need to invent bicycles. I can not join as I have no system knowledge in IT.
But if you have time, come here.
Objects - are they specific PBXs, EAs? As far as I understand, the main optimization GA of MT does not really care what objects are in front of it or what kind of nonsense the author put into his EA.
It seems to me that the topic of optimization in the interpretation of the topic was covered with such a fog that to this day nobody clearly understands what the actual point is.
This thread has seen everything. There were binary chromosomes, genes, genomes, species, populations, evolution, selection, interbreeding, multidimensional space and much more...
What has been missing here is clarity.
What is the purpose of optimization in trading? - To adjust TS parameters to obtain the maximal profitability for the period being tested. It is hard to imagine another application.
What are the criteria for evaluation?
1. Minimal number of calculations.
Accuracy.
Universality - nonsense. There is no need for any universal code. A clearly defined algorithm that solves a particular task.
What was presented by the topic-starter as "the universality of the algorithm" is just a generalization of a particular group of tasks.
Universality cannot be a property of an algorithm that simply takes a set of some numbers (no matter how it defines them) and passes them to the FF. Then it gets a value from the FF and uses it for further calculations. What universality is there if the mechanism of solving all optimization problems is exactly the same?
I agree. But if we discuss it further, I'm afraid we'll go in a nth circle in the branch.
Good luck with the Ticket to the Future project. Perhaps if successful, there will be no need to invent bicycles. I can not join as I have no system knowledge in IT.
But if you have the time, come here.
Thank you. I don't know yet, if help doesn't come, I'll abandon it as a public project.
I watch your thread regularly and enjoy it).
It seems to me that the topic of optimization as interpreted by the topic-starter is covered with such a fog that no one clearly understands what we are actually talking about.
This thread has seen everything. There were binary chromosomes, genes, genomes, species, populations, evolution, selection, interbreeding, multidimensional space and much more...
What has been missing here is clarity.
What is the purpose of optimization in trading? - To adjust TS parameters to obtain the maximal profitability for the period being tested. It is hard to imagine another application.
What are the criteria for evaluation?
1. Minimal number of calculations.
Accuracy.
Universality - nonsense. There is no need for any universal code. A clearly defined algorithm that solves a particular task.
What the topic starter has presented as the "universality of the algorithm" is just a generalization of a particular group of tasks.
Universality cannot be a property of an algorithm that simply takes a set of some numbers (no matter how it defines them) and passes them to the FF. Then it receives a value from the FF and uses it for further calculations. What universality is there if the mechanism of solving all optimization problems is exactly the same?
In many ways I agree with you. But since the topicstarter is in a sauna and can not answer us, I would not want to discuss his championship idea.
Maybe we'd better leisurely prepare our codes, without competitive interface, and then we'll see.