Algorithm Optimisation Championship. - page 16

New comment
 
Реter Konow:
What is the connection between the clear surface analogy and the equation example given? Where do they converge?
Sorry, I didn't understand the question. Please expand on the question.
 
By discussing the surface, the task is perfectly clear. Talking about an incomprehensible equation loses the essence of the matter. What does the equation have to do with it?
 
How do you apply a search strategy to a meaningless equation?
 
Реter Konow:
By discussing the surface, the problem is perfectly clear. Talking about an incomprehensible equation loses the essence of the matter. What does the equation have to do with it?

I can't imagine a surface in multidimensional space.

But that doesn't mean that you can't either. If you can imagine a surface in multidimensional space and this helps you in solving the problem, well, very well!

 
Реter Konow:
How do you apply a search strategy to a meaningless equation?
How does MT's in-house optimiser do it? That's exactly what it does day in and day out on millions of users' computers. :)
 
Реter Konow:
How to apply a search strategy to a meaningless equation?
The vector of coefficients (of the equation) will be fed to the input of the algorithm. The algorithm will change the coefficients. The output will be the known discrepancy between the solution with the found coefficients and the most accurate solution where the coefficients are perfectly matched.
 
Andrey Dik:
No, that will not work. You do realize that someone will be doing a complete overkill and then say they found a solution in 27 steps. We are not so naive people (though traders) that would believe in such noodles.

A full enumeration clearly loses out in time. And if there are many An variables, it can take days.

12*A1+4*A2+37*A3+....................................+5*An=S;

This is as far as the complete search is concerned.

About who solves it faster.

Yes, there is a question of who has more powerful hardware and he has an advantage. To make things equal, we can propose a test problem and solve it between the contestants on MQL shared hosting. There one day is free.

SZS. In general, it's just a thought, to attract more participants with different preparation. I wish you good luck with the contest.

 
Once we move away from the surface analogy, we all start speaking different languages.
 
Реter Konow:
As soon as we move away from the surface analogy, we all start speaking different languages.
Are we moving away?
 
Реter Konow:
As soon as we move away from the surface analogy, we all start speaking different languages.

The exact solution is the highest (or deepest) point on the surface in the input vector space. So wander around and look for it.

And what kind of complete search can there be if the data are continuous?

1...91011121314151617181920212223...133
New comment