Discussion of article "Population optimization algorithms: Saplings Sowing and Growing up (SSG)" - page 7

New comment
 
Nikolai Semko #:

but still, I think, finding such points had a probabilistic character.
My message is that there should be no input parameters requiring "optimisation" at the TC input. Such parameters turn the TS into a coin.
Even internal self-optimisation of internal parameters is also self-deception.

I admit that there is food that doesn't need to be cooked. But I have not been able to discover such food. And cooked food still allows you to not be hungry. Yes, it's not perfect, but it's better than nothing.

 
mytarmailS #:
The last 100 records will not correspond to different peaks, they will all be centred around one last found peak

The goal is the same for all of them - to find the maximum, but the ways of achieving the goals are different for all of them. some algorithms, like BFA, are forbidden to cluster all of them together in one peak.
 
mytarmailS #:
The last 100 records will not correspond to different peaks, they will all be centred around one last peak found

Almost like that (depends on the algorithm). This last peak will be the "most" - global. It will be discarded for further optimisations.

This way we will get a list of peaks: from Everest to the hill.

 
fxsaber #:

I suppose there are foods that don't need to be cooked. But I've never been able to discover such food. And boiled food does keep you from being hungry. It's not perfect, but it's better than nothing.

illusionary food creates the illusion of satiety

 
Nikolai Semko #:

but still, I think, finding such points
Even internal self-optimisation of internal parameters is also self-defeating.

Why?
 
Nikolai Semko #:

illusory food creates the illusion of satiety

This was a pretty good branch. Let's not spoil it.

 
gravity search is interesting in this sense, gravity can be set "in reverse", when all particles repel each other.
electromagnetic search is even cooler, there is a positive and negative charge, particles both repel and attract. let's see what kind of miracle will turn out, the article is being prepared.
 
fxsaber #:

Almost so (depends on the algorithm). This last peak will be the "most" - global. It will be discarded for subsequent optimisations.

Well, it's quite easy to implement

Just AO 2 should be forbidden to generate a solution similar to the solution of AO 1.

And then AO3 should be forbidden to plagiarise solutions from AO1 and AO2, etc.

By the way, how many dimensions of the search space do you have?
 
fxsaber #:

This was a pretty good thread. Let's not spoil it.

I realise it looks like I'm trolling. But I'm actually saying a very important thing and trying to save you time.

 
Andrey Dik #:
gravitational search is interesting in that sense, grvitation can be set "backwards" when all particles repel each other.
electromagnetic search is even cooler, there is a positive and negative charge, particles both repel and attract. let's see what kind of miracle will turn out, the article is being prepared.

If it works, then one optimisation + clustering of results will be enough.

12345678910111213
New comment