Gene Pool in a growing order

[luca.cornia]

Hello everyone,
I’m trying to create a diagrid structure on a cylinder like column and I’ll be using Galapagos to do some structure optimization on it.
I’m using Gene Pool to get a series of heights for the cross section to be placed in and another one to set the values for the radiuses of the sections.
Do you think it’s possibile to constraint the values in a Gene Pool for them to assume a monotonic increasing order?

[zhuangjia777]

I have a similar need like yours. I also want to know how to achieve monotonic items in gene pool.

[inno]

you can remap Gene Pool values using (static) domains, like this

but reordering the Gene Pool values (for instance by using the Sort component after the Gene Pool) or changing the remapping range of the Gene Pool values while Galapagos is running (for instance by remapping their values dinamically) both should render Galapagos useless

[luca.cornia]

Yeah I think so… I tried to simply remap the values output it but I feared that Galapagos would not work that way. That’s why I was thinking about restraining the Gene Pool values to be in an increasing order. I guessed that leaving some interval between values the solver could, in theory, change them a bit from the previous to the following one. Well, probably that is not the way it was designed to.
Thanks anyway mate, much appreciated

[luca.cornia]

I’ll simply apply the solver to a series of sliders, each one with min and max increasing with fixed ranges… maybe that’s the only way. I tried to avoid the countless wires around.
Thanks again guys, if I find a way I’ll post it

[luca.cornia]

In case you’re searching for some tips don’t fear to ask. I’m not sure I’ll be able to help but if I can I’ll gladly do that

[zhuangjia777]

I am thinking if mass adding the output of gene pool, and then remap addition results to your aim domain.

[inno]

that should not work for the reason described in this discussion (3 years old, but I think still on-spot)

Is there a way to make Galapagos work with Domains instead of sliders?

Galapagos works strictly with integer variables. If your slider goes from 0.0 to 2.0 with three decimal places, there’s a total of 2001 unique values it can assume. Galapagos treats this slider as an integer value between 0 and 2001. Any changes made to the slider accuracy or domain may result in a different number of discrete states and that would seriously complicate the algorithm. Also it would mean that state 258 would map to a different numerical value than it used to, so now the model an…

@DavidRutten

Also it would mean that state 258 would map to a different numerical value than it used to, so now the model and Galapagos’ understanding of the model are diverging.

[luca.cornia]

Do you mean like the process that Inno illustrated in the previous image?
Or you’re thinking about summing all the values in the gene pool in bulk? Sorry if I ask but I didn’t really grasped the meaning of that

[luca.cornia]

Oopsie, you’re right again. I’ll use integers but if I link lots and lots of sliders even if with little domains, like a range of 3 or 4, it will become a real pain for the solver.
I fear I’ll have to think really hard about a better way to resolve my dilemma.
@inno many thanks again, really!

[zhuangjia777]

I mean like this.

[luca.cornia]

Oh ok, didn’t know about the Mass Addition component. Nice trick, thanks for the idea

[inno]

you’re welcome

using integers or reals for the sliders should not really change anything, it’s just on Galapagos-side interpretation: it will consider the states the sliders can be, so [0 ; 99] has 100 different states like [0.00 ; 1.00]

@zhuangjia777

I mean like this.

I think by using Mass Addition you are actually dynamically remapping the values of the Gene Pools, falling into this situation: every time a slider changes, the end value of the Source domain also changes

@DavidRutten

Also it would mean that state 258 would map to a different numerical value than it used to, so now the model and Galapagos’ understanding of the model are diverging.