Flat Earth

Some challenges are hard :). Here we can discuss their solutions. Please find the thread for the appropriate challenge or create it if it doesn't exist.
User avatar
Nayus
Posts: 679
Joined: Thu Oct 06, 2016 2:03 am
Location: Argentina

Flat Earth

Post by Nayus » Sun Sep 16, 2018 11:59 pm

Flat Earth

Difficulty: Mad Scientist

Progress: Optional challenge

Description: This Petri dish is slippery and has no rim so cells fall and die when they reach the edge.
Can you nevertheless get 300 cells?


Win Conditions: A population of 300 or more user cells for 100 or more h with 5 or less cells inserted


General Strategy: Use the wall detection of the stereocytes

Record Solution: ?
Pixlo8
Posts: 8
Joined: Sat Oct 06, 2018 5:20 pm

Re: Flat Earth

Post by Pixlo8 » Tue Nov 27, 2018 12:22 am

I did it! This one has been giving me some trouble for a while, I ended up solving it actually by taking the prey from predator and evolving it to be able to hunt in slippery substrates, and then maxed the cell count by adding a lipocyte for longevity. They're pretty fun, here's a download link if you wanna check it out!

https://drive.google.com/file/d/1mFxrHF ... p=drivesdk
Screenshot_2018-11-26-19-21-55.png
User avatar
H4yw1r3
Posts: 532
Joined: Thu Jul 13, 2017 5:19 pm
Contact:

Re: Flat Earth

Post by H4yw1r3 » Tue Nov 27, 2018 10:10 am

It looks like we have almost the same layout for the genome, except mine only has 4 cells (Diamond Formation) and uses its "LipoEggs" to balance itself to its food. Your organism is so agressive when approaching pieces of food, while mine's passively swimming towards it.

Great job for finishing the challenge!
Are complex creatures too advanced for you to understand?
Well, let me flood you with them.
User avatar
wapcaplet
Posts: 321
Joined: Thu Dec 08, 2016 9:52 pm

Re: Flat Earth

Post by wapcaplet » Sat Dec 01, 2018 2:09 am

I have solved this challenge a few times using complicated omnidirectional swimmers that can detect both food bits and the substrate wall, but seeing Pixlo's screenshot inspired me to consider simpler designs (though unfortunately I couldn't download their substrate).

Then the workings of evolution on my design inspired me to try another design, which evolved to inspire another! And thus a whole new family of tumbling and spinning flat-earthers came to be.

This substrate has three different designed organisms:

https://www.dropbox.com/s/tkbj5x63xjdg4 ... trate?dl=0

Red was the first concept, a lazy lateral drifter similar to the orange one Pixlo made. This later evolved to have a rapid spinning behavior, which I tried to capitalize on in my design of the yellow trapezoidal model, and the simpler blue version that came next.

Once I had a semi-viable design, I knew I could rely on evolution to optimize and explore alternatives that I would never think of. And it sure did: my lazy lateral drifter became a lean, speedy spinner that could run circles around its ancestor. Something similar occurred with my subsequent yellow trapezoidal design, shedding a phagocyte and becoming the "dizzy shrimp" you see in the evolved substrate below:

https://www.dropbox.com/s/ns0mhlrvvf5g6 ... trate?dl=0


Evolution really favors a spinning motion for these, which makes a lot of sense if you look at how much surface area it can sweep its phagocytes across with a minimum of energy. Their approach to seeking food is haphazard and imprecise, but they more than make up for it by being in a near constant state of spinning, grabbing food in fairly wide radius.

The blue single-flagello species has some of this spinning behavior, but also enters a "resting" state when no food is sensed on one side, letting them conserve energy.

I was a little surprised to learn that wall-detection is not that important in this challenge, if your organism is efficient enough at seeking food. My red and blue spinners do not have wall detection, and I'm not convinced it's doing that much good for the yellow one either.

To others attempting to solve this challenge, I would suggest keeping your organism as simple as possible; start with a low swim speed and careful movements, and don't worry too much about the wall. Get a viable start, then let evolution do the fine-tuning for you. And don't get too dizzy!
Attachments
Capture+_2018-11-30-18-04-04~2.png
Evolved swimmers
Capture+_2018-11-30-18-18-57~2.png
Designed swimmers
Pixlo8
Posts: 8
Joined: Sat Oct 06, 2018 5:20 pm

Re: Flat Earth

Post by Pixlo8 » Sat Dec 01, 2018 7:45 am

H4yw1r3 wrote:
Tue Nov 27, 2018 10:10 am
It looks like we have almost the same layout for the genome, except mine only has 4 cells (Diamond Formation) and uses its "LipoEggs" to balance itself to its food. Your organism is so agressive when approaching pieces of food, while mine's passively swimming towards it.

Great job for finishing the challenge!
Thanks! And I actually went in and updated it, I optimized it to be able to work as a collective worm and avoid the edges.
https://drive.google.com/file/d/17QLBZ4 ... p=drivesdk
Screenshot_2018-12-01-02-37-36.png
Pixlo8
Posts: 8
Joined: Sat Oct 06, 2018 5:20 pm

Re: Flat Earth

Post by Pixlo8 » Sat Dec 01, 2018 7:52 am

wapcaplet wrote:
Sat Dec 01, 2018 2:09 am
I have solved this challenge a few times using complicated omnidirectional swimmers that can detect both food bits and the substrate wall, but seeing Pixlo's screenshot inspired me to consider simpler designs (though unfortunately I couldn't download their substrate).

Then the workings of evolution on my design inspired me to try another design, which evolved to inspire another! And thus a whole new family of tumbling and spinning flat-earthers came to be.

This substrate has three different designed organisms:

https://www.dropbox.com/s/tkbj5x63xjdg4 ... trate?dl=0

Red was the first concept, a lazy lateral drifter similar to the orange one Pixlo made. This later evolved to have a rapid spinning behavior, which I tried to capitalize on in my design of the yellow trapezoidal model, and the simpler blue version that came next.

Once I had a semi-viable design, I knew I could rely on evolution to optimize and explore alternatives that I would never think of. And it sure did: my lazy lateral drifter became a lean, speedy spinner that could run circles around its ancestor. Something similar occurred with my subsequent yellow trapezoidal design, shedding a phagocyte and becoming the "dizzy shrimp" you see in the evolved substrate below:

https://www.dropbox.com/s/ns0mhlrvvf5g6 ... trate?dl=0


Evolution really favors a spinning motion for these, which makes a lot of sense if you look at how much surface area it can sweep its phagocytes across with a minimum of energy. Their approach to seeking food is haphazard and imprecise, but they more than make up for it by being in a near constant state of spinning, grabbing food in fairly wide radius.

The blue single-flagello species has some of this spinning behavior, but also enters a "resting" state when no food is sensed on one side, letting them conserve energy.

I was a little surprised to learn that wall-detection is not that important in this challenge, if your organism is efficient enough at seeking food. My red and blue spinners do not have wall detection, and I'm not convinced it's doing that much good for the yellow one either.

To others attempting to solve this challenge, I would suggest keeping your organism as simple as possible; start with a low swim speed and careful movements, and don't worry too much about the wall. Get a viable start, then let evolution do the fine-tuning for you. And don't get too dizzy!
Wow! Cool to have inspired you. If you notice my other post I just added in wall detection funny enough. Mainly just because I like the worm designs, I think they're fun, and this one can get to 650 cells on the substrate so it's pretty decent. I think when a large number of cells are connected they can share nutrients so well they can accommodate the extra sterocyte, and their natural movement otherwise leads to many of them dying, so it really makes the worm design possible overall.

Interesting stuff isn't it?
Pixlo8
Posts: 8
Joined: Sat Oct 06, 2018 5:20 pm

Re: Flat Earth

Post by Pixlo8 » Sat Dec 01, 2018 6:10 pm

Just saw these links for my creations might not be working, here's a substrate showing all the versions I saved start fo finish from top to bottom
https://mega.nz/#!QkIRiCaY!nOJhyLKedfE_ ... ZS12_wMrjs
User avatar
wapcaplet
Posts: 321
Joined: Thu Dec 08, 2016 9:52 pm

Re: Flat Earth

Post by wapcaplet » Sat Dec 01, 2018 11:06 pm

Thanks, I have never been able to get cell lab files to download correctly in Google drive (though I use drive all the time for other stuff). The new link you shared works for me. Dropbox is another good choice.

I like how you used the lipocyte as a pivot point, or center of gravity. I see that your organisms also tend toward a tumbling or spinning motion, at least the non-worm ones do. The worm/snake design is good for keeping a lot of cells alive together, since they can share energy, but I've found it limits their mobility, especially when they get competing signals from food on both sides. But once they get moving, they can sweep a large area and grow quickly.

I'm working on an even more stripped down 3-mode version that can almost solve the challenge, using the spinning strategy. But now I am wondering about spinning snakes... Could that work too?
User avatar
wapcaplet
Posts: 321
Joined: Thu Dec 08, 2016 9:52 pm

Re: Flat Earth

Post by wapcaplet » Fri Dec 07, 2018 3:57 am

Well, here it is: a solution to "Flat Earth" using only three modes.
Capture+_2018-12-06-20-52-57_crop_543x762-217x305.png
Capture+_2018-12-06-20-52-57_crop_543x762-217x305.png (25.03 KiB) Viewed 90 times
Substrate: https://www.dropbox.com/s/jltla3idn6y7y ... trate?dl=0

The blue version was my design, and the purple version is after about 170,000h of evolution. I've put them both on the challenge substrate for comparison; the blue ones will soon go extinct, since the purple ones are so much more efficient.

The purples can win the challenge, sustaining around 320 cells (though they sometimes dip below 300). The blues can't, unless they get extremely lucky.

It was inspired by the natural spinning evolution of my previous organisms, but taken down to its bare essence: flagello, stereo, phago. The phagocyte used for feeding is in fact the egg of the next organism; when it splits, hopefully there is enough energy left in the stereocyte (being high priority) to produce a new egg; otherwise, the parent will die.

It could probably be made more efficient by adding a mode or two (while keeping the 3-cell layout) but I thought it would be fun and challenging to solve it with only 3 modes.

I love the wide swath of substrate an organism can scour with a spinning phagocyte. Because of the low viscosity here, it needs very little energy to keep spinning once started. In a way it's like having a (flat) donut-shaped phagocyte with a much larger surface area.

Steering only needs to be approximate. As long as the stereocyte can help it get anywhere near the food, the spinning will take care of the rest.
SlowMind
Posts: 49
Joined: Wed May 30, 2018 6:56 am
Location: Philippines

Re: Flat Earth

Post by SlowMind » Sun Dec 09, 2018 11:38 am

I made one with hidden flagellocyte

https://www.datafilehost.com/d/1a32d844
uhhhmmm...
Post Reply