Activity

samanantha.com

I made the reproduction algorithm better. When it’s time to reproduce, the organisms with the most amount of flags captured for each team reproduce with occasional mutations. I also added an invincibility feature in which the organism does not detect collisions for the first half second after respawning.

I think all that’s left is to make the UI prettier and create graphics for the players. If I’m up to the challenge, I might further improve the combat and reproduction algorithm, because right now, I’m rewarding selfish organisms (organisms that only focus on capturing flags rather than defending or something). If I can find a way to encourage both selfish and selfless organisms, that will make this simulation 10 times better.

Attachment
0
samanantha.com

I finished making the base of the game. I created the colliders (it’s a big buggy, but for the most part it works).

I need to fix my natural selection algorithm because right now, it sucks. I need to find a way to decide which organisms should reproduce and which ones should die.

Once I figure that out, I should focus more on the frontend. For my last project, I kind of neglected the frontend, but for this one, I want the animations/design to actually look good. I don’t have much frontend/design experience, so let’s see how this goes!

Attachment
0
samanantha.com

I have been procrastinating on creating the colliders because such a task seems daunting. So, what I did instead was create more input neurons and debugged the code so that the neural connections actually work.

I also coded the algorithm to detect if a player has the flag. Because it’s incredibly difficult to train a neural network to bring back the flag once it has it, I coded an algorithm that forces the player to bring the flag back to its base regardless of its output neurons. This feels like cheating, but it’s the only way to get this to work lol

To discourage players from getting stuck to the wall, I should make it so that if it stays at the wall for more than 5 seconds, it will die and respawn.

I need to create the respawn algorithm, killing algorithm, and the colliders. After I get the root of the game done, I need to figure out how I want to simulate natural selection here. I have a couple of ideas, but that is a problem for future me!

0
samanantha.com

I revamped the UI so that this simulator is more bearable to look at. I am really happy with the final result!!!

Attachment
0
samanantha.com

I created very basic player movements using only the current player’s position (x and y coordinates) as input neurons. I will try to create the colliders next time I have time to work on this. The colliders seem like a lot of work to implement, so the next devlog have a bunch of hours :)

Attachment
1

Comments

Tom
Tom 16 days ago

Awesome dude

samanantha.com

I created the flag sprites and began coding the internal neurons. I need to figure out how to detect colliders in js, though. The small black box on the top left corner of the field is the player. In the end, I will probably make an actual sprite of the player, but for now, it’ll be a black box :)

Attachment
0
samanantha.com

Shipped this project!

Hours: 19.8
Cookies: 🍪 286
Multiplier: 14.43 cookies/hr

This ended up being more difficult than I thought it would be, but I’m fairly satisfied with the final product. In the future, I would love to attempt to code in carnivores and herbivores to add another layer of complexity beyond the seasons in this simulation.

samanantha.com

I did some minor debugging (got the population and traits counter to work). The simulator is done!

Attachment
0
samanantha.com

I added seasons and a move to group neuron, which encourages organisms to move to a nearby organism. I also removed the ability to change the food spawn rate, so that that variable is only changed through the seasons.

Note to future self: the population counter/organism and their traits list is very faulty (they don’t display the proper population).

Also, something really interesting I observed: over time, the organisms. learn to not move around during winter. I honestly don’t know how that happened, but it somehow did. What they would do is that they would stay still and reproduce (their children spawned with full hunger regardless of parent’s hunger).

As for functionality, all I need to do now is to fix the bugs and make the website a bit prettier. Other than that, I am basically done!

Attachment
1

Comments

samanantha.com
samanantha.com 29 days ago

Another note to self: make it so that food that has been sitting around for a long time decays/disappears somehow

samanantha.com

I got rid of the carnivores because it was too buggy and too difficult to code. Instead, I added options to change the food spawn rate, spread a virus, and change the animation speed.

Next time, I want to add neurons that encourage these organisms to stay together. Another thing on my bucket list is to add seasons and natural disasters that impact the energy it takes to move, food scarcity, etc.

One bug that I was too lazy to fix right now: some squares will stay green (has the cellInfected class) but it isn’t alive (doesn’t have the cellAlive class).

The video quality is pretty terrible due to file size limits.

0
samanantha.com

Added carnivores and herbivores. However, I mainly attempted to fix some bugs, but ended up making it worse :)

Will fix later when I have time

Attachment
0
samanantha.com

Added a few more neurons, created a simple asexual reproduction algorithm with occasional mutations, and created graphs + species stats to understand what is happening in the simulation.

Attachment
0
samanantha.com

Coded the backbone logic and created very basic organism movements.

Attachment
0