Saturday 13 December 2014

Development Blog #0003 - Optimization

After scaling the simulation up to Rust's standard 4000mx4000m map, and adding 100 newmans, things slowed down a lot.  I re-did the visualization system.

Watch 100 newmans eat away at a full size map.  The map is 200 tiles across, where each tile is 20m.




Blue - newmans
Brown - trees
Gray - stone
Orange - metal
Yellow - animal


Great, they can eat the map now.  I'm going to invest a bit of time into making this look a bit prettier, so I can see when one square has both wood, stone, and a newman on it. 

After that, I'll implement the combat system.  Like the rest of the simulation, I'll do my best to estimate the value of weapons and armor, then calculate a newman's chance of winning a fight.  It's going to be pretty cool to see newmans eat eachother too.

After combat, I'll create the respawn system for newmans.  I'm choosing to go with the current build's spawning system, placing the newmans on the beach.  Wait, I don't have a beach yet!

After spawning is implemented, I'll make the map spawn look a lot like the current build.  It should be recognizable.  I'll also add background tiles to make it look like a proper model of Rust.


I'd love to make things look just like this:


But it's probably going to look more like this:



I'm not sure which direction I'll take from there.  I could do crafting, leading into base building, or I could do some simple AI - where the newmans would pick targets, and pursue a path to them.

When crafting and base building is in place, I should be dedicating all of my time to the AI.  That's going to be the hard part.  Luckily, I'm a computer science undergraduate with a high interest in AI, so we'll see if I can come up with a newman brain that models the average player accurately.

No comments:

Post a Comment