Evolving Game Development with Genetic Algorithms
summary
A playable web game called egg drop is built with React and conva, assets made via Logoist and Texture Packer, audio with howler js, and state logic in xate. A genetic algorithm evolves a population of hens (initially ~10, later streamed across the stage), using roulette selection, crossover and mutation; metrics are tracked and one run shows trends over 16 generations where hen speed and lay rates rise and player catch rate slowly falls, so the GA nudges difficulty but needs more tuning.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 1:35Genetic algorithms through nature
- 2:38Speaker introduction
- 4:18Atari and Kaboom inspiration
- 6:31Egg drop game concept
- 7:13React and Conva game
- 8:13Game mechanics and movement
- 10:46Player interaction and catching
- 11:27Hit detection and physics
- 12:11State machines in games
- 13:54xstate visualization
- 15:20AI tools for assets
- 18:51Vector graphics sprites
- 20:17Sprite animation assembly
- 21:03Game background design
- 21:45Typography and fonts
- 23:09Game title and UI
- 23:50Sound effects added
- 24:55Genetic algorithms basics
- 28:26Applying genetic algorithms to gameplay
- 30:33Population size optimization
- 32:16GA logic separation
- 34:03GA testing evolution
- 34:45GA performance analysis
- 36:10Project summary
- 36:51Further learning resources
- 37:35Closing thoughts
related talks
30mEvolving JavaScript: Cultivating Genetic Algorithms for Creative Coding
29mEvolving JavaScript: Cultivating Genetic Algorithms for Creative Coding
42m¡Viva La Evolución! Replicating Life with OTP
44mGame Development - The Ruby way
24mWe Will All Be Game Programmers
18mThink You Can Build a Game with AI? Think Again!
46mBuild Your Own Game Engine in the Browser No experience necessary, all experience
27mCSS as a Game Engine: Building Playful, Stateful Web Experiences Without JS