Welcome to Artificial Life — a 2D virtual evolution simulator where artificial creatures evolve by interacting with plants, eggs, other creatures, and their environment. You can run simulations over long periods, watch species emerge and adapt, and explore the dynamics of artificial ecosystems.
-
🧠 Real-Time Evolution
Creatures evolve using TWEANN (Topology and Weight Evolving Artificial Neural Networks), powered by a NEAT (NeuroEvolution of Augmenting Topologies) genetic algorithm. This allows complex behaviors to emerge organically without external programming. -
♻️ Continuous Simulation
Unlike traditional simulations with fixed endpoints, this one runs indefinitely. Expect millions of births and deaths, leading to emergent species, survival strategies, and even extinction events. -
🦎 Evolving Creatures
Each creature evolves both mentally and physically: attributes like digestion, vision, strength, armor, and size all change. Creatures may evolve toward herbivory, carnivory, or hybrid diets. -
⚡ High-Performance Engine
Optimized collision detection and spatial partitioning enable extremely fast simulations (hundreds of thousands of ticks/second), with multi-threading for long evolutionary runs.
🌍 World snapshot (12 ticks/sec)
🔍 Creature Info Tab (WIP)
👁️ Debug mode (vision cone + grid)
🧬 Brain viewer + forced mutation
⚙️ Simulation settings menu
📈 Graph of evolving traits
The neural architecture was implemented from scratch over ~1.5 months. RT-NEAT (Real-Time NEAT) was particularly challenging due to sparse documentation. Extensive parameter tuning ensures adaptive, realistic behavior in creatures.
Built for performance, the engine handles thousands of dynamically-sized entities using optimized spatial partitioning and collision detection. Designed to support billion-tick evolutionary cycles.
Creatures evolve both structurally and functionally: vision, diet, strength, and size all co-evolve with their neural networks. This allows for rich emergent diversity — predators, grazers, tanks, and everything in between.
The GUI is feature-rich and built for deep exploration:
- Real-time neural network visualization
- Graphs showing population stats & trait trends
- Creature inspector and dynamic mutation tools
- Toggle views for vision cones, hitboxes, and partitioning grids
- Save/load simulation states for long-term evolution
This project uses standard Java and requires no external dependencies.
- Java JDK 17+ installed
- IntelliJ IDEA (recommended) or another Java-compatible IDE
-
Clone this repository
git clone https://github.com/your-username/artificial-life.git cd artificial-life -
Open in IntelliJ IDEA
-
Go to File > Open, and select the project folder.
-
Wait for IntelliJ to index the files.
- Run the simulation
-
Open the main class (e.g.,
Main.java) -
Right-click and choose Run.
📝 No build tools like Maven or Gradle are needed for pre-release versions. Just open, build, and run.
-
Add support for player-controlled creatures
-
Refactor simulation into a proper MVC architecture
-
Improve world saving/loading
-
Eliminate all multi-threading bugs
-
Integrate reinforcement learning protocols
-
Improve UI appearance and interactivity
-
Optimize collision detection & memory usage
-
Add biomes, temperature, terrain height, etc. (Low Priority)
This project was inspired by many researchers and creators exploring artificial life, NEAT, and digital evolution:
-
Stanley, Kenneth O., & Miikkulainen, R. (2002).
Evolving Neural Networks through Augmenting Topologies.
Read the paper
This project is licensed under the MIT License - see the LICENSE file for details.
Thank you for exploring Artificial Life. This project is continually evolving, so check back often for updates as new features and optimizations are added.