Versatile, compact and easy to use Genetic Algorithm and Genetic Programming engine. Simple things are beautiful things, a handful of classes constitute the entire engine! Minimal amount of configuration and composition to get something running!
Licensed under the Apache 2.0 License.
- JRE 8: Java 8 Runtime for using the library and running demonstration programs.
- JDK 8: Java 8 Development Kit for compilation of library and programs using the library.
- MAVEN 3.3.9: Project and dependancy management tool from the Apache Software Foundation.
Ensure you have successfully compiled the project before running the demonstrations by executing:
mvn compile
mvn exec:exec -Dcircles
This demonstration creates a landscape of small circles and then attempts to find the largest circle that can fit amongst them without overlap. Press R to reset the demo or resize the window to make the environment smaller and more crowded.
mvn exec:exec -Dequations
This demonstration attempts to find the best way to arrange a formula of numbers to achieve a target value. There are two sets of numbers, a large set [25, 50, 75, 100] and a small set [1..10]. The target value is a whole number randomly chosen from the range [1..999], and the basic mathematical operators are used, namely +, -, * and /. The challenge here is to find a solution which uses ALL the numbers!
mvn exec:exec -Dpackman
This demonstration first generates a convex polygon and then attempts to fill this polygon with 10 non overlapping circles whilst maximizing polygon coverage. Press R to reset the demo and generate a new convex polygon.
mvn exec:exec -Drostering
This demonstration attempts to find an optimal solution to what is known as the nurse rostering problem, albeit with a vastly more simplified domain model of employees, shifts and abscense and a small handful of constraints.
mvn exec:exec -Dworms
A demonstration of Genetic Programming and emergent behaviour. This experiment attempts to evolve worm like creatures that navigate a procedurally generated 2D world, using cellular automata, avoiding obstacles whilst searching for food.
Minimize the window to prevent replay of current generations champion and thus speed up evolution cycles.
Maximize the window to cause replay of the current generations champion, please note that it might take a while after you maximize for the champion to replay depending on the overall fitness of that generation.
The environment is set to change every 5 minutes to ensure the creatures evolve a more holisitic survival strategy.
- Initial release