https://github.com/davidace/emc
(deprecated) Evolutionary Monte Carlo
https://github.com/davidace/emc
Last synced: 3 days ago
JSON representation
(deprecated) Evolutionary Monte Carlo
- Host: GitHub
- URL: https://github.com/davidace/emc
- Owner: DavidAce
- License: mit
- Created: 2015-06-14T07:34:46.000Z (almost 11 years ago)
- Default Branch: master
- Last Pushed: 2015-06-18T09:12:25.000Z (almost 11 years ago)
- Last Synced: 2025-03-01T00:58:45.589Z (over 1 year ago)
- Language: C++
- Homepage:
- Size: 855 KB
- Stars: 1
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README
- License: LICENSE
Awesome Lists containing this project
README
Here is a template ready to play with EMC. It compiles with g++ and gcc (the example works with g++ only).
To begin with, you need to specify the parameter ranges in the function "setupParameters". The number of parameters is set to 8 right now. If you need less paramets you can just set the remaining to the range 0 to 0, for instance.
Then you need to write an expression for the fitness H inside the function "fitnessTest". There are some comments there. In other words, you
need to design a mapping from parameters to a fitness, where a low fitness means good performance.
The fitness function H should ideally be a smooth function roughly between the values 0 and 20 approximately.
Read the EMC article PDF for some details on fitness function design.