This is a small java library that contains simple algorithms for discrete choices under agent-based models. It's basically a collection of the following:
- Bandit Algorithms
- Online/Recursive Regressions
- Evolutionary methods
The methods are described in a R&R paper, of which a discussion version is available here.
In that paper I split all these methods and use one at a time but in reality if you look at the code it's not too difficult to mix and match (say, use a SOFTMAX bandit algorithm whose memory is actually managed by a Kernel regression and learns from other people's choices as well).
It is complete in the sense that all the algorithms are in but it needs a bit of work in order to make it run on NETLOGO. Once that's settled I'll post this on MAVEN as well.