The Bandit
library includes various strategies (algorithms) for the multi-armed bandit problem. All the bandit stragegies are implemented as derived class from Gambler.
Strategies relying on the stochastic assumption could be found in the namespace Bandit.Stochastic.
The Bandit.dll
library relies on the MathNet.Numerics
LGPL library available at http://nmath.sourceforge.net.
Class | Description |
---|---|
Gambler | Base class for all the multi-armed bandit strategy. |