Comparison of different multi-armd bandit algorithm, based on simulated data.
The code contains Matlab implementation of following band algorithms:
- \epsilon greedy
- \epsilon_{n} greddy
- UCB
- Pursuit
- Softmax
- Exp3
This is one of my previous course project - http://faculty.cse.tamu.edu/nikolova/Teaching/CSCE689_fall2011/
The ComparisonOfBanditAlgorithm.pdf (my report) contains the all the algorithm details and comparisons.
Report bug to nadalwz1115@hotmail.com