|
k-Armed Bandit 1.0.0
A collection of k-armed bandits and assoicated agents for reinforcement learning
|
Go to the source code of this file.
Namespaces | |
| namespace | analysis |
Variables | |
| int | analysis.K = 10 |
| int | analysis.N = 2000 |
| int | analysis.M = 1000 |
| list | analysis.bandits = [] |
| analysis.single_bandit = bandit.Normal(k=K) | |
| list | analysis.agents |
| list | analysis.agent_names |
| analysis.rewards = numpy.zeros(shape=(len(agents), N, M), dtype=numpy.float) | |
| analysis._table | |
| analysis.a | |
| float | analysis.cumulative_mean_reward = 0.0 |
| analysis.action = test_agent.act() | |
| list | analysis.reward = bandits[n].select(index=action) |
| analysis.mean_rewards = numpy.mean(a=rewards, axis=1) | |
1.9.8