OraclePolicy: Policy: Oracle

In robinvanemden/contextual: Simulation and Analysis of Contextual Multi-Armed Bandit Policies

Description

OraclePolicy is a also known as a "cheating" or "godlike" policy, as it knows the reward probabilities at all times, and will always play the optimal arm. It is often used as a baseline to compare other policies to.

Usage

policy <- OraclePolicy()

Arguments

name

character string specifying this policy. name is, among others, saved to the History log and displayed in summaries and plots.

Methods

new()

Generates a new OraclePolicy object. Arguments are defined in the Argument section above.

set_parameters()

each policy needs to assign the parameters it wants to keep track of to list self$theta_to_arms that has to be defined in set_parameters()'s body. The parameters defined here can later be accessed by arm index in the following way: theta[[index_of_arm]]$parameter_name

get_action(context)

here, a policy decides which arm to choose, based on the current values of its parameters and, potentially, the current context.

set_reward(reward, context)

in set_reward(reward, context), a policy updates its parameter values based on the reward received, and, potentially, the current context.

References

Gittins, J., Glazebrook, K., & Weber, R. (2011). Multi-armed bandit allocation indices. John Wiley & Sons. (Original work published 1989)

See Also

Core contextual classes: Bandit, Policy, Simulator, Agent, History, Plot

Bandit subclass examples: BasicBernoulliBandit, ContextualLogitBandit, OfflineReplayEvaluatorBandit

Policy subclass examples: EpsilonGreedyPolicy, ContextualLinTSPolicy

robinvanemden/contextual documentation built on Aug. 12, 2019, 9:30 p.m.