AIToolbox
A library that offers tools for AI problem solving.
T3CPolicy.hpp
Go to the documentation of this file.
1 #ifndef AI_TOOLBOX_BANDIT_T3C_POLICY_HEADER_FILE
2 #define AI_TOOLBOX_BANDIT_T3C_POLICY_HEADER_FILE
3 
4 #include <AIToolbox/Bandit/Types.hpp>
5 #include <AIToolbox/Bandit/Experience.hpp>
6 #include <AIToolbox/Bandit/Policies/PolicyInterface.hpp>
7 #include <AIToolbox/Bandit/Policies/ThompsonSamplingPolicy.hpp>
8 
9 namespace AIToolbox::Bandit {
27  class T3CPolicy : public PolicyInterface {
28  public:
36  T3CPolicy(const Experience & exp, double beta, double var);
37 
43  virtual size_t sampleAction() const override;
44 
50  size_t recommendAction() const;
51 
65  virtual double getActionProbability(const size_t & a) const override;
66 
79  virtual Vector getPolicy() const override;
80 
86  const Experience & getExperience() const;
87 
88  private:
89  ThompsonSamplingPolicy policy_;
90  double beta_;
91  double var_;
92  };
93 }
94 
95 #endif
96 
AIToolbox::Bandit::ThompsonSamplingPolicy
This class implements a Thompson sampling policy.
Definition: ThompsonSamplingPolicy.hpp:19
AIToolbox::Bandit::T3CPolicy::getExperience
const Experience & getExperience() const
This function returns a reference to the underlying Experience we use.
Experience.hpp
AIToolbox::Bandit::PolicyInterface
Simple typedef for most of a normal Bandit's policy needs.
Definition: PolicyInterface.hpp:11
AIToolbox::Vector
Eigen::Matrix< double, Eigen::Dynamic, 1 > Vector
Definition: Types.hpp:16
AIToolbox::Bandit::T3CPolicy::getActionProbability
virtual double getActionProbability(const size_t &a) const override
This function returns the probability of taking the specified action.
AIToolbox::Bandit::T3CPolicy::sampleAction
virtual size_t sampleAction() const override
This function chooses an action using T3CPolicy.
AIToolbox::Bandit::T3CPolicy::T3CPolicy
T3CPolicy(const Experience &exp, double beta, double var)
Basic constructor.
AIToolbox::Bandit
Definition: Experience.hpp:6
PolicyInterface.hpp
AIToolbox::Bandit::T3CPolicy
This class implements the T3C sampling policy.
Definition: T3CPolicy.hpp:27
Types.hpp
AIToolbox::Bandit::Experience
This class computes averages and counts for a Bandit problem.
Definition: Experience.hpp:13
AIToolbox::Bandit::T3CPolicy::recommendAction
size_t recommendAction() const
This function returns the most likely best action until this point.
ThompsonSamplingPolicy.hpp
AIToolbox::Bandit::T3CPolicy::getPolicy
virtual Vector getPolicy() const override
This function returns a vector containing all probabilities of the policy.