ThompsonSamplingPolicy.hpp

Go to the documentation of this file.

#ifndef AI_TOOLBOX_FACTORED_BANDIT_THOMPSON_SAMPLING_POLICY_HEADER_FILE
#define AI_TOOLBOX_FACTORED_BANDIT_THOMPSON_SAMPLING_POLICY_HEADER_FILE
 
#include <random>
 
#include <AIToolbox/Factored/Bandit/Types.hpp>
#include <AIToolbox/Factored/Bandit/Experience.hpp>
#include <AIToolbox/Factored/Bandit/Policies/PolicyInterface.hpp>
#include <AIToolbox/Factored/Bandit/Algorithms/Utils/VariableElimination.hpp>
 
namespace AIToolbox::Factored::Bandit {
    class ThompsonSamplingPolicy : public PolicyInterface {
        public:
            ThompsonSamplingPolicy(const Experience & exp);
 
            virtual Action sampleAction() const override;
 
            virtual double getActionProbability(const Action & a) const override;
 
            static void setupGraph(const Experience & exp, VariableElimination::GVE::Graph & graph, RandomEngine & rnd);
 
            const Experience & getExperience() const;
 
        private:
            const Experience & exp_;
    };
}
 
#endif