Policy.hpp

Go to the documentation of this file.

#ifndef AI_TOOLBOX_POMDP_POLICY_HEADER_FILE
#define AI_TOOLBOX_POMDP_POLICY_HEADER_FILE
 
#include <tuple>
 
#include <AIToolbox/POMDP/Types.hpp>
#include <AIToolbox/PolicyInterface.hpp>
 
namespace AIToolbox::POMDP {
    class Policy : public PolicyInterface<size_t, Belief, size_t> {
        public:
            using Base = PolicyInterface<size_t, Belief, size_t>;
            Policy(size_t s, size_t a, size_t o);
 
            Policy(size_t s, size_t a, size_t o, const ValueFunction & v);
 
            // This may be implemented, but probably not since it would be mostly impossible to convert
            // from a POMDP policy format to another.
            // Policy(const PolicyInterface<Belief> & p);
 
            virtual size_t sampleAction(const Belief & b) const override;
 
            std::tuple<size_t, size_t> sampleAction(const Belief & b, unsigned horizon) const;
 
            std::tuple<size_t, size_t> sampleAction(size_t id, size_t o, unsigned horizon) const;
 
            virtual double getActionProbability(const Belief & b, const size_t & a) const override;
 
            double getActionProbability(const Belief & b, size_t a, unsigned horizon) const;
 
            size_t getO() const;
 
            size_t getH() const;
 
            const ValueFunction & getValueFunction() const;
 
        private:
            // H holds the available max horizon for this Policy.
            size_t O, H;
 
            ValueFunction policy_;
 
            friend std::istream& operator>>(std::istream &is, Policy & p);
    };
}
 
#endif