AIToolbox::PolicyInterface< State, Sampling, Action > Class Template Reference

This class represents the base interface for policies. More...

#include <AIToolbox/PolicyInterface.hpp>

Inheritance diagram for AIToolbox::PolicyInterface< State, Sampling, Action >:

Public Member Functions
	PolicyInterface (State s, Action a)
	Basic constructor. More...
virtual	~PolicyInterface ()
	Basic virtual destructor. More...
virtual Action	sampleAction (const Sampling &s) const =0
	This function chooses a random action for state s, following the policy distribution. More...
virtual double	getActionProbability (const Sampling &s, const Action &a) const =0
	This function returns the probability of taking the specified action in the specified state. More...
const State &	getS () const
	This function returns the number of states of the world. More...
const Action &	getA () const
	This function returns the number of available actions to the agent. More...

Protected Attributes
State	S
Action	A
RandomEngine	rand_

Detailed Description

template<typename State, typename Sampling, typename Action>
class AIToolbox::PolicyInterface< State, Sampling, Action >

This class represents the base interface for policies.

This class represents an interface that all policies must conform to. The interface is generic as different methods may have very different ways to store and compute policies, and this interface simply asks for a way to sample them.

This class is templatized in order to work with as many problem classes as possible. It doesn't really implement any specific functionality, but should be used as a guide in order to decide which methods a new policy should have.

The State and Sampling parameters are separate since in POMDPs we still have "integer" states, but we often want to sample from beliefs. Having two separate parameters allows for this. In the MDP case they are both the same.

Template Parameters

State	This defines the type that is used to store the state space.
Sampling	This defines the type that is used to sample from the policy.
Action	This defines the type that is used to handle actions.

Constructor & Destructor Documentation

PolicyInterface()

template<typename State , typename Sampling , typename Action >

AIToolbox::PolicyInterface< State, Sampling, Action >::PolicyInterface	(	State	s,
		Action	a
	)

Basic constructor.

Parameters

s	The number of states of the world.
a	The number of actions available to the agent.

~PolicyInterface()

template<typename State , typename Sampling , typename Action >

AIToolbox::PolicyInterface< State, Sampling, Action >::~PolicyInterface

virtual

Basic virtual destructor.

Member Function Documentation

getA()

template<typename State , typename Sampling , typename Action >

const Action & AIToolbox::PolicyInterface< State, Sampling, Action >::getA

This function returns the number of available actions to the agent.

Returns

The total number of actions.

getActionProbability()

template<typename State , typename Sampling , typename Action >

virtual double AIToolbox::PolicyInterface< State, Sampling, Action >::getActionProbability ( const Sampling &  s, const Action &  a  ) const

pure virtual

This function returns the probability of taking the specified action in the specified state.

Parameters

s	The selected state.
a	The selected action.

Returns

The probability of taking the selected action in the specified state.

Implemented in AIToolbox::EpsilonPolicyInterface< void, void, size_t >, AIToolbox::EpsilonPolicyInterface< State, State, Action >, AIToolbox::EpsilonPolicyInterface< size_t, size_t, size_t >, AIToolbox::MDP::WoLFPolicy, AIToolbox::MDP::PGAAPPPolicy, AIToolbox::MDP::QSoftmaxPolicy, AIToolbox::MDP::PolicyWrapper, AIToolbox::MDP::QGreedyPolicy, AIToolbox::MDP::BanditPolicyAdaptor< BanditPolicy >, and AIToolbox::EpsilonPolicyInterface< State, Sampling, Action >.

getS()

template<typename State , typename Sampling , typename Action >

const State & AIToolbox::PolicyInterface< State, Sampling, Action >::getS

This function returns the number of states of the world.

Returns

The total number of states.

sampleAction()

template<typename State , typename Sampling , typename Action >

virtual Action AIToolbox::PolicyInterface< State, Sampling, Action >::sampleAction ( const Sampling &  s ) const

pure virtual

This function chooses a random action for state s, following the policy distribution.

Parameters

s	The sampled state of the policy.

Returns

The chosen action.

Implemented in AIToolbox::EpsilonPolicyInterface< void, void, size_t >, AIToolbox::EpsilonPolicyInterface< State, State, Action >, AIToolbox::MDP::WoLFPolicy, AIToolbox::EpsilonPolicyInterface< size_t, size_t, size_t >, AIToolbox::MDP::PGAAPPPolicy, AIToolbox::MDP::QSoftmaxPolicy, AIToolbox::MDP::PolicyWrapper, AIToolbox::MDP::BanditPolicyAdaptor< BanditPolicy >, AIToolbox::MDP::QGreedyPolicy, and AIToolbox::EpsilonPolicyInterface< State, Sampling, Action >.

Member Data Documentation

A

template<typename State , typename Sampling , typename Action >

Action AIToolbox::PolicyInterface< State, Sampling, Action >::A

protected

rand_

template<typename State , typename Sampling , typename Action >

RandomEngine AIToolbox::PolicyInterface< State, Sampling, Action >::rand_

mutableprotected

S

template<typename State , typename Sampling , typename Action >

State AIToolbox::PolicyInterface< State, Sampling, Action >::S

protected

---

The documentation for this class was generated from the following file:

• include/AIToolbox/PolicyInterface.hpp