AIToolbox::Factored::MDP::TigerAntelope Class Reference

This class represents a 2-agent tiger antelope environment. More...

#include <AIToolbox/Factored/MDP/Environments/TigerAntelope.hpp>

Public Member Functions
	TigerAntelope (unsigned width, unsigned height)
	Basic constructor. More...
std::tuple< State, Rewards >	sampleSRs (const State &s, const Action &a) const
	This function allows to sample a new state and rewards. More...
bool	isTerminalState (const State &s) const
	This function returns whether a state is terminal. More...
State	getS () const
	This function returns the state space of the model. More...
Action	getA () const
	This function returns the action space of the model. More...
double	getDiscount () const
	This function returns the discount factor of the model. More...
size_t	getAntelopeState () const
	This function returns the state where the antelope is located. More...
const AIToolbox::MDP::GridWorld &	getGrid () const
	This function returns a reference to the internal GridWorld. More...
std::string	printState (const State &s) const
	This function returns a graphical representation of a State. More...

Detailed Description

This class represents a 2-agent tiger antelope environment.

The two tigers move in a torus grid which always has the antelope at its center. Their goal is to capture it; this can be done when both tigers are adjacent to the antelope, and only one of them moves onto it.

The antelope movement is simulated by shifting the whole world around, so that the antelope is always in the "center" of the state-space. This is done to reduce the size of the state-space from a 3d vector to a 2d vector.

Each tiger can move in one of the 4 cardinal directions, or stay still.

Constructor & Destructor Documentation

TigerAntelope()

AIToolbox::Factored::MDP::TigerAntelope::TigerAntelope	(	unsigned	width,
		unsigned	height
	)

Basic constructor.

Parameters

width	The width of the torus.
height	The height of the torus.

Member Function Documentation

getA()

Action AIToolbox::Factored::MDP::TigerAntelope::getA

(

)

const

This function returns the action space of the model.

getAntelopeState()

size_t AIToolbox::Factored::MDP::TigerAntelope::getAntelopeState

(

)

const

This function returns the state where the antelope is located.

getDiscount()

double AIToolbox::Factored::MDP::TigerAntelope::getDiscount

(

)

const

This function returns the discount factor of the model.

getGrid()

const AIToolbox::MDP::GridWorld& AIToolbox::Factored::MDP::TigerAntelope::getGrid

(

)

const

This function returns a reference to the internal GridWorld.

getS()

State AIToolbox::Factored::MDP::TigerAntelope::getS

(

)

const

This function returns the state space of the model.

isTerminalState()

bool AIToolbox::Factored::MDP::TigerAntelope::isTerminalState

(

const State &

s

)

const

This function returns whether a state is terminal.

Note that this function's return value is not defined for invalid states (for example a state with both tigers in the same place).

Parameters

s	The state to check.

Returns

Whether the state is terminal.

printState()

std::string AIToolbox::Factored::MDP::TigerAntelope::printState

(

const State &

s

)

const

This function returns a graphical representation of a State.

sampleSRs()

std::tuple<State, Rewards> AIToolbox::Factored::MDP::TigerAntelope::sampleSRs	(	const State &	s,
		const Action &	a
	)		const

This function allows to sample a new state and rewards.

Parameters

s	The state to start with.
a	The action to perform.

Returns

A new state and rewards.

---

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

• include/AIToolbox/Factored/MDP/Environments/TigerAntelope.hpp