mlpack::rl::DiscreteActionEnv Class Reference

To use the dummy environment, one may start by specifying the state and action dimensions. More...

#include <env_type.hpp>

Classes
class	Action
	Implementation of discrete action. More...
class	State
	Implementation of state of the dummy environment. More...

Public Member Functions
double	Sample (const State &, const Action &, State &)
	Dummy function to mimic sampling in an environment. More...
State	InitialSample ()
	Dummy function to mimic initial sampling in an environment. More...
bool	IsTerminal (const State &) const
	Dummy function to find terminal state. More...

Detailed Description

To use the dummy environment, one may start by specifying the state and action dimensions.

Eg:

DiscreteActionEnv::State::dimension = 4;
DiscreteActionEnv::Action::size = 2;

Now the DiscreteActionEnv class can be used as an EnvironmentType in RL methods just as any other mlpack's implementation of gym environments.

Member Function Documentation

InitialSample()

State mlpack::rl::DiscreteActionEnv::InitialSample ( )

inline

Dummy function to mimic initial sampling in an environment.

Returns

the dummy state.

IsTerminal()

bool mlpack::rl::DiscreteActionEnv::IsTerminal ( const State &  ) const

inline

Dummy function to find terminal state.

Parameters

*	(state) The current state.

Returns

It's of no use but so lets keep it false.

Sample()

double mlpack::rl::DiscreteActionEnv::Sample ( const State &  , const Action &  , State &    )

inline

Dummy function to mimic sampling in an environment.

Parameters

*	(state) The current state.
*	(action) The current action.
*	(nextState) The next state.

Returns

It's of no use so lets keep it 0.

---

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

• src/mlpack/methods/reinforcement_learning/environment/env_type.hpp