RLGlueAgent

java.lang.Object
- burlap.behavior.singleagent.interfaces.rlglue.RLGlueAgent

All Implemented Interfaces:

Environment, org.rlcommunity.rlglue.codec.AgentInterface
```
public class RLGlueAgent
extends java.lang.Object
implements Environment, org.rlcommunity.rlglue.codec.AgentInterface
```
An RLGlue Agent that is implemented as a BURLAP Environment, so that any BURLAP RL agent can control the RLGlue agent.
To initialize this BURLAP environment, use the loadAgent() or loadAgent(String, String) methods so that RLGlue knows to tell this class about the environment this class is wrapping. After the load method, you can get the corresponding BURLAP domain that represents it by calling the getDomain() method. You can also get the RLGlue preferred discount factor using the getDiscountFactor() method.
As with normal BURLAP Environment implementations, you should call the resetEnvironment() method whenever a terminal state is reached. This will cause this class to block until RLGlue has started a new episode with a new initial state, unless the RLGlue experiment has finished, in which case the method will return immediately. If you want to check whether the RLGLue experiment has finished manually, use the rlGlueExperimentFinished method.

Author:

James MacGlashan.

Nested Class Summary

Nested Classes
Modifier and Type	Class and Description
`static class`	`RLGlueAgent.MutableInt` A mutable int wrapper
`static class`	`RLGlueAgent.StateReference` A wrapper that maintains a reference to a `State` or null.

Field Summary

Fields
Modifier and Type	Field and Description
`protected State`	`curState` The current state of the environment
`protected boolean`	`curStateIsTerminal` Whether the current state is a terminal state
`protected int`	`debugCode` Debug code used for printing debug information.
`protected double`	`discount` The RLGlue specified discount factor
`protected Domain`	`domain` The BURLAP `Domain` specifying the RLGlue problem representation and action space.
`protected RLGlueAgent.MutableInt`	`domainSet` A variable for synchronized checking if the domain has been set.
`protected double`	`lastReward` The last reward received
`protected RLGlueAgent.MutableInt`	`nextAction` Maintains the current RLGlue action to be performed.
`protected RLGlueAgent.StateReference`	`nextStateReference` Maintains the field to wait for the next state received from RLGlue
`protected boolean`	`printDebug` Whether to print debug statements.
`protected boolean`	`rlGlueExperimentFinished`

Constructor Summary

Constructors
Constructor and Description

RLGlueAgent()

Constructors
Constructor and Description
`RLGlueAgent()`

Method Summary

All Methods Static Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`void`	`agent_cleanup()`
`void`	`agent_end(double v)`
`void`	`agent_init(java.lang.String arg0)`
`java.lang.String`	`agent_message(java.lang.String s)`
`org.rlcommunity.rlglue.codec.types.Action`	`agent_start(org.rlcommunity.rlglue.codec.types.Observation observation)`
`org.rlcommunity.rlglue.codec.types.Action`	`agent_step(double v, org.rlcommunity.rlglue.codec.types.Observation observation)`
`void`	`blockUntilStateReceived()` Blocks the calling thread until a state is provided by the RLGlue server or the RLGlue experiment has ended.
`State`	`currentObservation()` Returns the current observation of the environment as a `State`.
`EnvironmentOutcome`	`executeAction(Action ga)` Executes the specified action in this environment
`int`	`getDebugCode()`
`double`	`getDiscountFactor()` Returns the discount factor for this environment.
`Domain`	`getDomain()` Returns the domain for this environment.
`static org.rlcommunity.rlglue.codec.types.Action`	`getRLGlueAction(int id)` Returns the corresponding RLGlue action for the given action id.
`boolean`	`isInTerminalState()` Returns whether the environment is in a terminal state that prevents further action by the agent.
`double`	`lastReward()` Returns the last reward returned by the environment
`void`	`loadAgent()` Loads this RLGlue `AgentInterface` into RLGlue and runs its event loop in a separate thread.
`void`	`loadAgent(java.lang.String hostAddress, java.lang.String portString)` Loads this RLGlue `AgentInterface` into RLGlue using the specified host address and port nd runs its event loop in a separate thread.
`void`	`resetEnvironment()` Resets this environment to some initial state, if the functionality exists.
`boolean`	`rlGlueExperimentFinished()` Returns true if the RLGlue experiment is finished; false otherwise.
`void`	`toggleDebug(boolean printDebug)` Toggles whether debug information should be printed

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Field Detail
  - debugCode
```
protected int debugCode
```
    Debug code used for printing debug information.
  - printDebug
```
protected boolean printDebug
```
    Whether to print debug statements.
  - nextAction
```
protected final RLGlueAgent.MutableInt nextAction
```
    Maintains the current RLGlue action to be performed.
  - nextStateReference
```
protected RLGlueAgent.StateReference nextStateReference
```
    Maintains the field to wait for the next state received from RLGlue
  - curState
```
protected State curState
```
    The current state of the environment
  - curStateIsTerminal
```
protected boolean curStateIsTerminal
```
    Whether the current state is a terminal state
  - lastReward
```
protected double lastReward
```
    The last reward received
  - domain
```
protected Domain domain
```
    The BURLAP Domain specifying the RLGlue problem representation and action space.
  - discount
```
protected double discount
```
    The RLGlue specified discount factor
  - domainSet
```
protected RLGlueAgent.MutableInt domainSet
```
    A variable for synchronized checking if the domain has been set.
  - rlGlueExperimentFinished
```
protected boolean rlGlueExperimentFinished
```
- Constructor Detail
  - RLGlueAgent
```
public RLGlueAgent()
```
- Method Detail
  - loadAgent
```
public void loadAgent()
```
    Loads this RLGlue AgentInterface into RLGlue and runs its event loop in a separate thread.
  - loadAgent
```
public void loadAgent(java.lang.String hostAddress,
                      java.lang.String portString)
```
    Loads this RLGlue AgentInterface into RLGlue using the specified host address and port nd runs its event loop in a separate thread.
    
    Parameters:
    
    hostAddress - the RLGlue host address.
    
    portString - the port on which to connect to RLGlue.
  - toggleDebug
```
public void toggleDebug(boolean printDebug)
```
    Toggles whether debug information should be printed
    
    Parameters:
    
    printDebug - whether to print debug logs or not
  - getDebugCode
```
public int getDebugCode()
```
  - getDomain
```
public Domain getDomain()
```
    Returns the domain for this environment. This method will block until the domain for this environment is set by RLGLue via the agent_init(String) method, which means you ought to have called loadAgent() or loadAgent(String, String) before calling this method, otherwise RLGlue will not know to set the environment.
    
    Returns:
    
    the BURLAP Domain specification for this RLGlue environment.
  - getDiscountFactor
```
public double getDiscountFactor()
```
    Returns the discount factor for this environment. This method will block until the domain for this environment is set by RLGLue via the agent_init(String) method, which means you ought to have called loadAgent() or loadAgent(String, String) before calling this method, otherwise RLGlue will not know to set the environment.
    
    Returns:
    
    the discount factor to use for this RLGlue problem.
  - rlGlueExperimentFinished
```
public boolean rlGlueExperimentFinished()
```
    Returns true if the RLGlue experiment is finished; false otherwise.
    
    Returns:
    
    true if the RLGlue experiment is finished; false otherwise
  - blockUntilStateReceived
```
public void blockUntilStateReceived()
```
    Blocks the calling thread until a state is provided by the RLGlue server or the RLGlue experiment has ended.
  - agent_init
```
public void agent_init(java.lang.String arg0)
```
    Specified by:
    
    agent_init in interface org.rlcommunity.rlglue.codec.AgentInterface
  - agent_start
```
public org.rlcommunity.rlglue.codec.types.Action agent_start(org.rlcommunity.rlglue.codec.types.Observation observation)
```
    Specified by:
    
    agent_start in interface org.rlcommunity.rlglue.codec.AgentInterface
  - agent_step
```
public org.rlcommunity.rlglue.codec.types.Action agent_step(double v,
                                                            org.rlcommunity.rlglue.codec.types.Observation observation)
```
    Specified by:
    
    agent_step in interface org.rlcommunity.rlglue.codec.AgentInterface
  - agent_end
```
public void agent_end(double v)
```
    Specified by:
    
    agent_end in interface org.rlcommunity.rlglue.codec.AgentInterface
  - agent_cleanup
```
public void agent_cleanup()
```
    Specified by:
    
    agent_cleanup in interface org.rlcommunity.rlglue.codec.AgentInterface
  - agent_message
```
public java.lang.String agent_message(java.lang.String s)
```
    Specified by:
    
    agent_message in interface org.rlcommunity.rlglue.codec.AgentInterface
  - currentObservation
```
public State currentObservation()
```
    Description copied from interface: Environment
    
    Returns the current observation of the environment as a State.
    
    Specified by:
    
    currentObservation in interface Environment
    
    Returns:
    
    the current observation of the environment as a State.
  - executeAction
```
public EnvironmentOutcome executeAction(Action ga)
```
    Description copied from interface: Environment
    
    Executes the specified action in this environment
    
    Specified by:
    
    executeAction in interface Environment
    
    Parameters:
    
    ga - the Action that is to be performed in this environment.
    
    Returns:
    
    the resulting observation and reward transition from applying the given GroundedAction in this environment.
  - lastReward
```
public double lastReward()
```
    Description copied from interface: Environment
    
    Returns the last reward returned by the environment
    
    Specified by:
    
    lastReward in interface Environment
    
    Returns:
    
    the last reward returned by the environment
  - isInTerminalState
```
public boolean isInTerminalState()
```
    Description copied from interface: Environment
    
    Returns whether the environment is in a terminal state that prevents further action by the agent.
    
    Specified by:
    
    isInTerminalState in interface Environment
    
    Returns:
    
    true if the current environment is in a terminal state; false otherwise.
  - resetEnvironment
```
public void resetEnvironment()
```
    Description copied from interface: Environment
    
    Resets this environment to some initial state, if the functionality exists.
    
    Specified by:
    
    resetEnvironment in interface Environment
  - getRLGlueAction
```
public static org.rlcommunity.rlglue.codec.types.Action getRLGlueAction(int id)
```
    Returns the corresponding RLGlue action for the given action id.
    
    Parameters:
    
    id - the action id
    
    Returns:
    
    An RLGlue action for the corresponding aciton id.

Class RLGlueAgent

Nested Class Summary

Field Summary

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Field Detail

debugCode

printDebug

nextAction

nextStateReference

curState

curStateIsTerminal

lastReward

domain

discount

domainSet

rlGlueExperimentFinished

Constructor Detail

RLGlueAgent

Method Detail

loadAgent

loadAgent

toggleDebug

getDebugCode

getDomain

getDiscountFactor

rlGlueExperimentFinished

blockUntilStateReceived

agent_init

agent_start

agent_step

agent_end

agent_cleanup

agent_message

currentObservation

executeAction

lastReward

isInTerminalState

resetEnvironment

getRLGlueAction