clarion.mcs
Class MCS

java.lang.Object
  extended byclarion.mcs.MCS

public class MCS
extends java.lang.Object


Field Summary
protected  short[][][] acsInputDims
          the info of the filtered ACS input dimensions.
protected  GeneralNet[][] acsInputFilterNets
          the BP nets for ACS filters.
protected  short[][] acsLearnMethods
          the ACS learning methods.
protected  GeneralNet[][] acsLearnNets
          the BP nets for ACS learning.
protected  int acsNetNum
          the number of ACS action-decision networks.
protected  short[][][] acsOutputDims
          the info of the filtered ACS output dimensions.
protected  GeneralNet[][] acsOutputFilterNets
          the BP nets for ACS filters.
protected  short[][] acsReasonMethods
          the ACS reasoning methods.
protected  GeneralNet[][] acsReasonNets
          the BP nets for ACS reasoning.
protected  double[] buffer
          monitor buffer input.
protected  StochasticDecider decider
          the action decider.
protected  double[] drives
          drive input.
protected  GeneralNet evalNet
          the BP net for evaluation.
protected  Global global
          global CLARION settings.
protected  double[] goal
          goal input.
protected  int goalAction
          the index of goal action.
protected  GeneralNet gsNet
          the BP net for goal action.
protected  double[] input
          overall input.
protected  MonitorBuf monitorBuf
          the monitor buffer.
protected  short[][] nacsInputDims
          the info of the filtered NACS input dimensions.
protected  GeneralNet[] nacsInputFilterNets
          the BP nets for NACS filters.
protected  short[] nacsLearnMethods
          the NACS learning methods.
protected  GeneralNet[] nacsLearnNets
          the BP nets for NACS learning.
protected  short[][] nacsOutputDims
          the info of the filtered NACS output dimensions.
protected  GeneralNet[] nacsOutputFilterNets
          the BP nets for NACS filters.
protected  short[] nacsReasonMethods
          the NACS reasoning methods.
protected  GeneralNet[] nacsReasonNets
          the BP nets for NACS reasoning.
protected  short[] offsets
           
protected  double[] output
           
protected  double reinforcement
          the reinforcement.
protected  double[] state
          sensory input.
protected  double[] wm
          working memory input.
 
Constructor Summary
MCS(ACS acs, NACS nacs, Drives drives, Global global)
          constructor.
 
Method Summary
 short[] getAcsInputDims(int netIdx, int compIdx)
          Returns the info of the filtered ACS input dimensions with the specific network and the component in that network.
 int getAcsLearnMethod(int netIdx, int compIdx)
          Returns the learning method of the specific network and the component in that ACS network.
 short[] getAcsOutputDims(int netIdx, int compIdx)
          Returns the info of the filtered ACS output dimensions with the specific network and the component in that network.
 int getAcsReasonMethod(int netIdx, int compIdx)
          Returns the reasoning method of the specific network and the component in that ACS network.
 int getGoalAction()
          Returns the goal action.
 MonitorBuf getMonitorBuffer()
          Returns the monitor buffer.
 short[] getNacsInputDims(int compIdx)
          Returns the info of the filtered NACS input dimensions with the specific component.
 int getNacsLearnMethod(int compIdx)
          Returns the learning method of the specific component in NACS subsystem.
 short[] getNacsOutputDims(int compIdx)
          Returns the info of the filtered NACS output dimensions with the specific component.
 int getNacsReasonMethod(int compIdx)
          Returns the reasoning method of the specific component in NACS subsystem.
protected  void getOutputs(GeneralNet net)
          Computes the outputs from a specific BP decision network.
 double getReinforcement()
          Returns the reinforcement.
 void setBuffer(double[] arr)
          Sets the monitor buffer input.
 void setDrvies(double[] arr)
          Sets the drive input.
 void setGoal(double[] arr)
          Sets the goal input.
 void setState(short[] arr)
          Sets the sensory input.
 void setWM(double[] arr)
          Sets the working memory input.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

global

protected Global global
global CLARION settings.


acsNetNum

protected int acsNetNum
the number of ACS action-decision networks.


goalAction

protected int goalAction
the index of goal action.


reinforcement

protected double reinforcement
the reinforcement.


acsInputDims

protected short[][][] acsInputDims
the info of the filtered ACS input dimensions.


acsOutputDims

protected short[][][] acsOutputDims
the info of the filtered ACS output dimensions.


nacsInputDims

protected short[][] nacsInputDims
the info of the filtered NACS input dimensions.


nacsOutputDims

protected short[][] nacsOutputDims
the info of the filtered NACS output dimensions.


acsReasonMethods

protected short[][] acsReasonMethods
the ACS reasoning methods.


nacsReasonMethods

protected short[] nacsReasonMethods
the NACS reasoning methods.


acsLearnMethods

protected short[][] acsLearnMethods
the ACS learning methods.


nacsLearnMethods

protected short[] nacsLearnMethods
the NACS learning methods.


offsets

protected short[] offsets

input

protected double[] input
overall input.


output

protected double[] output

state

protected double[] state
sensory input.


goal

protected double[] goal
goal input.


wm

protected double[] wm
working memory input.


drives

protected double[] drives
drive input.


buffer

protected double[] buffer
monitor buffer input.


evalNet

protected GeneralNet evalNet
the BP net for evaluation.


gsNet

protected GeneralNet gsNet
the BP net for goal action.


acsInputFilterNets

protected GeneralNet[][] acsInputFilterNets
the BP nets for ACS filters.


acsOutputFilterNets

protected GeneralNet[][] acsOutputFilterNets
the BP nets for ACS filters.


nacsInputFilterNets

protected GeneralNet[] nacsInputFilterNets
the BP nets for NACS filters.


nacsOutputFilterNets

protected GeneralNet[] nacsOutputFilterNets
the BP nets for NACS filters.


acsReasonNets

protected GeneralNet[][] acsReasonNets
the BP nets for ACS reasoning.


nacsReasonNets

protected GeneralNet[] nacsReasonNets
the BP nets for NACS reasoning.


acsLearnNets

protected GeneralNet[][] acsLearnNets
the BP nets for ACS learning.


nacsLearnNets

protected GeneralNet[] nacsLearnNets
the BP nets for NACS learning.


monitorBuf

protected MonitorBuf monitorBuf
the monitor buffer.


decider

protected StochasticDecider decider
the action decider.

Constructor Detail

MCS

public MCS(ACS acs,
           NACS nacs,
           Drives drives,
           Global global)
constructor.

Parameters:
acs - the ACS subsystem.
nacs - the NACS subsystem.
drives - the used drives.
global - the global settings.
Method Detail

setState

public void setState(short[] arr)
Sets the sensory input.

Parameters:
arr - the sensory input to set.

setGoal

public void setGoal(double[] arr)
Sets the goal input.

Parameters:
arr - the goal input to set.

setWM

public void setWM(double[] arr)
Sets the working memory input.

Parameters:
arr - the working input to set.

setDrvies

public void setDrvies(double[] arr)
Sets the drive input.

Parameters:
arr - the drive input to set.

setBuffer

public void setBuffer(double[] arr)
Sets the monitor buffer input.

Parameters:
arr - the monitor buffer input to set.

getMonitorBuffer

public MonitorBuf getMonitorBuffer()
Returns the monitor buffer.

Returns:
the monitor buffer.

getReinforcement

public double getReinforcement()
Returns the reinforcement.

Returns:
the reinforcement.

getGoalAction

public int getGoalAction()
Returns the goal action.

Returns:
the goal action.

getAcsInputDims

public short[] getAcsInputDims(int netIdx,
                               int compIdx)
Returns the info of the filtered ACS input dimensions with the specific network and the component in that network.

Parameters:
netIdx - the ACS network index.
compIdx - the component index.
Returns:
the info of the filtered ACS input dimensions.

getAcsOutputDims

public short[] getAcsOutputDims(int netIdx,
                                int compIdx)
Returns the info of the filtered ACS output dimensions with the specific network and the component in that network.

Parameters:
netIdx - the ACS network index.
compIdx - the component index.
Returns:
the info of the filtered ACS output dimensions.

getNacsInputDims

public short[] getNacsInputDims(int compIdx)
Returns the info of the filtered NACS input dimensions with the specific component.

Parameters:
compIdx - the component index.
Returns:
the info of the filtered NACS input dimensions.

getNacsOutputDims

public short[] getNacsOutputDims(int compIdx)
Returns the info of the filtered NACS output dimensions with the specific component.

Parameters:
compIdx - the component index.
Returns:
the info of the filtered NACS output dimensions.

getAcsReasonMethod

public int getAcsReasonMethod(int netIdx,
                              int compIdx)
Returns the reasoning method of the specific network and the component in that ACS network.

Parameters:
netIdx - the ACS network index.
compIdx - the component index.
Returns:
the reasoning method.

getNacsReasonMethod

public int getNacsReasonMethod(int compIdx)
Returns the reasoning method of the specific component in NACS subsystem.

Parameters:
compIdx - the component index.
Returns:
the reasoning method.

getAcsLearnMethod

public int getAcsLearnMethod(int netIdx,
                             int compIdx)
Returns the learning method of the specific network and the component in that ACS network.

Parameters:
netIdx - the ACS network index.
compIdx - the component index.
Returns:
the learning method.

getNacsLearnMethod

public int getNacsLearnMethod(int compIdx)
Returns the learning method of the specific component in NACS subsystem.

Parameters:
compIdx - the component index.
Returns:
the learning method.

getOutputs

protected void getOutputs(GeneralNet net)
Computes the outputs from a specific BP decision network.

Parameters:
net - the specific decision network.