All Packages  Class Hierarchy  This Package  Previous  Next  Index

Class sim.gradDesc.IDD

java.lang.Object
   |
   +----sim.Experiment
           |
           +----sim.gradDesc.GradDesc
                   |
                   +----sim.gradDesc.IDD

public class IDD
extends GradDesc
Incremental Delta Delta (reference "Multi-Agent Residual Advantage Learning With General Function Approximation" on publication list at http://www-anw.cs.umass.edu/~mharmon).

This code is (c) 1997 Mance Harmon <mharmon@acm.org>, http://www-anw.cs.umass.edu/~mharmon
The source and object code may be redistributed freely. If the code is modified, please state so in the comments.

Version:
0.9, 22 July 97 This code will become 1.0 when autoparse is done.
Author:
Mance Harmon

Variable Index

 o alphas
the learning rates, alpha = e^beta(i)
 o betas
the beta matrix
 o errFun
the error function to minimize
 o error
a noisy estimate of the error being gradient descended on
 o logSmoothedError
log base 10 of the smoothed error
 o pAlphas
a pointer to alphas
 o pBetas
a pointer to the betas
 o rndCopy
a copy of the generator passed to evaluate().
 o smoothedError
an exponentially smoothed estimate of the error
 o smoothingFactor
the constant used to smooth the error (near 1 = long halflife)
 o theta
the learning rate, a small positive number
 o time
current time (time increments once per weight change
 o tolerance
stop learning when smoothed error < tolerance
 o weightChange1
the change in the weights at time t
 o weightChange2
the change in the weights at time t+1

Constructor Index

 o IDD()

Method Index

 o BNF(int)
Return a parameter array if BNF(), parse(), and unparse() are to be automated, null otherwise.
 o initialize(int)
Initialize, either partially or completely.
 o parse(Parser, int)
Parse the input file to get the parameters for this object.
 o run()
repeatedly change x until f(x) reaches a local minimum
 o setWatchManager(WatchManager, String)
Register all variables with this WatchManager.
 o unparse(Unparser, int)
Output a description of this object that can be parsed with parse().

Variables

 o errFun
 protected ErrFun errFun
the error function to minimize

 o theta
 protected NumExp theta
the learning rate, a small positive number

 o alphas
 protected MatrixD alphas
the learning rates, alpha = e^beta(i)

 o pAlphas
 protected PMatrixD pAlphas
a pointer to alphas

 o betas
 protected MatrixD betas
the beta matrix

 o pBetas
 protected PMatrixD pBetas
a pointer to the betas

 o weightChange1
 protected MatrixD weightChange1
the change in the weights at time t

 o weightChange2
 protected MatrixD weightChange2
the change in the weights at time t+1

 o error
 protected PDouble error
a noisy estimate of the error being gradient descended on

 o smoothedError
 protected PDouble smoothedError
an exponentially smoothed estimate of the error

 o smoothingFactor
 protected NumExp smoothingFactor
the constant used to smooth the error (near 1 = long halflife)

 o tolerance
 protected NumExp tolerance
stop learning when smoothed error < tolerance

 o logSmoothedError
 protected PDouble logSmoothedError
log base 10 of the smoothed error

 o time
 protected PInt time
current time (time increments once per weight change

 o rndCopy
 protected Random rndCopy
a copy of the generator passed to evaluate(). Is used to generate the same inputs when evaluate is called a second time.

Constructors

 o IDD
 public IDD()

Methods

 o setWatchManager
 public void setWatchManager(WatchManager wm,
                             String name)
Register all variables with this WatchManager. This will be called after all parsing is done.

Overrides:
setWatchManager in class GradDesc
 o run
 public void run()
repeatedly change x until f(x) reaches a local minimum

Overrides:
run in class GradDesc
 o BNF
 public String BNF(int lang)
Return a parameter array if BNF(), parse(), and unparse() are to be automated, null otherwise.

Overrides:
BNF in class GradDesc
See Also:
getParameters public
 o unparse
 public void unparse(Unparser u,
                     int lang)
Output a description of this object that can be parsed with parse().

Overrides:
unparse in class GradDesc
See Also:
Parsable
 o parse
 public Object parse(Parser p,
                     int lang) throws ParserException
Parse the input file to get the parameters for this object.

Throws: ParserException
parser didn't find the required token
Overrides:
parse in class GradDesc
 o initialize
 public void initialize(int level)
Initialize, either partially or completely.

Overrides:
initialize in class Experiment
See Also:
initialize

All Packages  Class Hierarchy  This Package  Previous  Next  Index