All Packages  Class Hierarchy  This Package  Previous  Next  Index

Class rsf.operator.ElwInv

java.lang.Object
   |
   +----rsf.operator.ElwInv

public class ElwInv

extends Object

implements Operator

1/x[i]
.
 If 
x[1]
 is close to zero, I replace it by 
eps
.




  


  
 
	ElwInv(RsfSpace)
  
 
 eps = java.lang.Float.MIN_VALUE

  
 
	ElwInv(RsfSpace, Float)
  
 
  
 
	ElwInv(RsfSpace, float)
  
 
 Any element 
e
 with 
Abs(e) < eps
 
 is replaced by 
eps
 or 
-eps
 
 as appropiate.


  


  
 
	addImage(Vector, Vector)
  
 
  
 
	copy()
  
 
  
 
	elwInv(Vector)
  
 
  
 
	elwInv(Vector, float)
  
 
  
 
	elwInv(Vector, float, float)
  
  In general this should not be invoked.
  
 
	elwInvDivByZeroIsZero(Vector)
  
 
  
 
	getDomain()
  
 
  
 
	getRange()
  
 
  
 
	image(Vector)
  
 
  
 
	image(Vector, Vector)
  
 
  
 
	main(String[])
  
 
  
 
	residual(Vector, Vector, Vector)
  
 



  



ElwInv
 public ElwInv(RsfSpace spc,
               float eps)


  
 Any element 
e
 with 
Abs(e) < eps
 
 is replaced by 
eps
 or 
-eps
 
 as appropiate. This prevents a division by zero.  
 A practical eps value might be 1.0e-12f.
 No catching of NaN or INFINITE numbers yet



ElwInv
 public ElwInv(RsfSpace spc,
               Float eps)


ElwInv
 public ElwInv(RsfSpace spc)


  
 eps = java.lang.Float.MIN_VALUE




  


copy
 public JamObject copy()


image
 public Vector image(Vector d)


image
 public void image(Vector d,
                   Vector r)


addImage
 public void addImage(Vector d,
                      Vector r)


residual
 public void residual(Vector d,
                      Vector b,
                      Vector r)


getDomain
 public Space getDomain()


getRange
 public Space getRange()


main
 public static void main(String args[])


elwInv
 public static void elwInv(Vector data)


elwInvDivByZeroIsZero
 public static void elwInvDivByZeroIsZero(Vector data)


elwInv
 public static void elwInv(Vector data,
                           float eps)


elwInv
 public static void elwInv(Vector data,
                           float eps,
                           float upper)


  
 In general this should not be invoked. Maybe I should make it private. 
 A user may set eps and upper to nonsensical values.


  

    
 Parameters:
    
 : - eps:  absolute value of the lower limit of floats inverted
               correctly.
    
 upper: - absolute value assigned to the inverse off input values
               that are equal or smaller than eps.
  


All Packages  Class Hierarchy  This Package  Previous  Next  Index