Package pal.math
Interface OrthogonalHints
-
- All Known Implementing Classes:
MutationRateModelTree
public interface OrthogonalHintsProvides a means for giving an Orthogonal base optimiser (IE, OrthognalMinimum) hints about the function that may alow it to optimise better.- Version:
- $Id: OrthogonalHints.java,v 1.2 2002/04/21 00:13:31 matt Exp $
- Author:
- Matthew Goode
-
-
Nested Class Summary
Nested Classes Modifier and Type Interface Description static classOrthogonalHints.Utils
-
Method Summary
All Methods Instance Methods Abstract Methods Modifier and Type Method Description intgetInternalParameterBoundaries(int parameter, double[] storage)A boundary is a value of a parameter for which values lower than the boundary and values higher than the boundary are better treated as two separate functions (IE, they are only piecewise connected), and minimisation should be performed over both ranges individually (and then the true minimum taken as the minimuma of the ranges)OrderEnumeratorgetSuggestedOrdering(OrderEnumerator defaultOrdering)If there is a "best" ordering to use it can be specified here, if not should return null
-
-
-
Method Detail
-
getSuggestedOrdering
OrderEnumerator getSuggestedOrdering(OrderEnumerator defaultOrdering)
If there is a "best" ordering to use it can be specified here, if not should return null- Parameters:
defaultOrdering- The ordering suggested by the optimiser, may be null!- Returns:
- null, or default ordering if no known best ordering
-
getInternalParameterBoundaries
int getInternalParameterBoundaries(int parameter, double[] storage)A boundary is a value of a parameter for which values lower than the boundary and values higher than the boundary are better treated as two separate functions (IE, they are only piecewise connected), and minimisation should be performed over both ranges individually (and then the true minimum taken as the minimuma of the ranges)- Returns:
- the number of boundary locations stored in storage, or -1 if not enough room, or 0 if there are no boundaries (other than the normal parameter range)
-
-