DistanceCostFunction.md

---

---

class DistanceCostFunction( ErrorDistribution )	[source]

To calculate a distance based cost function

For one observation with n counts it holds

     f( d ) = exp( -SUM( d / s ) )

where d are the distances and s is the scale

The function is mostly used to calculate the likelihood L of traveling-salesman-like problems

Author Do Kester.

DistanceCostFunction( copy=None )	[source]

Constructor.

Parameters

• copy : DistanceCostFunction
       distribution to be copied.

copy( )	[source]

Return copy of this.

acceptWeight( )	[source]

True if the distribution accepts weights. Always false for this distribution.

logLikelihood_alt( problem, allpars )	[source]

Return the negative sum of the distances.

Parameters

• problem : Problem
       to be solved
• allpars : array_like
       list of all parameters in the problem

logLdata( problem, allpars )	[source]

Return the individual distances (multiplied by the weights).

Parameters

• problem : Problem
       to be solved
• allpars : array_like
       list of all parameters in the problem

partialLogL( model, param, fitIndex )	[source]

Does not work for this class

Parameters

• model : Model
       model to calculate mock data
• param : array_like
       parameters of the model
• fitIndex : array_like
       indices of the params to be fitted

Methods inherited from ErrorDistribution

• getGaussianScale( problem, allpars=None )
• getResiduals( problem, allpars=None )
• getChisq( problem, allpars=None )
• toSigma( scale )
• isBound( )
• keepFixed( fixed=None )
• setPriors( priors )
• setLimits( limits )
• domain2Unit( dval, ks )
• unit2Domain( uval, ks )
• logCLhood( problem, allpars )
• logLhood( problem, allpars )
• partialLogL_alt( problem, allpars, fitIndex )
• numPartialLogL( problem, allpars, fitIndex )
• updateLogL( problem, allpars, parval=None )
• hyparname( k )