Navigation

Global	Documentation
Home	Manual
Readme	Glossary
Notes	Design
Examples	Classes
	Trouble
	Style

class NearEngine( OrderEngine )	Source

The NearEngine searches the nearest neighbour (NN) for a point and goes there first.

This is NOT a random engine as it mostly(??) steps uphill.

It belongs to the class of generalized travelling salesman problems where the parameters of the problem is an ordered list.

The walker is kept when the logLikelihood > lowLhood

Attributes from Engine

walkers, errdis, maxtrials, nstep, slow, rng, report, phantoms, verbose

Author Do Kester.

NearEngine( walkers, errdis, copy=None, **kwargs )

Constructor.

Parameters

• walkers : SampleList
       walkers to be diffused
• errdis : ErrorDistribution
       error distribution to be used
• copy : NearEngine
       to be copied
• kwargs : dict for Engine
       "phantoms", "slow", "seed", "verbose"

copy( )

Return copy of this.

executeOnce( kw, lowLhood, dims=[0,1] )

Execute the NearEngine one time.

TBC: This engine seems to take a lot of CPU.

Parameters

• kw : int
       id of walker to diffuse
• lowLhood : float
       lower limit in logLikelihood
• dims : list of 2 ints
       dimensions to process over

Methods inherited from OrderEngine

• execute( kw, lowLhood, append=False, iteration=0 )
• calculateUnitRange( )

Methods inherited from Engine

• bestBoost( problem, myFitter=None )
• setWalker( kw, problem, allpars, logL, walker=None, fitIndex=None )
• noBoost( walker )
• doBoost( walker )
• domain2Unit( problem, dval, kpar=None )
• unit2Domain( problem, uval, kpar=None )
• startJourney( unitStart )
• calcJourney( unitDistance )
• reportJourney( )
• makeIndex( np, val )
• reportCall( )
• reportSuccess( )
• reportReject( )
• reportFailed( )
• reportBest( )
• printReport( best=False )
• successRate( )
• getUnitMinmax( problem, lowLhood )
• getUnitRange( problem, lowLhood )