Base object for Monte-Carlo Global Optimization methods. More...

Inheritance diagram for ObjCryst::MonteCarloObj:

Collaboration diagram for ObjCryst::MonteCarloObj:

Public Member Functions
	MonteCarloObj (const string name="")
	Constructor.

	MonteCarloObj (const bool internalUseOnly)
	Constructor. More...

virtual	~MonteCarloObj ()
	Destructor.

void	SetAlgorithmSimulAnnealing (const AnnealingSchedule scheduleTemp, const REAL tMax, const REAL tMin, const AnnealingSchedule scheduleMutation=ANNEALING_CONSTANT, const REAL mutMax=16., const REAL mutMin=.125, const long nbTrialRetry=0, const REAL minCostRetry=0.)
	Set the refinement method to simulated Annealing. More...

void	SetAlgorithmParallTempering (const AnnealingSchedule scheduleTemp, const REAL tMax, const REAL tMin, const AnnealingSchedule scheduleMutation=ANNEALING_CONSTANT, const REAL mutMax=16., const REAL mutMin=.125)
	Set the refinement method to Parallel Tempering. More...

virtual void	Optimize (long &nbSteps, const bool silent=false, const REAL finalcost=0, const REAL maxTime=-1)
	Launch optimization (a single run) for N steps. More...

virtual void	MultiRunOptimize (long &nbCycle, long &nbSteps, const bool silent=false, const REAL finalcost=0, const REAL maxTime=-1)
	Launch optimization for multiple runs of N steps. More...

void	RunSimulatedAnnealing (long &nbSteps, const bool silent=false, const REAL finalcost=0, const REAL maxTime=-1)

void	RunParallelTempering (long &nbSteps, const bool silent=false, const REAL finalcost=0, const REAL maxTime=-1)

void	RunRandomLSQMethod (long &nbCycle)

virtual void	XMLOutput (ostream &os, int indent=0) const
	Output a description of the object in XML format to a stream. More...

virtual void	XMLInput (istream &is, const XMLCrystTag &tag)
	Input in XML format from a stream, restoring the set of refined objects and the associated cost functions. More...

virtual const string	GetClassName () const
	Get the name for this class type.

LSQNumObj &	GetLSQObj ()
	Access to the builtin LSQ optimization object.

const LSQNumObj &	GetLSQObj () const
	Access to the builtin LSQ optimization object.

virtual void	InitLSQ (const bool useFullPowderPatternProfile=true)
	Prepare mLSQ for least-squares refinement during the global optimization. More...

Public Member Functions inherited from ObjCryst::OptimizationObj
	OptimizationObj (const string name="")
	Constructor.

virtual	~OptimizationObj ()
	Destructor.

virtual void	RandomizeStartingConfig ()
	Randomize starting configuration. More...

void	FixAllPar ()
	Fix all parameters.

void	SetParIsFixed (const string &parName, const bool fix)
	Fix one parameter.

void	SetParIsFixed (const RefParType *type, const bool fix)
	Fix one family of parameters.

void	UnFixAllPar ()
	UnFix All parameters.

void	SetParIsUsed (const string &parName, const bool use)
	Set a parameter to be used.

void	SetParIsUsed (const RefParType *type, const bool use)
	Set a family of parameters to be used.

void	SetLimitsRelative (const string &parName, const REAL min, const REAL max)
	Change the relative limits for a parameter from its name.

void	SetLimitsRelative (const RefParType *type, const REAL min, const REAL max)
	Change the relative limits for a family of parameter.

void	SetLimitsAbsolute (const string &parName, const REAL min, const REAL max)
	Change the absolute limits for a parameter from its name.

void	SetLimitsAbsolute (const RefParType *type, const REAL min, const REAL max)
	Change the absolute limits for a family of parameter.

virtual REAL	GetLogLikelihood () const
	The optimized (minimized, actually) function. More...

void	StopAfterCycle ()
	Stop after the current cycle. USed for interactive refinement.

virtual void	DisplayReport ()
	Show report to the user during refinement. Used for GUI update.

void	AddRefinableObj (RefinableObj &)
	Add a refined object. All sub-objects are also added.

RefinableObj &	GetFullRefinableObj (const bool rebuild=true)
	Get the RefinableObj with all the parameters from all refined objects. More...

const string &	GetName () const
	Get the name for this object.

void	SetName (const string &)
	Set the name for this object.

virtual void	Print () const
	Print some information about this object.

void	RestoreBestConfiguration ()
	Restore the Best configuration.

bool	IsOptimizing () const
	Are we busy optimizing ?

void	TagNewBestConfig ()
	During a global optimization, tell all objects that the current config is the latest "best" config.

REAL	GetLastOptimElapsedTime () const
	Get the elapsed time (in seconds) during the last optimization.

MainTracker &	GetMainTracker ()
	Get the MainTracker.

const MainTracker &	GetMainTracker () const
	Get the MainTracker.

RefObjOpt &	GetXMLAutoSaveOption ()

const RefObjOpt &	GetXMLAutoSaveOption () const

const REAL &	GetBestCost () const
	Access to current best cost.

REAL &	GetBestCost ()
	Access to current best cost.

virtual void	BeginOptimization (const bool allowApproximations=false, const bool enableRestraints=false)
	Begin optimization for all objects.

virtual void	EndOptimization ()
	End optimization for all objects.

virtual long &	NbTrialPerRun ()
	Number of trial per run.

virtual const long &	NbTrialPerRun () const
	Number of trial per run.

Protected Member Functions
virtual void	NewConfiguration (const RefParType *type=gpRefParTypeObjCryst)
	Make a random change in the configuration. More...

virtual void	InitOptions ()
	Initialization of options.

Protected Member Functions inherited from ObjCryst::OptimizationObj
void	PrepareRefParList ()

virtual void	UpdateDisplay ()
	Update Display (if any display is available), when a new 'relevant' configuration is reached. More...

void	BuildRecursiveRefObjList ()
	(Re)build OptimizationObj::mRecursiveRefinedObjList, if an object has been added or modified. More...

Protected Attributes
RefObjOpt	mGlobalOptimType
	Method used for the global optimization. More...

REAL	mCurrentCost
	Current value of the cost function.

RefObjOpt	mSaveTrackedData
	Option to save the evolution of tracked data (cost functions, likelihhod, individual parameters,...)

REAL	mTemperature
	Current temperature for annealing.

REAL	mTemperatureMax
	Beginning temperature for annealing.

REAL	mTemperatureMin
	Lower temperature.

RefObjOpt	mAnnealingScheduleTemp
	Schedule for the annealing.

REAL	mTemperatureGamma
	Gamma for the 'gamma' temperature schedule.

REAL	mMutationAmplitude
	Mutation amplitude. More...

REAL	mMutationAmplitudeMax
	Mutation amplitude at the beginning of the optimization.

REAL	mMutationAmplitudeMin
	Mutation amplitude at the end of the optimization.

RefObjOpt	mAnnealingScheduleMutation
	Schedule for the annealing.

REAL	mMutationAmplitudeGamma
	Gamma for the 'gamma' Mutation amplitude schedule.

long	mNbTrialRetry
	Number of trials before testing if we are below the given minimum cost. More...

REAL	mMinCostRetry
	Cost to reach unless an automatic randomization and retry is done.

LSQNumObj	mLSQ
	Least squares object.

RefObjOpt	mAutoLSQ
	Option to run automatic least-squares refinements.

Protected Attributes inherited from ObjCryst::OptimizationObj
RefinableObj	mRefParList
	The refinable par list used during refinement. More...

string	mName
	Name of the GlobalOptimization object.

string	mSaveFileName
	File name where refinement info is saved (NOT USED so far...)

long	mNbTrialPerRun
	Number of trial per run, to be saved/restored in XML output.

long	mNbTrial
	Number of trials so far.

REAL	mBestCost
	Best value of the cost function so far.

long	mBestParSavedSetIndex
	Index of the 'best' saved parameter set.

unsigned long	mContext
	The current 'context', in the case the optimization is run in different parallel contexts.

map< unsigned long, map< const RefinableObj *, LogLikelihoodStats > >	mvContextObjStats
	Statistics for each context (mutable for dynamic update during optimization)

map< const RefinableObj *, DynamicObjWeight >	mvObjWeight
	Weights for each objects in each context (mutable for dynamic update during optimization)

std::vector< pair< long, REAL > >	mvSavedParamSet
	List of saved parameter sets. More...

bool	mIsOptimizing
	True if a refinement is being done. For multi-threaded environment.

bool	mStopAfterCycle
	If true, then stop at the end of the cycle. Used in multi-threaded environment.

ObjRegistry< RefinableObj >	mRefinedObjList
	The refined objects.

ObjRegistry< RefinableObj >	mRecursiveRefinedObjList
	The refined objects, recursively including all sub-objects. More...

RefObjOpt	mXMLAutoSave
	Periodic save of complete environment as an xml file.

REAL	mLastOptimTime
	The time elapsed after the last optimization, in seconds.

MainTracker	mMainTracker
	MainTracker object to track the evolution of cost functions, likelihood, and individual parameters. More...

Detailed Description

Base object for Monte-Carlo Global Optimization methods.

The algorithm is quite simple, whith two type of optimizations, either simulated Annealing or Parallel Tempering, the latter being recommanded for most real-world optimizations.

Definition at line 336 of file GlobalOptimObj.h.

Constructor & Destructor Documentation

ObjCryst::MonteCarloObj::MonteCarloObj ( const bool internalUseOnly )

Constructor.

Using internalUseOnly=true will avoid registering the the object to any registry, and thus (for example) no display will be created, nor will this object be automatically be saved.

Definition at line 433 of file GlobalOptimObj.cpp.

Member Function Documentation

void ObjCryst::MonteCarloObj::InitLSQ ( const bool useFullPowderPatternProfile = true )

virtual

Prepare mLSQ for least-squares refinement during the global optimization.

Parameters

useFullPowderPatternProfile if true, the refinement will use the full profile version of powder patterns, otherwise only the integrated powder pattern will be used (faster).

Definition at line 2089 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::MultiRunOptimize	(	long &	nbCycle,
		long &	nbSteps,
		const bool	silent = `false`,
		const REAL	finalcost = `0`,
		const REAL	maxTime = `-1`
	)

virtual

Launch optimization for multiple runs of N steps.

Parameters

nbCycle	the number of runs (cycles) to perform. The structure is randomized at the beginning of each cycle. If nbCycle==-1, this will run indefinitely. The nbCycle parameter is decreased after each run.
nbSteps	the number of steps to go. This number is modified (decreases!) as the refinement goes on.
silent	: if true, absolutely no message should be printed (except debugging)
finalcost	the optimization will stop if overall cost fallse below this value
maxTime	the optimization will stop after the given number of seconds has been spent optimizing (ignored if <0).

Implements ObjCryst::OptimizationObj.

Definition at line 576 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::NewConfiguration ( const RefParType * type = gpRefParTypeObjCryst )

protectedvirtual

Make a random change in the configuration.

This just generates a new configuration with random changes (according to current parameters). The new configuration is not tested in this function vs temperature: this should be done in the OptimizationObj::Optimize() function, which also chooses whether to revert to the previous configuration.

Random moves are made by the objects and not by this function, because the new configuration can be specific (like, for example, permutations between some of the parameters (atoms)).

Parameters

type	can be used to restrict the move to a given category of parameters.

Definition at line 2024 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::Optimize	(	long &	nbSteps,
		const bool	silent = `false`,
		const REAL	finalcost = `0`,
		const REAL	maxTime = `-1`
	)

virtual

Launch optimization (a single run) for N steps.

Parameters

nbSteps	the number of steps to go. This number is modified (decreases!) as the refinement goes on.
silent	: if true, absolutely no message should be printed (except debugging)
finalcost	the optimization will stop if overall cost fallse below this value
maxTime	the optimization will stop after the given number of seconds has been spent optimizing (ignored if <0).

Implements ObjCryst::OptimizationObj.

Definition at line 499 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::RunParallelTempering	(	long &	nbSteps,
		const bool	silent = `false`,
		const REAL	finalcost = `0`,
		const REAL	maxTime = `-1`
	)

Do a single Parallel Tempering run. This is called by Optimize(...) and MultiRunOptimize(), which must also prepare the optimization (PrepareRefParList(), etc..).

Definition at line 1147 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::RunSimulatedAnnealing	(	long &	nbSteps,
		const bool	silent = `false`,
		const REAL	finalcost = `0`,
		const REAL	maxTime = `-1`
	)

Do a single simulated annealing run. This is called by Optimize(...) and MultiRunOptimize(), which must also prepare the optimization (PrepareRefParList(), etc..).

Definition at line 718 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::SetAlgorithmParallTempering	(	const AnnealingSchedule	scheduleTemp,
		const REAL	tMax,
		const REAL	tMin,
		const AnnealingSchedule	scheduleMutation = `ANNEALING_CONSTANT`,
		const REAL	mutMax = `16.`,
		const REAL	mutMin = `.125`
	)

Set the refinement method to Parallel Tempering.

The refinement begins at max and finishes at min temperature.

Parameters

scheduleTemp	temperature schedule
tMax,tMin	Max and Min temperatures. See AnnealingSchedule.
scheduleMutation	the mutation schedule. For each new configuration, the variation of each refinable parameter is less than its RefinablePar::GlobalOptimStep(), multiplied by the current mutation amplitude. By default this mutation is equal to 1., but making bigger steps can be a good idea at the beginning of the refinement. Thus you can choose a schedule for the amplitude, exactly like for the temperature.
mutMax,mutMin	Max and Min mutation amplitudes.

Warning: do not use the 'smart' option for the temperature schedule, it is not yet implemented. Later it will be used to set the temperatures as a function of the amplitude schedule, so that we keep accepted move between 30% and 70%.

Note: this will be removed when we separate the different algorithms in different classes.

Definition at line 481 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::SetAlgorithmSimulAnnealing	(	const AnnealingSchedule	scheduleTemp,
		const REAL	tMax,
		const REAL	tMin,
		const AnnealingSchedule	scheduleMutation = `ANNEALING_CONSTANT`,
		const REAL	mutMax = `16.`,
		const REAL	mutMin = `.125`,
		const long	nbTrialRetry = `0`,
		const REAL	minCostRetry = `0.`
	)

Set the refinement method to simulated Annealing.

Note that Parellel Tempering is more efficient to get out of local minima, so you sould rather use that method.

The refinement begins at max and finishes at min temperature.

Parameters

scheduleTemp	temperature schedule. See AnnealingSchedule.
tMax,tMin	Max and Min temperatures.
scheduleMutation	the mutation schedule. For each new configuration, the variation of each refinable parameter is less than its RefinablePar::GlobalOptimStep(), multiplied by the current mutation amplitude. By default this mutation is equal to 1., but making bigger steps is a good idea at the beginning of the refinement (for higher temperatures). See AnnealingSchedule. See AnnealingSchedule.
mutMax,mutMin	Max and Min mutation amplitudes.
minCostRetry,nbTrialRetry	if after nbTrialRetry, the cost function is still above minCostRetry, then start again from a random configuration. No randomization is made if nbTrialRetry <= 0.
maxNbTrialSinceBest	if more than maxNbTrialSinceBest trials have been made since the best configuration was recorded, then revert to that configuration. This should be large enough to have an ergodic search (the default is never to revert..)

Warning: do not use the 'smart' option for the temperature schedule, it is not yet implemented. Later it will be used to set the temperatures as a function of the amplitude schedule, so that we accept between 30% and 70% moves.

Note: this will be removed when we separate the different algorithms in different classes.

Definition at line 460 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::XMLInput	(	istream &	is,
		const XMLCrystTag &	tag
	)

virtual

Input in XML format from a stream, restoring the set of refined objects and the associated cost functions.

Note that the corresponding objects must have been loaded in memory before, else shit happens.

Implements ObjCryst::OptimizationObj.

Definition at line 1930 of file GlobalOptimObj.cpp.

void ObjCryst::MonteCarloObj::XMLOutput	(	ostream &	os,
		int	indent = `0`
	)		const

virtual

Output a description of the object in XML format to a stream.

This saves the list of refined object and the cost functions, as well as options for the refinement. The refined objects are not saved, so this must be done somewhere else (they must be reloaded before this object).

Implements ObjCryst::OptimizationObj.

Definition at line 1869 of file GlobalOptimObj.cpp.

Member Data Documentation

RefObjOpt ObjCryst::MonteCarloObj::mGlobalOptimType

protected

Method used for the global optimization.

Should be removed when we switch to using several classes for different algorithms.

Definition at line 465 of file GlobalOptimObj.h.

REAL ObjCryst::MonteCarloObj::mMutationAmplitude

protected

Mutation amplitude.

From .25 to 64. Random moves will have a maximum amplitude equal to this amplitude multiplied by the Global optimization step defined for each RefinablePar. Large amplitude should be used at the beginning of the refinement (high temeratures).

Definition at line 492 of file GlobalOptimObj.h.

long ObjCryst::MonteCarloObj::mNbTrialRetry

protected

Number of trials before testing if we are below the given minimum cost.

If <=0, this will be ignored.

Definition at line 504 of file GlobalOptimObj.h.

The documentation for this class was generated from the following files:

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation