Open Graph Drawing Framework  v.2007.11

Overview Class Hierarchy Class Index Class List Members Namespaces Source Files Search for:

ogdf::FMMMLayout Class Reference

The fast multipole multilevel layout algorithm. More...

#include <FMMMLayout.h>

Inheritance diagram for ogdf::FMMMLayout:

List of all members.

Public Types
enum	PageFormatType { pfPortrait, pfLandscape, pfSquare }
	Possible page formats. More...
enum	QualityVsSpeed { qvsGorgeousAndEfficient, qvsBeautifulAndFast, qvsNiceAndIncredibleSpeed }
	Trade-off between run-time and quality. More...
enum	EdgeLengthMeasurement { elmMidpoint, elmBoundingCircle }
	Specifies how the length of an edge is measured. More...
enum	AllowedPositions { apAll, apInteger, apExponent }
	Specifies which positions for a node are allowed. More...
enum	TipOver { toNone, toNoGrowingRow, toAlways }
	Specifies in which case it is allowed to tip over drawings of connected components. More...
enum	PreSort { psNone, psDecreasingHeight, psDecreasingWidth }
	Specifies how connected components are sorted before the packing algorithm is applied. More...
enum	GalaxyChoice { gcUniformProb, gcNonUniformProbLowerMass, gcNonUniformProbHigherMass }
	Specifies how sun nodes of galaxies are selected. More...
enum	MaxIterChange { micConstant, micLinearlyDecreasing, micRapidlyDecreasing }
	Specifies how MaxIterations is changed in subsequent multilevels. More...
enum	InitialPlacementMult { ipmSimple, ipmAdvanced }
	Specifies how the initial placement is generated. More...
enum	ForceModel { fmFruchtermanReingold, fmEades, fmNew }
	Specifies the force model. More...
enum	RepulsiveForcesMethod { rfcExact, rfcGridApproximation, rfcNMM }
	Specifies how to calculate repulsive forces. More...
enum	StopCriterion { scFixedIterations, scThreshold, scFixedIterationsOrThreshold }
	Specifies the stop criterion. More...
enum	InitialPlacementForces { ipfUniformGrid, ipfRandomTime, ipfRandomRandIterNr }
	Specifies how the initial placement is done. More...
enum	ReducedTreeConstruction { rtcPathByPath, rtcSubtreeBySubtree }
	Specifies how the reduced bucket quadtree is constructed. More...
enum	SmallestCellFinding { scfIteratively, scfAluru }
	Specifies how to calculate the smallest quadratic cell surrounding particles of a node in the reduced bucket quadtree. More...
Public Member Functions
	FMMMLayout ()
	Creates an instance of the layout algorithm.
	~FMMMLayout ()
The algorithm call
void	call (GraphAttributes &GA)
	Calls the algorithm for graph GA and returns the layout information in AG.
void	call (GraphAttributes &AG, const EdgeArray< double > &edgeLength)
	Extended algorithm call: Allows to pass desired lengths of the edges.
void	call (GraphAttributes &AG, char *ps_file)
	Extended algorithm call: Calls the algorithm for graph AG.
void	call (GraphAttributes &AG, const EdgeArray< double > &edgeLength, char *ps_file)
	Extend algorithm call: Allows to pass desired lengths of the edges.
Further information.
double	getCpuTime ()
	Returns the runtime (=CPU-time) of the layout algorithm in seconds.
High-level options
Allow to specify the most relevant parameters.
bool	useHighLevelOptions () const
	Returns the current setting of option useHighLevelOptions.
void	useHighLevelOptions (bool uho)
	Sets the option useHighLevelOptions to uho.
PageFormatType	pageFormat () const
	Returns the current setting of option pageFormat.
void	pageFormat (PageFormatType t)
	Sets the option pageRatio to t.
double	unitEdgeLength () const
	Returns the current setting of option unitEdgeLength.
void	unitEdgeLength (double x)
	Sets the option unitEdgeLength to x.
bool	newInitialPlacement () const
	Returns the current setting of option newInitialPlacement.
void	newInitialPlacement (bool nip)
	Sets the option newInitialPlacement to nip.
QualityVsSpeed	qualityVersusSpeed () const
	Returns the current setting of option qualityVersusSpeed.
void	qualityVersusSpeed (QualityVsSpeed qvs)
	Sets the option qualityVersusSpeed to qvs.
General low-level options
The low-level options in this and the following sections are meant for experts or interested people only.
void	randSeed (int p)
	Sets the seed of the random number generator.
int	randSeed () const
	Returns the seed of the random number generator.
EdgeLengthMeasurement	edgeLengthMeasurement () const
	Returns the current setting of option edgeLengthMeasurement.
void	edgeLengthMeasurement (EdgeLengthMeasurement elm)
	Sets the option edgeLengthMeasurement to elm.
AllowedPositions	allowedPositions () const
	Returns the current setting of option allowedPositions.
void	allowedPositions (AllowedPositions ap)
	Sets the option allowedPositions to ap.
int	maxIntPosExponent () const
	Returns the current setting of option maxIntPosExponent.
void	maxIntPosExponent (int e)
	Sets the option maxIntPosExponent to e.
Options for the divide et impera step
double	pageRatio () const
	Returns the current setting of option pageRatio.
void	pageRatio (double r)
	Sets the option pageRatio to r.
int	stepsForRotatingComponents () const
	Returns the current setting of option stepsForRotatingComponents.
void	stepsForRotatingComponents (int n)
	Sets the option stepsForRotatingComponents to n.
TipOver	tipOverCCs () const
	Returns the current setting of option tipOverCCs.
void	tipOverCCs (TipOver to)
	Sets the option tipOverCCs to to.
double	minDistCC () const
	Returns the minimal distance between connected components.
void	minDistCC (double x)
	Sets the minimal distance between connected components to x.
PreSort	presortCCs () const
	Returns the current setting of option presortCCs.
void	presortCCs (PreSort ps)
	Sets the option presortCCs to ps.
Options for the multilevel step
int	minGraphSize () const
	Returns the current setting of option minGraphSize.
void	minGraphSize (int n)
	Sets the option minGraphSize to n.
GalaxyChoice	galaxyChoice () const
	Returns the current setting of option galaxyChoice.
void	galaxyChoice (GalaxyChoice gc)
	Sets the option galaxyChoice to gc.
int	randomTries () const
	Returns the current setting of option randomTries.
void	randomTries (int n)
	Sets the option randomTries to n.
MaxIterChange	maxIterChange () const
	Returns the current setting of option maxIterChange.
void	maxIterChange (MaxIterChange mic)
	Sets the option maxIterChange to mic.
int	maxIterFactor () const
	Returns the current setting of option maxIterFactor.
void	maxIterFactor (int f)
	Sets the option maxIterFactor to f.
InitialPlacementMult	initialPlacementMult () const
	Returns the current setting of option initialPlacementMult.
void	initialPlacementMult (InitialPlacementMult ipm)
	Sets the option initialPlacementMult to ipm.
Options for the force calculation step
ForceModel	forceModel () const
	Returns the used force model.
void	forceModel (ForceModel fm)
	Sets the used force model to fm.
double	springStrength () const
	Returns the strength of the springs.
void	springStrength (double x)
	Sets the strength of the springs to x.
double	repForcesStrength () const
	Returns the strength of the repulsive forces.
void	repForcesStrength (double x)
	Sets the strength of the repulsive forces to x.
RepulsiveForcesMethod	repulsiveForcesCalculation () const
	Returns the current setting of option repulsiveForcesCalculation.
void	repulsiveForcesCalculation (RepulsiveForcesMethod rfc)
	Sets the option repulsiveForcesCalculation to rfc.
StopCriterion	stopCriterion () const
	Returns the stop criterion.
void	stopCriterion (StopCriterion rsc)
	Sets the stop criterion to rsc.
double	threshold () const
	Returns the threshold for the stop criterion.
void	threshold (double x)
	Sets the threshold for the stop criterion to x.
int	fixedIterations () const
	Returns the fixed number of iterations for the stop criterion.
void	fixedIterations (int n)
	Sets the fixed number of iterations for the stop criterion to n.
double	forceScalingFactor () const
	Returns the scaling factor for the forces.
void	forceScalingFactor (double f)
	Sets the scaling factor for the forces to \ f.
bool	coolTemperature () const
	Returns the current setting of option coolTemperature.
void	coolTemperature (bool b)
	Sets the option coolTemperature to b.
double	coolValue () const
	Returns the current setting of option coolValue.
void	coolValue (double x)
	Sets the option coolValue to x.
InitialPlacementForces	initialPlacementForces () const
	Returns the current setting of option initialPlacementForces.
void	initialPlacementForces (InitialPlacementForces ipf)
	Sets the option initialPlacementForces to ipf.
Options for the postprocessing step
bool	resizeDrawing () const
	Returns the current setting of option resizeDrawing.
void	resizeDrawing (bool b)
	Sets the option resizeDrawing to b.
double	resizingScalar () const
	Returns the current setting of option resizingScalar.
void	resizingScalar (double s)
	Sets the option resizingScalar to s.
int	fineTuningIterations () const
	Returns the number of iterations for fine tuning.
void	fineTuningIterations (int n)
	Sets the number of iterations for fine tuning to n.
double	fineTuneScalar () const
	Returns the curent setting of option fineTuneScalar.
void	fineTuneScalar (double s)
	Sets the option fineTuneScalar to s.
bool	adjustPostRepStrengthDynamically () const
	Returns the current setting of option adjustPostRepStrengthDynamically.
void	adjustPostRepStrengthDynamically (bool b)
	Sets the option adjustPostRepStrengthDynamically to b.
double	postSpringStrength () const
	Returns the strength of the springs in the postprocessing step.
void	postSpringStrength (double x)
	Sets the strength of the springs in the postprocessing step to x.
double	postStrengthOfRepForces () const
	Returns the strength of the repulsive forces in the postprocessing step.
void	postStrengthOfRepForces (double x)
	Sets the strength of the repulsive forces in the postprocessing step to x.
Options for repulsive force approximation methods
int	frGridQuotient () const
	Returns the current setting of option frGridQuotient.
void	frGridQuotient (int p)
	Sets the option frGridQuotient to p.
ReducedTreeConstruction	nmTreeConstruction () const
	Returns the current setting of option nmTreeConstruction.
void	nmTreeConstruction (ReducedTreeConstruction rtc)
	Sets the option nmTreeConstruction to rtc.
SmallestCellFinding	nmSmallCell () const
	Returns the current setting of option nmSmallCell.
void	nmSmallCell (SmallestCellFinding scf)
	Sets the option nmSmallCell to scf.
int	nmParticlesInLeaves () const
	Returns the current setting of option nmParticlesInLeaves.
void	nmParticlesInLeaves (int n)
	Sets the option nmParticlesInLeaves to n.
int	nmPrecision () const
	Returns the precision p for the p-term multipole expansions.
void	nmPrecision (int p)
	Sets the precision for the multipole expansions to \ p.
Private Member Functions
void	call_DIVIDE_ET_IMPERA_step (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E)
	Calls the divide (decomposition into connected components) and impera (drawing and packing of the componenets) step.
void	call_MULTILEVEL_step_for_subGraph (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E, int comp_index)
	Calls the multilevel step for subGraph G.
void	call_FORCE_CALCULATION_step (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E, int act_level, int max_level)
	Calls the force calculation step for G, A, E.
void	call_POSTPROCESSING_step (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E, NodeArray< DPoint > &F, NodeArray< DPoint > &F_attr, NodeArray< DPoint > &F_rep, NodeArray< DPoint > &last_node_movement)
	Calls the postprocessing step.
void	initialize_all_options ()
	All parameter options are set to the default values.
void	update_low_level_options_due_to_high_level_options_settings ()
	Updates several low level parameter options due to the settings of the high level parameter options.
void	import_NodeAttributes (const Graph &G, GraphAttributes &GA, NodeArray< NodeAttributes > &A)
	Imports for each node v of G its width, height and position(given from GA) in A.
void	import_EdgeAttributes (const Graph &G, const EdgeArray< double > &edgeLength, EdgeArray< EdgeAttributes > &E)
	Imports for each edge e of G its desired length given via edgeLength.
void	init_ind_ideal_edgelength (const Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E)
	Sets the individual ideal edge length for each edge e.
void	set_radii (const Graph &G, NodeArray< NodeAttributes > &A)
	The radii of the surrounding circles of the bounding boxes are computed.
void	export_NodeAttributes (Graph &G_reduced, NodeArray< NodeAttributes > &A_reduced, GraphAttributes &GA)
	Exports for each node v in G_reduced the position of the original_node in G.
void	make_simple_loopfree (const Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes >E, Graph &G_reduced, NodeArray< NodeAttributes > &A_reduced, EdgeArray< EdgeAttributes > &E_reduced)
	Creates a simple and loopfree copy of G and stores the corresponding node / edge attributes.
void	delete_parallel_edges (const Graph &G, EdgeArray< EdgeAttributes > &E, Graph &G_reduced, List< edge > &S, EdgeArray< double > &new_edgelength)
	Deletes parallel edges of G_reduced.
void	update_edgelength (List< edge > &S, EdgeArray< double > &new_edgelength, EdgeArray< EdgeAttributes > &E_reduced)
	Sets for each edge e of G_reduced in S its edgelength to new_edgelength[e].
double	get_post_rep_force_strength (int n)
	Returns the value for the strength of the repulsive forces.
void	make_positions_integer (Graph &G, NodeArray< NodeAttributes > &A)
	Makes the node positions integers.
void	create_postscript_drawing (GraphAttributes &AG, char *ps_file)
	Creates a simple drawing of AG in postscript format and saves it in file ps_file.
void	create_maximum_connected_subGraphs (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E, Graph G_sub[], NodeArray< NodeAttributes > A_sub[], EdgeArray< EdgeAttributes > E_sub[], NodeArray< int > &component)
	Constructs the list of connected components of G.
void	pack_subGraph_drawings (NodeArray< NodeAttributes > &A, Graph G_sub[], NodeArray< NodeAttributes > A_sub[])
	The drawings of the subgraphs are packed.
void	calculate_bounding_rectangles_of_components (List< Rectangle > &R, Graph G_sub[], NodeArray< NodeAttributes > A_sub[])
	The bounding rectangles of all connected componenents of G are calculated and stored in R.
Rectangle	calculate_bounding_rectangle (Graph &G, NodeArray< NodeAttributes > &A, int componenet_index)
	The bounding rectangle of the componenet_index-th. component of G is returned.
void	rotate_components_and_calculate_bounding_rectangles (List< Rectangle > &R, Graph G_sub[], NodeArray< NodeAttributes > A_sub[])
double	calculate_area (double width, double height, int comp_nr)
void	export_node_positions (NodeArray< NodeAttributes > &A, List< Rectangle > &R, Graph G_sub[], NodeArray< NodeAttributes > A_sub[])
void	delete_all_subGraphs (Graph G_sub[], NodeArray< NodeAttributes > A_sub[], EdgeArray< EdgeAttributes > E_sub[])
	Frees dynamically allocated memory for the connected component subgraphs.
int	get_max_mult_iter (int act_level, int max_level, int node_nr)
void	calculate_forces (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E, NodeArray< DPoint > &F, NodeArray< DPoint > &F_attr, NodeArray< DPoint > &F_rep, NodeArray< DPoint > &last_node_movement, int iter, int fine_tuning_step)
	The forces are calculated here.
void	init_boxlength_and_cornercoordinate (Graph &G, NodeArray< NodeAttributes > &A)
	The length of the computational box in the first iteration is set (down left corner is at (0,0).
void	create_initial_placement (Graph &G, NodeArray< NodeAttributes > &A)
	The initial placements of the nodes are created by using initialPlacementForces().
void	init_F (Graph &G, NodeArray< DPoint > &F)
	Sets all entries of F to (0,0).
void	make_initialisations_for_rep_calc_classes (Graph &G, NodeArray< NodeAttributes > &A, NodeArray< DPoint > &F_rep)
	Make initializations for the data structures that are used in the choosen class for rep. force calculation.
void	calculate_repulsive_forces (Graph &G, NodeArray< NodeAttributes > &A, NodeArray< DPoint > &F_rep)
	Calculates repulsive forces for each node.
void	deallocate_memory_for_rep_calc_classes ()
	Deallocates dynamically allocated memory of the choosen rep. calculation class.
void	calculate_attractive_forces (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E, NodeArray< DPoint > &F_attr)
	Calculates attractive forces for each node.
double	f_attr_scalar (double d, double ind_ideal_edge_length)
	Returns the attractive force scalar.
void	add_attr_rep_forces (Graph &G, NodeArray< DPoint > &F_attr, NodeArray< DPoint > &F_rep, NodeArray< DPoint > &F, int iter, int fine_tuning_step)
	Add attractive and repulsive forces for each node.
void	move_nodes (Graph &G, NodeArray< NodeAttributes > &A, NodeArray< DPoint > &F)
	Move the nodes.
void	update_boxlength_and_cornercoordinate (Graph &G, NodeArray< NodeAttributes > &A)
	Computes a new tight computational square-box.
double	max_radius (int iter)
	Describes the max. radius of a move in one time step, depending on the number of iterations.
void	set_average_ideal_edgelength (Graph &G, EdgeArray< EdgeAttributes > &E)
	The average_ideal_edgelength for all edges is computed.
double	get_average_forcevector_length (Graph &G, NodeArray< DPoint > &F)
void	prevent_oscilations (Graph &G, NodeArray< DPoint > &F, NodeArray< DPoint > &last_node_movement, int iter)
double	angle (DPoint &P, DPoint &Q, DPoint &R)
	Calculates the angle between PQ and PS in [0,2pi).
void	init_last_node_movement (Graph &G, NodeArray< DPoint > &F, NodeArray< DPoint > &last_node_movement)
	last_node_movement is initialized to F (used after first iteration).
void	adapt_drawing_to_ideal_average_edgelength (Graph &G, NodeArray< NodeAttributes > &A, EdgeArray< EdgeAttributes > &E)
void	restrict_force_to_comp_box (DPoint &force)
void	init_time ()
	Sets time_total to zero.
Private Attributes
bool	m_useHighLevelOptions
	The option for using high-level options.
PageFormatType	m_pageFormat
	The option for the page format.
double	m_unitEdgeLength
	The unit edge length.
bool	m_newInitialPlacement
	The option for new initial placement.
QualityVsSpeed	m_qualityVersusSpeed
	The option for quality-vs-speed trade-off.
int	m_randSeed
	The random seed.
EdgeLengthMeasurement	m_edgeLengthMeasurement
	The option for edge length measurement.
AllowedPositions	m_allowedPositions
	The option for allowed positions.
int	m_maxIntPosExponent
	The option for the used exponent.
double	m_pageRatio
	The desired page ratio.
int	m_stepsForRotatingComponents
	The number of rotations.
TipOver	m_tipOverCCs
	Option for tip-over of connected components.
double	m_minDistCC
	The separation between connected components.
PreSort	m_presortCCs
	The option for presorting connected components.
int	m_minGraphSize
	The option for minimal graph size.
GalaxyChoice	m_galaxyChoice
	The selection of galaxy nodes.
int	m_randomTries
	The number of random tries.
MaxIterChange	m_maxIterChange
	The option for how to change MaxIterations.
int	m_maxIterFactor
	The factor used for decreasing MaxIterations.
InitialPlacementMult	m_initialPlacementMult
	The option for creating initial placement.
ForceModel	m_forceModel
	The used force model.
double	m_springStrength
	The strengths of springs.
double	m_repForcesStrength
	The strength of repulsive forces.
RepulsiveForcesMethod	m_repulsiveForcesCalculation
	Option for how to calculate repulsive forces.
StopCriterion	m_stopCriterion
	The stop criterion.
double	m_threshold
	The threshold for the stop criterion.
int	m_fixedIterations
	The fixed number of iterations for the stop criterion.
double	m_forceScalingFactor
	The scaling factor for the forces.
bool	m_coolTemperature
	The option for how to scale forces.
double	m_coolValue
	The value by which forces are decreased.
InitialPlacementForces	m_initialPlacementForces
	The option for how the initial placement is done.
bool	m_resizeDrawing
	The option for resizing the drawing.
double	m_resizingScalar
	Parameter for resizing the drawing.
int	m_fineTuningIterations
	The number of iterations for fine tuning.
double	m_fineTuneScalar
	Parameter for scaling forces during fine tuning.
bool	m_adjustPostRepStrengthDynamically
	The option adjustPostRepStrengthDynamically.
double	m_postSpringStrength
	The strength of springs during postprocessing.
double	m_postStrengthOfRepForces
	The strength of repulsive forces during postprocessing.
int	m_frGridQuotient
	The grid quotient.
ReducedTreeConstruction	m_NMTreeConstruction
	The option for how to construct reduced bucket quadtree.
SmallestCellFinding	m_NMSmallCell
	The option for how to calculate smallest quadtratic cells.
int	m_NMParticlesInLeaves
	The maximal number of particles in a leaf.
int	m_NMPrecision
	The precision for multipole expansions.
double	max_integer_position
	The maximum value for an integer position.
double	cool_factor
	Needed for scaling the forces if coolTemperature is true.
double	average_ideal_edgelength
	Measured from center to center.
double	boxlength
	Holds the length of the quadratic comput. box.
int	number_of_components
	The number of components of the graph.
DPoint	down_left_corner
	Holds down left corner of the comput. box.
NodeArray< double >	radius
	Holds the radius of the surrounding circle for each node.
double	time_total
	The runtime (=CPU-time) of the algorithm in seconds.
FruchtermanReingold	FR
	Class for repulsive force calculation (Fruchterman, Reingold).
NMM	NM
	Class for repulsive force calculation.

---

Detailed Description

The fast multipole multilevel layout algorithm.

The class FMMMLayout implements a force-directed graph drawing method suited also for very large graphs. It is based on a combination of an efficient multilevel scheme and a strategy for approximating the repusive forcesin the system by rapidly evaluating potential fields.

The implementation is based on the following publication:

Stefan Hachul, Michael Jünger: Drawing Large Graphs with a Potential-Field-Based Multilevel Algorithm. 12th International Symposium on Graph Drawing 1998, New York (GD '04), LNCS 3383, pp. 285-295, 2004.

Optional parameters

The following options are the most important. You can set useHighLevelOptions to true and just need to asjust a few parameters. However, you can also adjust all parameters that the implementation uses (see below), but this requires good knowledge of the algorithm.

Option	Type	Default	Description
useHighLevelOptions	bool	false	Wether high-level options are used or not.
pageFormat	PageFormatType	pfSquare	The desired apsect ratio of the layout.
unitEdgeLength	double	100.0	The unit edge length.
newInitialPlacement	bool	false	Specifies if initial placement of nodes is varied.
qualityVersusSpeed	QualityVsSpeed	qvsBeautifulAndFast	Indicates if the algorithm is tuned either for best quality or best speed.

If you want to do more detailed fine-tuning, you can adjust all parameters used by the algorithm. Please refer to the paper cited above for better understanding of the algorithm.

General
randSeed	int	100	The seed of the random number generator.
edgeLengthMeasurement	EdgeLengthMeasurement	elmBoundingCircle	Indicates how the length of an edge is measured.
allowedPositions	AllowedPositions	apInteger	Defines which positions for a node are allowed.
maxIntPosExponent	int	40	Defines the exponent used if allowedPositions == apExponent.
Divide et impera step
pageRatio	double	1.0	The desired page ratio.
stepsForRotatingComponents	int	10	The number of rotations per connected component.
tipOverCCs	TipOver	toNoGrowingRow	Specifies when it is allowed to tip over drawings.
minDistCC	double	100	The minimal distance between connected components.
presortCCs	PreSort	psDecreasingHeight	Defines if the connected components are sorted before the packing algorithm is applied.
Multilevel step
minGraphSize	int	50	Determines the number of nodes of a graph in the for which no more collapsing of galaxies is performed
galaxyChoice	GalaxyChoice	gcNonUniformProbLowerMass	Defines how sun nodes of galaxies are selected.
randomTries	int	20	Defines the number of tries to get a random node with minimal star mass.
maxIterChange	MaxIterChange	micLinearlyDecreasing	Defines how MaxIterations is changed in subsequent multilevels.
maxIterFactor	int	10	Defines the factor used for decrasing MaxIterations.
initialPlacementMult	InitialPlacementMult	ipmAdvanced	Defines how the initial placement is generated.
Force calculation step
forceModel	ForceModel	fmNew	The used force model.
springStrength	double	1.0	The strength of the springs.
repForcesStrength	double	1.0	The strength of the repulsive forces.
repulsiveForcesCalculation	RepulsiveForcesMethod	rfcNMM	Defines how to calculate repulsive forces.
stopCriterion	StopCriterion	scFixedIterationsOrThreshold	The stop criterion.
threshold	double	0.01	The threshold for the stop criterion.
fixedIterations	int	30	The fixed number of iterations for the stop criterion.
forceScalingFactor	double	0.05	The scaling factor for the forces.
coolTemperature	bool	false	Use coolValue for scaling forces.
coolValue	double	0.99	The value by which forces are decreased.
initialPlacementForces	InitialPlacementForces	ipfRandomRandIterNr	Defines how the initial placement is done.
Force calculation step
resizeDrawing	bool	true	Specifies if the resulting drawing is resized.
resizingScalar	double	1	Defines a parameter to scale the drawing if resizeDrawing is true.
fineTuningIterations	int	20	The number of iterations for fine tuning.
fineTuneScalar	double	0.2	Defines a parameter for scaling the forces in the fine-tuning iterations.
adjustPostRepStrengthDynamically	bool	true	If set to true, the strength of the repulsive force field is calculated.
postSpringStrength	double	2.0	The strength of the springs in the postprocessing step.
postStrengthOfRepForces	double	0.01	The strength of the repulsive forces in the postprocessing step.
Repulsive force approximation methods
frGridQuotient	int	2	The grid quotient.
nmTreeConstruction	ReducedTreeConstruction	rtcSubtreeBySubtree	Defines how the reduced bucket quadtree is constructed.
nmSmallCell	SmallestCellFinding	scfIteratively	Defines how the smallest quadratic cell that surrounds the particles of a node in the reduced bucket quadtree is calculated.
nmParticlesInLeaves	int	25	The maximal number of particles that are contained in a leaf of the reduced bucket quadtree.
nmPrecision	int	4	The precision p for the p-term multipole expansions.

Running time

The running time of the algorithm is O(n log n + m) for graphs with n nodes and m edges. The required space is linear in the input size.

Definition at line 256 of file FMMMLayout.h.

---

Member Enumeration Documentation