DynArcLookUp< GR > Class Template Reference

Detailed Description

template<typename GR>
class lemon::DynArcLookUp< GR >

Using this class, you can find an arc in a digraph from a given source to a given target in amortized time O(logd), where d is the out-degree of the source node.

It is possible to find all parallel arcs between two nodes with the operator() member.

This is a dynamic data structure. Consider to use ArcLookUp or AllArcLookUp if your digraph is not changed so frequently.

This class uses a self-adjusting binary search tree, the Splay tree of Sleator and Tarjan to guarantee the logarithmic amortized time bound for arc look-ups. This class also guarantees the optimal time bound in a constant factor for any distribution of queries.

Template Parameters

GR	The type of the underlying digraph.

See Also

ArcLookUp

AllArcLookUp

#include <lemon/core.h>

Inherits ObserverBase< GR, GR::Arc >.

Public Types
typedef GR	Digraph
	The Digraph type.

Public Member Functions
	DynArcLookUp (const Digraph &g)
	Constructor. More...
Arc	operator() (Node s, Node t, Arc p=INVALID) const
	Find an arc between two nodes. More...

Constructor & Destructor Documentation

DynArcLookUp ( const Digraph &  g )

inline

Constructor.

It builds up the search database.

Member Function Documentation

Arc operator() ( Node  s, Node  t, Arc  p = INVALID  ) const

inline

Find an arc between two nodes.

Parameters

s	The source node.
t	The target node.
p	The previous arc between s and t. It it is INVALID or not given, the operator finds the first appropriate arc.

Returns

An arc from s to t after p or INVALID if there is no more.

For example, you can count the number of arcs from u to v in the following way.

* DynArcLookUp<ListDigraph> ae(g);
* ...
* int n = 0;
* for(Arc a = ae(u,v); a != INVALID; a = ae(u,v,a)) n++;
* 

Finding the arcs take at most O(logd) amortized time, specifically, the time complexity of the lookups is equal to the optimal search tree implementation for the current query distribution in a constant factor.

Note

This is a dynamic data structure, therefore the data structure is updated after each graph alteration. Thus although this data structure is theoretically faster than ArcLookUp and AllArcLookUp, it often provides worse performance than them.