Class DirectedScaleFreeGraphGenerator<V,E>

java.lang.Object
org.jgrapht.generate.DirectedScaleFreeGraphGenerator<V,E>
Type Parameters:
V - the graph vertex type
E - the graph edge type
All Implemented Interfaces:
GraphGenerator<V,E,V>

public class DirectedScaleFreeGraphGenerator<V,E> extends Object implements GraphGenerator<V,E,V>
A generator for directed scale-free graphs.

This generator creates a directed scale-free graph according to a power law, as described in Bollobás et al. The paper can be cited as Béla Bollobás, Christian Borgs, Jennifer Chayes, and Oliver Riordan. "Directed scale-free graphs." Proceedings of the fourteenth annual ACM-SIAM symposium on Discrete algorithms. Society for Industrial and Applied Mathematics, 2003.

In This generator, the graph continues to grow one edge per step, according to the probabilities alpha, beta, and gamma (which sum up to 1).

  • alpha is the probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in.
  • beta is the probability that the new edge is from an existing vertex v to an existing vertex w, where v and w are chosen independently, v according to d_out + delta_out, and w according to d_in + delta_in.
  • gamma is the probability that the new edge is from an existing vertex v to a new vertex w, where v is chosen according to d_out + delta_out.

In their original paper, the graph continues to grow according to a certain power law until a certain number of edges is reached irrespective to the number of nodes.
However, because the target number of edges is not known beforehand, in this implementation, we added another feature that enables the user to grow the curve according to that power law until the target number of edges or target number of nodes is reached.

  • Nested Class Summary

    Nested Classes
    Modifier and Type
    Class
    Description
    private static enum 
    An enum to indicate the vertex selection using its inDegree or outDegree
  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    private boolean
    Control whether the generated graph may contain loops.
    private boolean
    Control whether the generated graph many contain multiple (parallel) edges between the same two vertices
    private final float
    probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in criterion.
    private final float
    The probability that the new edge is (from a new vertex v to an existing vertex w) plus the probability that the new edge is (from an existing vertex v to an existing vertex w), where v and w are chosen independently, v according to d_out + delta_out, and w according to d_in + delta_in criteria.
    private final float
    In-degree bias used for Alpha and Beta
    private final float
    Out-degree bias used for Beta and Gamma
    private int
    Maximum number of consecutive failed attempts to add an edge.
    private final Random
     
    private final int
    Target total number of edges to reach.
    private final int
    Target total number of targetNodes to reach.
    This has lower priority than targetEdges.
  • Constructor Summary

    Constructors
    Constructor
    Description
    DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes)
    Constructs a Generator.
    DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, long seed)
    Constructs a Generator using a seed for the random number generator.
    DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, long seed, boolean allowingMultipleEdges, boolean allowingSelfLoops)
    Constructs a Generator using a seed for the random number generator and sets the two relaxation options allowingMultipleEdges and allowingSelfLoops.
    DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, Random rng)
    Construct a new generator using the provided random number generator.
    DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, Random rng, boolean allowingMultipleEdges, boolean allowingSelfLoops)
    Construct a new generator using the provided random number generator and sets the two relaxation options allowingMultipleEdges and allowingSelfLoops.
  • Method Summary

    Modifier and Type
    Method
    Description
    void
    generateGraph(Graph<V,E> target, Map<String,V> resultMap)
    Generates an instance of the Graph.
    int
    Returns the maximum allowed number of consecutive failed attempts to add an edge.
    boolean
    Returns whether the generated graph may contain multiple (parallel) edges between the same two vertices.
    boolean
    Returns whether the generated graph may contain multiple (parallel) edges between the same two vertices
    private V
    pickAVertex(Graph<V,E> target, Set<V> allNewNodes, Set<E> allNewEdgesSet, DirectedScaleFreeGraphGenerator.Direction direction, float bias)
    Select a vertex from the currently available vertices, using the passed bias.
    void
    setAllowingMultipleEdges(boolean allowingMultipleEdges)
    Sets whether the generated graph may contain multiple (parallel) edges between the same two vertices
    void
    setAllowingSelfLoops(boolean allowingSelfLoops)
    Sets whether the generated graph may contain multiple (parallel) edges between the same two vertices
    void
    setMaxFailures(int maxFailures)
    Sets the maximum allowed number of consecutive failed attempts to add an edge (must be non negative).

    Methods inherited from class java.lang.Object

    clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    Methods inherited from interface org.jgrapht.generate.GraphGenerator

    generateGraph
  • Field Details

    • rng

      private final Random rng
    • alpha

      private final float alpha
      probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in criterion.
    • alphaPlusBeta

      private final float alphaPlusBeta
      The probability that the new edge is (from a new vertex v to an existing vertex w) plus the probability that the new edge is (from an existing vertex v to an existing vertex w), where v and w are chosen independently, v according to d_out + delta_out, and w according to d_in + delta_in criteria. This equals 1 - gamma. Gamma refers to the probability that the new edge is from an existing vertex v to a new vertex w.
    • deltaIn

      private final float deltaIn
      In-degree bias used for Alpha and Beta
    • deltaOut

      private final float deltaOut
      Out-degree bias used for Beta and Gamma
    • targetEdges

      private final int targetEdges
      Target total number of edges to reach. It has a higher priority than targetNodes. Zero is a valid value.
      If negative number, the user does not care about the total number of edges and is interested only in the number of nodes, therefore, targetNodes will be considered instead. Otherwise, targetEdges will be considered and targetEdges will be ignored.
    • targetNodes

      private final int targetNodes
      Target total number of targetNodes to reach.
      This has lower priority than targetEdges. It will not be used unless targetEdges given is a negative number.
    • maxFailures

      private int maxFailures
      Maximum number of consecutive failed attempts to add an edge.
    • allowingMultipleEdges

      private boolean allowingMultipleEdges
      Control whether the generated graph may contain loops.
    • allowingSelfLoops

      private boolean allowingSelfLoops
      Control whether the generated graph many contain multiple (parallel) edges between the same two vertices
  • Constructor Details

    • DirectedScaleFreeGraphGenerator

      public DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes)
      Constructs a Generator.
      Parameters:
      alpha - The probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in.
      gamma - The probability that the new edge is from an existing vertex v to a new vertex w, where v is chosen according to d_out + delta_out.
      deltaIn - The in-degree bias used for Alpha and Beta.
      deltaOut - The out-degree bias used for Beta and Gamma.
      targetEdges - Target total number of edges to reach. It has a higher priority than targetNodes. Zero is a valid value.
      If negative number, the user does not care about the total number of edges and is interested only in the number of nodes, therefore, targetNodes will be considered instead. Otherwise, targetEdges will be considered and targetEdges will be ignored.
      targetNodes - Target number of nodes to reach. Zero is a valid value.
      This parameter has lower priority than targetEdges and will be used only if targetEdges given is a negative number.
    • DirectedScaleFreeGraphGenerator

      public DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, long seed)
      Constructs a Generator using a seed for the random number generator.
      Parameters:
      alpha - The probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in.
      gamma - The probability that the new edge is from an existing vertex v to a new vertex w, where v is chosen according to d_out + delta_out.
      deltaIn - The in-degree bias used for Alpha and Beta.
      deltaOut - The out-degree bias used for Beta and Gamma.
      targetEdges - Target total number of edges to reach. It has a higher priority than targetNodes. Zero is a valid value.
      If negative number, the user does not care about the total number of edges and is interested only in the number of nodes, therefore, targetNodes will be considered instead. Otherwise, targetEdges will be considered and targetEdges will be ignored.
      targetNodes - Target number of nodes to reach. Zero is a valid value.
      This parameter has lower priority than targetEdges and will be used only if targetEdges given is a negative number.
      seed - The seed to feed to the random number generator.
    • DirectedScaleFreeGraphGenerator

      public DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, long seed, boolean allowingMultipleEdges, boolean allowingSelfLoops)
      Constructs a Generator using a seed for the random number generator and sets the two relaxation options allowingMultipleEdges and allowingSelfLoops.
      Parameters:
      alpha - The probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in.
      gamma - The probability that the new edge is from an existing vertex v to a new vertex w, where v is chosen according to d_out + delta_out.
      deltaIn - The in-degree bias used for Alpha and Beta.
      deltaOut - The out-degree bias used for Beta and Gamma.
      targetEdges - Target total number of edges to reach. It has a higher priority than targetNodes. Zero is a valid value.
      If negative number, the user does not care about the total number of edges and is interested only in the number of nodes, therefore, targetNodes will be considered instead. Otherwise, targetEdges will be considered and targetEdges will be ignored.
      targetNodes - Target number of nodes to reach. Zero is a valid value.
      This parameter has lower priority than targetEdges and will be used only if targetEdges given is a negative number.
      seed - The seed to feed to the random number generator.
      allowingMultipleEdges - whether the generator allows multiple parallel edges between the same two vertices (v, w).
      allowingSelfLoops - whether the generator allows self loops from the a vertex to itself.
    • DirectedScaleFreeGraphGenerator

      public DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, Random rng)
      Construct a new generator using the provided random number generator.
      Parameters:
      alpha - The probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in.
      gamma - The probability that the new edge is from an existing vertex v to a new vertex w, where v is chosen according to d_out + delta_out.
      deltaIn - The in-degree bias used for Alpha and Beta.
      deltaOut - The out-degree bias used for Beta and Gamma.
      targetEdges - Target total number of edges to reach. It has a higher priority than targetNodes. Zero is a valid value.
      If negative number, the user does not care about the total number of edges and is interested only in the number of nodes, therefore, targetNodes will be considered instead. Otherwise, targetEdges will be considered and targetEdges will be ignored.
      targetNodes - Target number of nodes to reach. Zero is a valid value.
      This parameter has lower priority than targetEdges and will be used only if targetEdges given is a negative number.
      rng - The Random object to use.
    • DirectedScaleFreeGraphGenerator

      public DirectedScaleFreeGraphGenerator(float alpha, float gamma, float deltaIn, float deltaOut, int targetEdges, int targetNodes, Random rng, boolean allowingMultipleEdges, boolean allowingSelfLoops)
      Construct a new generator using the provided random number generator and sets the two relaxation options allowingMultipleEdges and allowingSelfLoops.
      Parameters:
      alpha - The probability that the new edge is from a new vertex v to an existing vertex w, where w is chosen according to d_in + delta_in.
      gamma - The probability that the new edge is from an existing vertex v to a new vertex w, where v is chosen according to d_out + delta_out.
      deltaIn - The in-degree bias used for Alpha and Beta.
      deltaOut - The out-degree bias used for Beta and Gamma.
      targetEdges - Target total number of edges to reach. It has a higher priority than targetNodes. Zero is a valid value.
      If negative number, the user does not care about the total number of edges and is interested only in the number of nodes, therefore, targetNodes will be considered instead. Otherwise, targetEdges will be considered and targetEdges will be ignored.
      targetNodes - Target number of nodes to reach. Zero is a valid value.
      This parameter has lower priority than targetEdges and will be used only if targetEdges given is a negative number.
      rng - The Random object to use.
      allowingMultipleEdges - whether the generator allows multiple parallel edges between the same two vertices (v, w).
      allowingSelfLoops - whether the generator allows self loops from the a vertex to itself.
  • Method Details

    • generateGraph

      public void generateGraph(Graph<V,E> target, Map<String,V> resultMap)
      Generates an instance of the Graph.
      Specified by:
      generateGraph in interface GraphGenerator<V,E,V>
      Parameters:
      target - the target graph
      resultMap - not used by this generator, can be null
      Throws:
      TooManyFailuresException - When the method fails maxFailures times to add a new edge to the growing graph.
      IllegalArgumentException - When the graph does not support Multiple edges or self loop while the generator does.
    • pickAVertex

      private V pickAVertex(Graph<V,E> target, Set<V> allNewNodes, Set<E> allNewEdgesSet, DirectedScaleFreeGraphGenerator.Direction direction, float bias)
      Select a vertex from the currently available vertices, using the passed bias.
      Parameters:
      target - The target graph
      allNewNodes - All (new) nodes in the target graph
      allNewEdgesSet - All (new) edges in the target graph
      direction - DirectedScaleFreeGraphGenerator.Direction.IN for inDegree or DirectedScaleFreeGraphGenerator.Direction.IN for outDegree
      bias - deltaIn or deltaOut value according to #directioIn
      Returns:
      the selected node.
    • getMaxFailures

      public int getMaxFailures()
      Returns the maximum allowed number of consecutive failed attempts to add an edge.
      Returns:
      maxFailure field.
    • setMaxFailures

      public void setMaxFailures(int maxFailures)
      Sets the maximum allowed number of consecutive failed attempts to add an edge (must be non negative).
      Parameters:
      maxFailures - Maximum allowed (non negative) number of consecutive failed attempts to add an edge.
    • isAllowingMultipleEdges

      public boolean isAllowingMultipleEdges()
      Returns whether the generated graph may contain multiple (parallel) edges between the same two vertices.
      Returns:
      whether the generated graph may contain multiple (parallel) edges between the same two vertices
    • setAllowingMultipleEdges

      public void setAllowingMultipleEdges(boolean allowingMultipleEdges)
      Sets whether the generated graph may contain multiple (parallel) edges between the same two vertices
      Parameters:
      allowingMultipleEdges - whether the generated graph may contain multiple (parallel) edges between the same two vertices
    • isAllowingSelfLoops

      public boolean isAllowingSelfLoops()
      Returns whether the generated graph may contain multiple (parallel) edges between the same two vertices
      Returns:
      whether the generated graph many contain multiple (parallel) edges between the same two vertices
    • setAllowingSelfLoops

      public void setAllowingSelfLoops(boolean allowingSelfLoops)
      Sets whether the generated graph may contain multiple (parallel) edges between the same two vertices
      Parameters:
      allowingSelfLoops - whether the generated graph many contain multiple (parallel) edges between the same two vertices