Project 3 Part II Dijkstras shortest path DSP

1
Project 3Part II Dijkstras shortest path (DSP)
Use a min-heap to implement DSP
V set of all vertices S set of completed
vertices. vertices in H V -S - Root of the
heap has the smallest distance from the source s
2
Implementation issues

• heap class
• It is a min-heap.
• Refer to Lafores max-heap implementation.
• First, define a new Node class
• E.g.
• Class Node
• int vertex_id
• double distance
• .
• heap functions
• remove() remove the root, that is, the node
  with minimum key.
• trickeDown(int heap_id) It can be used by remove
  function.
• insert( int vertex_id, double dist) insert a
  new node into the heap
• trickleUp( int heap_id) It can be used by insert
  function.
• decreaseDistance( int vertex_id, double dist)
  change a nodes key value.
• Note

3
Use a map to store vertices locations in the heap

• class heap
• private Node heapArray
• private int currentSize
• private int map
• .
• mapi
• mapi changes as the node for vertex i is moving
  in the heap.

4
Incorporate DSP into myGraph class

• public void DSP(string s)
• String shortestPath( String s, String t,
  boolean byHops)
• If byHops true, use breath first traversal
• If byHops false, call DSP.
• Just modeify old shortestpath method.
• int SPTree
• Like BFTree and DFTree, store the shortest paths
  generated by DSP.
• SPTreei the predecessor of vertex i on the
  shortest path from s to i.