Go to the documentation of this file.00001
00002
00003
00004
00005
00006 from priodict import priorityDictionary
00007
00008 def Dijkstra(G,start,end=None):
00009 """
00010 Find shortest paths from the start vertex to all
00011 vertices nearer than or equal to the end.
00012
00013 The input graph G is assumed to have the following
00014 representation: A vertex can be any object that can
00015 be used as an index into a dictionary. G is a
00016 dictionary, indexed by vertices. For any vertex v,
00017 G[v] is itself a dictionary, indexed by the neighbors
00018 of v. For any edge v->w, G[v][w] is the length of
00019 the edge. This is related to the representation in
00020 <http://www.python.org/doc/essays/graphs.html>
00021 where Guido van Rossum suggests representing graphs
00022 as dictionaries mapping vertices to lists of neighbors,
00023 however dictionaries of edges have many advantages
00024 over lists: they can store extra information (here,
00025 the lengths), they support fast existence tests,
00026 and they allow easy modification of the graph by edge
00027 insertion and removal. Such modifications are not
00028 needed here but are important in other graph algorithms.
00029 Since dictionaries obey iterator protocol, a graph
00030 represented as described here could be handed without
00031 modification to an algorithm using Guido's representation.
00032
00033 Of course, G and G[v] need not be Python dict objects;
00034 they can be any other object that obeys dict protocol,
00035 for instance a wrapper in which vertices are URLs
00036 and a call to G[v] loads the web page and finds its links.
00037
00038 The output is a pair (D,P) where D[v] is the distance
00039 from start to v and P[v] is the predecessor of v along
00040 the shortest path from s to v.
00041
00042 Dijkstra's algorithm is only guaranteed to work correctly
00043 when all edge lengths are positive. This code does not
00044 verify this property for all edges (only the edges seen
00045 before the end vertex is reached), but will correctly
00046 compute shortest paths even for some graphs with negative
00047 edges, and will raise an exception if it discovers that
00048 a negative edge has caused it to make a mistake.
00049 """
00050
00051 D = {}
00052 P = {}
00053 Q = priorityDictionary()
00054 Q[start] = 0
00055
00056 for v in Q:
00057 D[v] = Q[v]
00058 if v == end: break
00059
00060 for w in G[v]:
00061 vwLength = D[v] + G[v][w]
00062 if w in D:
00063 if vwLength < D[w]:
00064 raise ValueError, \
00065 "Dijkstra: found better path to already-final vertex"
00066 elif w not in Q or vwLength < Q[w]:
00067 Q[w] = vwLength
00068 P[w] = v
00069
00070 return (D,P)
00071
00072 def shortestPath(G,start,end):
00073 """
00074 Find a single shortest path from the given start vertex
00075 to the given end vertex.
00076 The input has the same conventions as Dijkstra().
00077 The output is a list of the vertices in order along
00078 the shortest path.
00079 """
00080
00081 D,P = Dijkstra(G,start,end)
00082 Path = []
00083 while 1:
00084 Path.append(end)
00085 if end == start: break
00086 end = P[end]
00087 Path.reverse()
00088 return Path
00089
00090