def Dijkstra(G,start,end=None):
#Init
D = {} #Dctionary of final distances
P = {} #Dictionary of predecessors
Q = priodict.priorityDictionary()
Q[start] = 0
#While there are elements in the priority queue
for v in Q:
D[v] = Q[v]
if v == end:
break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,Q,P)
def dijkstra(G, start, end=None):
"""
Find shortest paths from the start vertex to all vertices nearer than or equal to the end.
For any vertex v, G[v] is itself a dictionary, indexed by the neighbors of v. For any edge v->w, G[v][w] is the length of the edge.
The output is a pair (D,P) where D[v] is the distance from start to v and P[v] is the predecessor of v along the shortest path from s to v.
"""
D = {} # dictionary of final distances.
P = {} # dictionary of predecessors.
Q = priorityDictionary() # estimated distances of non-final vertices.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w][0]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = (v, G[v][w][1])
return (D, P)
def __DijkstraFromSource(graph, sourceNode):
utils.Utils.isStr(sourceNode)
if sourceNode not in graph:
raise ValueError('the source node is not in the graph')
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # estimated distances of non-final vertices
Q[sourceNode] = 0
for anode in Q:
D[anode] = Q[anode]
graph[sourceNode].routingPath[anode]=node.Path([],0.0)
for neighbor in graph[anode].connections:
sourceToNeighborLenght = D[anode] + graph[anode].connections[neighbor]
if neighbor in D:
if sourceToNeighborLenght < D[neighbor]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif neighbor not in Q or sourceToNeighborLenght < Q[neighbor]:
Q[neighbor] = sourceToNeighborLenght
P[neighbor] = anode
graph[sourceNode].routingPath[anode].value = Q[anode]
end = anode;
while 1:
graph[sourceNode].routingPath[anode].path.append(end)
if end == sourceNode: break
end = P[end]
graph[sourceNode].routingPath[anode].path.reverse()
def djikstraExtended(self,s,t):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[s] = 0
Path = []
for v in Q:
D[v] = Q[v]
if v == t:
while 1:
if t == s: break
Path.append(P[t][1])
t = P[t][0]
Path.reverse()
return Path,Q
for edgew in self.adj(v):
w=edgew.v
vwLength = D[v] + edgew.length
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = [v, edgew]
return Path,Q
def astar(start, goal, data):
closedset = set()
openset = set()
openset.add(start)
came_from = {}
g_score = {}
g_score[start] = 0
f_score = priodict.priorityDictionary()
f_score[start] = g_score[start] + heuristic_cost_estimate(start, goal)
while len(openset) != 0:
current = f_score.smallest()
del f_score[current]
if current == goal:
return reconstruct_path(came_from, goal)
openset.remove(current)
closedset.add(current)
for neighbor in neighbor_nodes(current, data):
tentative_g_score = g_score[current] + dist_between(current, neighbor)
if neighbor in closedset and tentative_g_score >= g_score[neighbor]:
continue
if neighbor not in openset or tentative_g_score < g_score[neighbor]:
came_from[neighbor] = current
g_score[neighbor] = tentative_g_score
f_score[neighbor] = g_score[neighbor] + heuristic_cost_estimate(neighbor, goal)
if neighbor not in openset:
openset.add(neighbor)
return None
def astar (start,end,forbidden):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
G={}
for v in Q:
D[v] = Q[v]
if v == end:
break
if not G.has_key(v):
values= state_trans_fuction(v,Ux,sX)
hold_v = copy.deepcopy(values)
for x in values:
if G.has_key(x) :
#hold_v.remove(x)
continue
if x in forbidden:
hold_v.remove(x)
print x,"forbidden"
#if x in hold_v and x in forbidden:
# hold_v.remove(x)
print hold_v
G[v]=hold_v
for w in G[v]:
if w in forbidden:
pass
vwLength = D[v] + 1 + abs((end[0]-w[0]))+abs((end[1]-w[1]))
if w not in D or vwLength < D[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def dijkstra(G,start,end=None):
"""
Taken from http://code.activestate.com/recipes/119466-dijkstras-algorithm-for-shortest-paths/
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def Dijkstra(graph, start, end=None):
final_distances = {} # dictionary of final distances
predecessors = {}
estimated_distances = priorityDictionary() # heap binary tree initialization
estimated_distances[start] = 0
oneHop = False
hops = 0
for vertex in estimated_distances:
final_distances[vertex] = estimated_distances[vertex]
if (vertex == end and oneHop):
break
oneHop = True
hops = hops+1
for edge in graph[vertex]:
path_distance = final_distances[vertex] + graph[vertex][edge]
if edge in final_distances:
if path_distance < final_distances[edge]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
else:
if (final_distances[edge] == 0): #scenario node - node
final_distances[edge] = path_distance
if edge not in estimated_distances or path_distance < estimated_distances[edge]:
estimated_distances[edge] = path_distance
predecessors[edge] = vertex
return (final_distances,predecessors)
def dijkstra(map, start, end):
if start == end:
return 0
if not (map.has_key(start)) or not (map.has_key(end)):
return -1
distances = {}
nodes = {}
prelim_distances = priorityDictionary()
prelim_distances[start] = 0
for vertex in prelim_distances:
distances[vertex] = prelim_distances[vertex]
if vertex == end:
break
for edge in map[vertex]:
dist = distances[vertex] + map[vertex][edge]
if edge in distances:
if dist < distances[edge]:
return Error
elif edge not in prelim_distances or dist < prelim_distances[edge]:
prelim_distances[edge] = dist
nodes[edge] = vertex
return (distances, nodes)
def shortest_path(self, start, end, parameter):
"""
Find a single shortest path from the given start vertex
to the given end vertex using Dijkstra's algorithm.
Operates on this Network's graph, constructed using getGraph()
parameter = one of ["latency" | "cost"]
"""
# Run Dijkstra.
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
E = {}
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end:
break
edges = self.get_edges(v)
for edge in edges:
w = edge.sink
if parameter == "latency":
vwLength = D[v] + edge.latency
elif parameter == "cost":
vwLength = D[v] + edge.cost
if w in D:
if vwLength < D[w]:
raise ValueError("Dijkstra: found better path to already-final vertex")
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
E[w] = edge
# Get the shortest path.
Path = []
walker = end
while 1:
Path.append(walker)
if walker == start: break
try:
walker = P[walker]
except KeyError:
return None
Path.reverse()
# check sanity.
if Path[0] != start:
dprint("Path[0] is %s, start is %s", (Path[0], start))
raise InvalidPath
if Path[-1] != end:
dprint("Path[-1] is %s, end is %s", (Path[-1], end))
raise InvalidPath
# return Path
E[start] = None
return [E[s] for s in Path]
def Dijkstra(G,start,end=None,option="length"):
"""
Modified from David Eppstein's codes.
G: graph object form snap.py
start: nid of the start node (int)
end: nid of the end node (int)
option can be "length", "time" and "prob"
Find shortest paths from the start vertex to all
vertices nearer than or equal to the end.
The output is a pair (D,P) where D[v] is the distance
from start to v and P[v] is the predecessor of v along
the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly
when all edge lengths are positive. This code does not
verify this property for all edges (only the edges seen
before the end vertex is reached), but will correctly
compute shortest paths even for some graphs with negative
edges, and will raise an exception if it discovers that
a negative edge has caused it to make a mistake.
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
NI = G.GetNI(v)
for Id in NI.GetOutEdges():#w in G[v]:
eid = G.GetEId(v, Id)
#if eid == -1: raise
if option == "length":
edgeweight = 1
elif option == "time":
edgeweight = G.GetIntAttrDatE(eid, "AveTime")
elif option == "prob":
prob = G.GetFltAttrDatE(eid, "Prob")
if prob <= 0.001:
prob = 0.001
edgeweight = - log(prob)
vwLength = D[v] + edgeweight#G[v][w]
if Id in D:
if vwLength < D[Id]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif Id not in Q or vwLength < Q[Id]:
Q[Id] = vwLength
P[Id] = v
return (D,P)
def Dijkstra(G,start,end=None):
"""
Find shortest paths from the start vertex to all vertices nearer than or equal to the end.
The input graph G is assumed to have the following representation:
A vertex can be any object that can be used as an index into a dictionary.
G is a dictionary, indexed by vertices. For any vertex v, G[v] is itself a dictionary,
indexed by the neighbors of v. For any edge v->w, G[v][w] is the length of the edge.
This is related to the representation in <http://www.python.org/doc/essays/graphs.html>
where Guido van Rossum suggests representing graphs as dictionaries mapping vertices
to lists of outgoing edges, however dictionaries of edges have many advantages over lists:
they can store extra information (here, the lengths), they support fast existence tests,
and they allow easy modification of the graph structure by edge insertion and removal.
Such modifications are not needed here but are important in many other graph algorithms.
Since dictionaries obey iterator protocol, a graph represented as described here could
be handed without modification to an algorithm expecting Guido's graph representation.
Of course, G and G[v] need not be actual Python dict objects, they can be any other
type of object that obeys dict protocol, for instance one could use a wrapper in which vertices
are URLs of web pages and a call to G[v] loads the web page and finds its outgoing links.
The output is a pair (D,P) where D[v] is the distance from start to v and P[v] is the
predecessor of v along the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly when all edge lengths are positive.
This code does not verify this property for all edges (only the edges examined until the end
vertex is reached), but will correctly compute shortest paths even for some graphs with negative
edges, and will raise an exception if it discovers that a negative edge has caused it to make a mistake.
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # estimated distances of non-final vertices
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
try:
for w in G[v]:
if w == None:
print v, G[v]
assert(0)
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
except:
#print len(Q), v, D[v]
pass
return (D,P)
def compute_optimal_steiner(G, in_arcs, out_arcs, U, max_dim):
M = {}
W = []
W_it = powerset(U)
for el in W_it:
W.append(set(el))
#Each subset is identified by its index in W
for i in range(G.shape[0]):
for w in W:
M[(i, W.index(w))] = float('inf')
M[(i, W.index(set([])))] = 0
for u in U:
M[(u, W.index(set([u])))] = 0
for w in W:
Q_w = priorityDictionary()
for i in range(G.shape[0]):
M[(i, W.index(w))] = min(map(lambda x: M[(i, W.index(set(x)))] + M[(i, W.index(w - set(x)))], powerset(w)))
Q_w[i] = M[(i, W.index(w))]
for v_min in Q_w:
for arc in in_arcs[v_min]:
v = arc[0]
if(M[(v, W.index(w))] > M[(v_min, W.index(w))] + 1):
M[(v, W.index(w))] = M[(v_min, W.index(w))] + 1
Q_w[v] = M[(v, W.index(w))]
#I can prune the search here if, for a given subset, the size of the minimum tree exceeds max_dim
below_bound = False
for i in range(G.shape[0]):
if(M[(i, W.index(w))] <= max_dim):
below_bound = True
break
if(not(below_bound)):
return []
root = 0
for v in range(1, G.shape[0]):
if (M[(v, W.index(set(U)))] < M[(root, W.index(set(U)))]):
root = v
solution = construct_tree(root, M, out_arcs, W, set(U))
#print "W:"
#print W
#print "M:"
#print M
return solution
def _shortest_path_to(self, dest):
"""Returns the next exit to the shortest path from self to dest
and the distance of the shortest path from self to dest."""
# TODO: remove the duplicate exits in the graph
if dest is self:
return None, 0
## if not dest.exits: # small optimization
## return None, None # no path exists
# add start and end to the graph
G = self.world.g
for v in (self, dest):
G[v] = {}
for e in v.exits:
G[v][e] = G[e][v] = int_distance(v.x, v.y, e.x, e.y)
start = self
end = dest
# apply Dijkstra's algorithm (with priority list)
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = (0,)
for v in Q:
D[v] = Q[v][0]
if v == end:
break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
pass
elif w not in Q or vwLength < Q[w][0]:
Q[w] = (vwLength, int(w.id)) # the additional value makes the result "cross-machine deterministic"
P[w] = v
# restore the graph
for v in (start, end):
del G[v]
for e in v.exits:
del G[e][v]
# exploit the results
if end not in P:
return None, None # no path exists
Path = []
while 1:
Path.append(end)
if end == start:
break
end = P[end]
Path.reverse()
return Path[1], D[dest]
def Dijkstra(G,start,end=None):
"""
Find shortest paths from the start vertex to all vertices nearer than or equal to the end.
The input graph G is assumed to have the following representation:
A vertex can be any object that can be used as an index into a dictionary.
G is a dictionary, indexed by vertices. For any vertex v, G[v] is itself a dictionary,
indexed by the neighbors of v. For any edge v->w, G[v][w] is the length of the edge.
This is related to the representation in <http://www.python.org/doc/essays/graphs.html>
where Guido van Rossum suggests representing graphs as dictionaries mapping vertices
to lists of outgoing edges, however dictionaries of edges have many advantages over lists:
they can store extra information (here, the lengths), they support fast existence tests,
and they allow easy modification of the graph structure by edge insertion and removal.
Such modifications are not needed here but are important in many other graph algorithms.
Since dictionaries obey iterator protocol, a graph represented as described here could
be handed without modification to an algorithm expecting Guido's graph representation.
Of course, G and G[v] need not be actual Python dict objects, they can be any other
type of object that obeys dict protocol, for instance one could use a wrapper in which vertices
are URLs of web pages and a call to G[v] loads the web page and finds its outgoing links.
The output is a pair (D,P) where D[v] is the distance from start to v and P[v] is the
predecessor of v along the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly when all edge lengths are positive.
This code does not verify this property for all edges (only the edges examined until the end
vertex is reached), but will correctly compute shortest paths even for some graphs with negative
edges, and will raise an exception if it discovers that a negative edge has caused it to make a mistake.
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # estimated distances of non-final vertices
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end:
break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def MST_Kruskal(graph):
"""
An implementation of Kruskal and Prim's Minimum Spanning tree algorithm - implemented after the description in Cormen
"""
result_graph = {}
set_container = {}
# Initially the nodes only know themselves
for node in graph:
set_container[node] = [node]
result_graph[node] = {}
# we make the priority queue
Q = priorityDictionary()
# We make sure the edge is only there once
already_there = {}
for outer_node in graph:
inner_nodes = graph[outer_node]
for inner_node in inner_nodes:
if already_there.get((inner_node, outer_node)):
continue
already_there[(inner_node, outer_node)] = 1
already_there[(outer_node, inner_node)] = 1
Q[(outer_node, inner_node)] = inner_nodes[inner_node]
for edge in Q:
p1 = edge[0]
p2 = edge[1]
if set_container[p1] != set_container[p2]:
result_graph[p1][p2] = Q[edge]
result_graph[p2][p1] = Q[edge]
set1 = set_container[p1]
set2 = set_container[p2]
for node in set2:
if not node in set1:
set1.append(node)
for node in set1:
set_container[node] = set1
return (result_graph, set_container)
def MST_Kruskal(graph):
"""
An implementation of Kruskal and Prim's Minimum Spanning tree algorithm - implemented after the description in Cormen
"""
result_graph = {}
set_container = {}
# Initially the nodes only know themselves
for node in graph:
set_container[node] = [node]
result_graph[node] = {}
# we make the priority queue
Q = priorityDictionary()
# We make sure the edge is only there once
already_there = {}
for outer_node in graph:
inner_nodes = graph[outer_node]
for inner_node in inner_nodes:
if already_there.get((inner_node, outer_node)):
continue
already_there[(inner_node, outer_node)] = 1
already_there[(outer_node, inner_node)] = 1
Q[(outer_node, inner_node)] = inner_nodes[inner_node]
for edge in Q:
p1 = edge[0]
p2 = edge[1]
if set_container[p1] != set_container[p2]:
result_graph[p1][p2] = Q[edge]
result_graph[p2][p1] = Q[edge]
set1 = set_container[p1]
set2 = set_container[p2]
for node in set2:
if not node in set1:
set1.append(node)
for node in set1:
set_container[node] = set1
return (result_graph, set_container)
def priorityGraph(self, graph=None):
"""
return a priority graph. Each sub dictionnary of the graph is a
priority dictionnary as defined in priorityDictionary
"""
if graph == None:
G = self.graph
else:
G = graph
Gp = {}
for n1 in G.keys():
d = priorityDictionary()
for n2 in G[n1].keys():
d[n2] = G[n1][n2]
Gp[n1] = d
return Gp
def priorityGraph(self, graph=None):
"""
return a priority graph. Each sub dictionnary of the graph is a
priority dictionnary as defined in priorityDictionary
"""
if graph == None:
G = self.graph
else:
G = graph
Gp = {}
for n1 in G.keys():
d = priorityDictionary()
for n2 in G[n1].keys():
d[n2] = G[n1][n2]
Gp[n1] = d
return Gp
def __init__(self,start,goal,state_trans,forbidden=set()):
self.OPEN = priorityDictionary()
self.INCONS= set()
self.CLOSED = set()
self.s_start = State(start,start,goal)
self.s_start.set_rhs(constants.INF)
self.s_start.set_g(constants.INF)
self.s_goal = State(goal,start,goal)
self.s_goal.set_g(constants.INF)
self.s_goal.set_rhs(0)
self.eps = 2.5 # TODO make it change during iterations.
self.PREC = {}
self.G = {}
self.path_= {}
self.G[self.s_start] = self.s_start
self.forbidden = forbidden#TODO add "obstacles"
self.OPEN[self.s_goal] = self.keys(self.s_goal)
self.state_trans= state_trans
def Dijkstra(G, start, end=None):
D = {}
P = {}
Q = priorityDictionary()
Q[start] = G[start[0]][start[1]][0]
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v[0]][v[1]][1]:
vwLength = D[v] + G[w[0]][w[1]][0]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return D[v]
def dijkstra(G, source):
Q = priorityDictionary()
prev = {}
distances = {}
Q[source] = 0.0
for v in Q:
distances[v] = Q[v]
if v == None:
break
for w in G.matrix[v]:
temp = distances[v] + G.matrix[v][w]
if w in distances:
if temp < distances[w]:
raise ValueError("Error")
elif w not in Q or temp < Q[w]:
Q[w] = temp
prev[w] = v
return (distances, prev)
def Dijkstra(G, start, end=None):
"""Find shortest paths from the start vertex to all vertices nearer than or equal to the end.
The input graph G is assumed to have the following representation:
A vertex can be any object that can be used as an index into a dictionary.
G is a dictionary, indexed by vertices. For any vertex v, G[v] is itself a dictionary, indexed by
the neighbors of v. For any edge v->w, G[v][w] is the length of the edge.
Of course, G and G[v] need not be actual Python dict objects, they can be any other type of object
that obeys dict protocol, for instance one could use a wrapper in which vertices are URLs of web pages
and a call to G[v] loads the web page and finds its outgoing links.
The output is a pair (D, P) where D[v] is the distance from start to v and P[v] is the predecessor
of v along the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly when all edge lengths are positive.
This code does not verify this property for all edges (only the edges examined until the end vertex
is reached), but will correctly compute shortest paths even for some graphs with negative edges, and
will raise an exception if it discovers that a negative edge has caused it to make a mistake.
"""
D = {} # dictionary of real shortest distances
P = {} # dictionary of predecessors
Q = priorityDictionary(
) # estimated distances from "unsettled" or "unprocessed" vertices
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end:
break
for w in G[v]:
vw_dist = D[v] + G[v][w]
if w in D:
if vw_dist < D[w]:
raise ValueError(
"Dijkstra: found better path to already-processed vertex"
)
elif w not in Q or vw_dist < Q[w]:
Q[w] = vw_dist
P[w] = v
return D, P
def Dijkstra(G, start, end=None):
D = {}
P = {}
Q = priorityDictionary()
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def Dijkstra(G,s,e=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # estimated distances of non-final vertices
Q[s] = 0
for v in Q:
D[v] = Q[v]
if v == e: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def Dijkstra(G,start,end=None):
D = {}
P = {}
Q = priorityDictionary()
Q[start] = G[start[0]][start[1]][0]
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v[0]][v[1]][1]:
vwLength = D[v] + G[w[0]][w[1]][0]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return D[v]
def Dijkstra(G,start,end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
print(G)
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
print("valueerror")
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def Dijkstra(G,start,end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def djsktra(G, start, end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary()
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def route(G,starts,ends=[]):
"""
Build the shortest path tree in G from starts. Stops when any vertex in 'ends' is met. Uses Dijkstra's algorithm.
@param G: graph
@param starts: singleton or list of starting node ids
@param ends: singleton or list of ending node ids
@returns: a tuple (D,P,v)
D: dictionary of distances (from starts)
P: dictionary of predecessors (encoding the tree)
v: final vertex (useful to determine which vertex of ends we reached)
Code modified from the PADS library.
http://www.ics.uci.edu/~eppstein/PADS/
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary()
if type(starts)==list:
for s in starts:
Q[s]=0
else:
Q[starts]=0
if not type(ends)==list: ends=[ends]
for v in Q:
D[v] = Q[v]
if v in ends: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
print v,w,vwLength,D[v],G[v][w],D[w]
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P,v)
def Dijkstra(Graph,ilk,son=None):
uzaklik={}#son uzakliklari tutacak
gelis={}#noktaya gelinen yeri tutar.
nodedist = priorityDictionary()#nodların yaklasik hesaplamasini tutar
nodedist[ilk]=0
for node in nodedist:
uzaklik[node] = nodedist[node]
if node == son:
break
for node2 in Graph[node][1]:
sonrasininUzakligi = uzaklik[node] + Graph[node][1][node2]
if node2 in uzaklik:
if sonrasininUzakligi < uzaklik[node2]:
raise ValueError
elif node2 not in nodedist or sonrasininUzakligi < nodedist[node2]:
nodedist[node2] = sonrasininUzakligi
gelis[node2] = node
return (uzaklik,gelis)
def Dijkstra(graph, start_node, end_node=None):
distances = {} # dictionary of final distances
predecessors = {} # dictionary of predecessors
priority_dictionary = priorityDictionary(
) # estimated distances of non-final vertices
priority_dictionary[start_node] = 0.0
for vertex in priority_dictionary:
distances[vertex] = priority_dictionary[vertex]
if vertex == end_node:
break
for w in graph[vertex]:
vwLength = float(
format(distances[vertex] + graph[vertex][w], ".2f"))
if w in distances:
if vwLength < distances[w]:
pass
elif w not in priority_dictionary or vwLength < priority_dictionary[
w]:
priority_dictionary[w] = vwLength
predecessors[w] = vertex
return (distances, predecessors)
def Dijkstra(g,start,end=None):
d = {} # dictionary of final distances
p = {} # dictionary of predecessors
q = priorityDictionary() # est.dist. of non-final vert.
q[start] = 0
for v in q:
d[v] = q[v]
if v == end: break
for w in g[v]:
vwLength = d[v] + g[v][w]
if w in d:
if vwLength < d[w]:
raise ValueError
elif w not in q or vwLength < q[w]:
q[w] = vwLength
p[w] = [v]
elif w not in q or vwLength == q[w]:
q[w] = vwLength
p[w] += [v]
return (d,p)
def Dijkstra(self, start, end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end:
break
for w in self.neighbours[v]:
vwLength = D[v] + self.edge_weights[(v, w)]
if w in D:
if vwLength < D[w]:
error = "found, better path to already-final vertex"
raise ValueError, error
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def Dijkstra(G,start,end=None):#TODO test that if there is no way between start and end???
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def shortestPaths(nodes, edges, startDist, initialNodes):
'''simple shortest paths algorithm, using advanced data structure. calculates
all distances from the starting set of nodes to all other nodes. returns
nodes to distance mapping. nodes is a list, edges is a dict from nodes to
other nodes with distance as a tuple, startdist is a float, initialnodes
is a list.'''
currentNodes = priorityDictionary() #holds data on nodes left to process
nodeDist = {}
for initialNode in initialNodes:
currentNodes[initialNode] = startDist
while len(currentNodes) > 0:
currentNode = currentNodes.smallest()
lastDist = currentNodes.pop(currentNodes.smallest())
if currentNode not in nodeDist or lastDist < nodeDist[currentNode]:
#update the dist, add neighbors to heap
nodeDist[currentNode] = lastDist
for neighborNode,nbDist in edges[currentNode]:
newDist = lastDist + nbDist
if neighborNode not in currentNodes or \
newDist <= currentNodes[neighborNode]:
currentNodes[neighborNode] = newDist #updates prio dict
return nodeDist
def Dijkstra(G, start, end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
n = len(G)
for v in Q:
D[v] = Q[v]
if v == end: break
for w in range(n):
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def Dijkstra(G,start,end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
vwLength=0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
print("wielgachny wynik")
raise ValueError
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
# print('Dijkstra:')
# print(D)
return D,P
def UnitDijkstra(G,start,end=None):
"""
Same as the other, now just with unit distances
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for w in G[v]:
vwLength = D[v] + 1
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D,P)
def Dijkstra(graph,start,end=None):
final_distances = {} # dictionary of final distances
predecessors = {} # dictionary of predecessors
estimated_distances = priorityDictionary() # est.dist. of non-final vert.
estimated_distances[start] = 0
for vertex in estimated_distances:
final_distances[vertex] = estimated_distances[vertex]
if vertex == end: break
for edge in graph[vertex]:
path_distance = final_distances[vertex] + graph[vertex][edge]
if edge in final_distances:
if path_distance < final_distances[edge]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif edge not in estimated_distances or path_distance < estimated_distances[edge]:
estimated_distances[edge] = path_distance
predecessors[edge] = vertex
return (final_distances,predecessors)
def Dijkstra(G, start, end=None):
min_max = float('inf')
for s in start:
D = {}
P = {}
Q = priorityDictionary()
Q[s] = G[s[0]][s[1]][0]
for v in Q:
D[v] = Q[v]
if v in end: break
for w in G[v[0]][v[1]][1]:
vwLength = D[v] + G[w[0]][w[1]][0]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
if D[v] < min_max:
min_max = D[v]
return min_max
def Dijkstra(G, start, end=None):
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # estimated distances of non-final vertices
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end:
break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError(
"Dijkstra: se encotro un mejor camino al nodo destino."
)
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def Dijkstra(G, start, end=None):
#Init
D = {} #Dctionary of final distances
P = {} #Dictionary of predecessors
Q = priodict.priorityDictionary()
Q[start] = 0
#While there are elements in the priority queue
for v in Q:
D[v] = Q[v]
if v == end:
break
for w in G[v]:
vwLength = D[v] + G[v][w]
if w in D:
if vwLength < D[w]:
raise ValueError
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
return (D, P)
def Dijkstra(G,start,end=None):
# dict of final distances
D = {}
# dict of predecessors
P = {}
# estimated distances of non-final vertices
estimates = priorityDictionary()
estimates[start] = 0
for vert in estimates:
D[vert] = estimates[vert]
if vert == end:
break
for w in G[vert]:
length = D[vert] + G[vert][w]
if w in D:
if length < D[w]:
raise ValueError("Dijkstra: found better path to already-final vertex")
elif w not in estimates or length < estimates[w]:
estimates[w] = length
P[w] = vert
return (D,P)
def Dijkstra(G,start,end=None):
min_max = float('inf')
for s in start:
D = {}
P = {}
Q = priorityDictionary()
Q[s] = G[s[0]][s[1]][0]
for v in Q:
D[v] = Q[v]
if v in end: break
for w in G[v[0]][v[1]][1]:
vwLength = D[v] + G[w[0]][w[1]][0]
if w in D:
if vwLength < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwLength < Q[w]:
Q[w] = vwLength
P[w] = v
if D[v] < min_max:
min_max = D[v]
return min_max
def dijkstra(map, start, end):
if start == end:
return 0
if not (map.has_key(start)) or not (map.has_key(end)):
return -1
distances = {}
nodes = {}
prelim_distances = priorityDictionary()
prelim_distances[start] = 0
for vertex in prelim_distances:
distances[vertex] = prelim_distances[vertex]
if vertex == end: break
for edge in map[vertex]:
dist = distances[vertex] + map[vertex][edge]
if edge in distances:
if dist < distances[edge]:
return Error
elif edge not in prelim_distances or dist < prelim_distances[edge]:
prelim_distances[edge] = dist
nodes[edge] = vertex
return (distances, nodes)
def correctgraph(self,conn):
cur=conn.cursor()
nodeDict=priorityDictionary()
for node in self.edges:
if len(self.edges[node])<=2:
nodeDict[node]=min(map(lambda x:x.sectId, self.edges[node]))
print len(nodeDict)
count=0
visited=[]
for node in nodeDict:
count+=1
#~ print count
cur.execute("update network_dumppoints_copy set split_id=%s \
where st_startpoint(geomline)='%s' or st_endpoint(geomline)='%s'" % (nodeDict[node],node.geom,node.geom))
for nbr in self.adj(node):
if nbr.u==node:
if nbr.v not in visited and nbr.v in nodeDict:
nodeDict[nbr.v]=nodeDict[node]
else:
if nbr.u not in visited and nbr.u in nodeDict:
nodeDict[nbr.u]=nodeDict[node]
visited.append(node)
conn.commit()
def Dijkstra(graph,start,end=None):
distances = {} # dictionary of final distances
predec = {} # dictionary of predecessors
queue = priorityDictionary() # est.dist. of non-final vert.
queue[start] = 0
def shortest_path_from_pts_to_intensity_dijk(point_0_s, point_1_int, dat, dist_func=lambda x:1.0/x, min_thr=0, max_thr=None, dist_diff=None, thr_distance=100000, zero_distance=None):
'''
find shortest path between point_0_s and points having intensity point_1_int in a 3d array dat.
shortest_path_serial_dijk(point_0_s, point_1_int, dat, dist_func=lambda x:1.0/x, min_thr=0, max_thr=None, dist_diff=None, thr_distance=100000, zero_distance=None):
point_0_s: list of starting points [(x, y, z), ...]
point_1_int: ending point intensity intensity
dat: numpy 3d array
dist_func = lambda x:1.0/x: distance between a and b := euclidean_distance(a, b) * dist_func(b)
min_thr = 0: ignore voxels having intensity < min_thr
max_thr = 0: ignore voxels having intensity > max_thr
dist_diff = None: if not None, distance between a and b := euclidean_distance(a, b) * dist_diff(a, b)
thr_distance = 100000: the distance from/to the thresholded intensity voxels
zero_distance = None: if not None, voxels having intensity zero_distance having distance 0 to their neighbors
'''
nx, ny, nz = dat.shape
d1 = 1.0
d2 = np.sqrt(2)
d3 = np.sqrt(3)
nbd_sets = [ (nbd_d1, d1), (nbd_d2, d2), (nbd_d3, d3) ]
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
for point_0 in point_0_s:
Q[point_0] = 0
for v in Q:
D[v] = Q[v]
if dat[v] == point_1_int:
point_1 = v
break
for nbd_d, dist_base in nbd_sets:
for nbd in get_nbds_bdchk(v, nbd_d, nx, ny, nz):
if dat[nbd] == point_1_int:
#print 'found point_2_int=%s' % point_1_int
#print len(Q)
distance = 0.01
elif (zero_distance is not None and dat[nbd] == zero_distance):
distance = 0.01
elif (min_thr is not None and dat[nbd] <= min_thr) or (max_thr is not None and max_thr <= dat[nbd]):
distance = thr_distance
elif dist_diff is not None:
distance = dist_base * dist_diff(dat[nbd], dat[v])
else:
distance = dist_base * dist_func(dat[nbd])
if distance < 0.0:
print 'distance < 0'
return []
vwLength = D[v] + distance
if nbd in D:
if vwLength < D[nbd]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif nbd not in Q or vwLength < Q[nbd]:
#if nbd not in Q or vwLength < Q[nbd]:
Q[nbd] = vwLength
P[nbd] = v
Path = []
end = point_1
start = point_0_s
while 1:
Path.append(end)
if end in start: break
end = P[end]
if len(Path) > 40:
break
Path.reverse()
return Path
def a_star(position,angle, map, previous_start, previous_policy, depth=float('inf')):
# start:(int, int),map:MATRIX,
# goals:coordinates of all goals
#position of current position(start), and apple (goal)
#TODO:
# Check if the action of start state has been calculated
# return 0
goals_location = helper.object_position(Constants.GOAL_TYPE)
goals_states = [helper.positionTOstate(*p) for p in goals_location]
# for i in range(len(map)):
# for j in range(len(map[i])):
# if map[i][j] == 'g':
# goals_states.append((i, j))
# print "goals_location", goals_location
# print 'goals_states', goals_states
start = helper.positionTOstate(position[0],position[1])
if start in previous_policy:
at = previous_policy.index(start)
if at != len(previous_policy):
previous_policy = previous_policy[at:]
print 'previous policy:', previous_policy
print 'next step', start, 'to', previous_policy[1]
return angle_between_position(position,
helper.stateTOposition(*previous_policy[1]))
else:
previous_policy = []
unvisted = priorityDictionary(sort_by = lambda x: x[1])
unvisted[start] = 0
gScores = defaultdict(lambda : float('inf')) # cost from start to state
fScores = defaultdict(None) # ~cost from state to goal
pred = defaultdict(lambda : (float('inf'), float('inf')))
visited = set()
gScores[start] = 0
for curr in unvisted:
visited.add(curr)
# fScores[curr] = heuristic(curr, goals_states, map, angle) + gScores[curr]
# if curr in goals_states:
# break
# add positions arround curr that are not visited
for sur in get_surrounding(curr, map):
if distance(sur, start, angle, map) > depth:
continue
if curr == start:
dist = distance(curr, sur, angle, map)
else:
dist = distance(curr, sur,
angle_between_state(pred[curr], curr, map),
map)
if sur in visited and\
gScores[sur] < gScores[curr] + dist:
continue
# calculate f values for new-added blocks
# if gScores[sur]+heuristic(sur, goals_states, map, angle) > \
# gScores[curr]+heuristic(curr, goals_states, map, angle):
else:
pred[sur] = curr
gScores[sur] = gScores[curr] + dist
fScores[sur] = heuristic(sur, goals_states, map, angle) + gScores[sur]
if sur not in visited:
unvisted[sur] = fScores[sur]
# helper.print_dict(gScores, name='gScore')
# helper.print_dict(fScores, name='fScore')
#find the angle of close_list
dest = start
minF = float('inf')
for state, f in fScores.items():
if f < minF and state != start:
minF = f
dest = state
previous_policy = retreive_path(start, dest, pred)
next_block = previous_policy[1]
print 'facing', angle, start, 'to', previous_policy[-1], ':', previous_policy
print 'next step', start, 'to', next_block, '==', position, 'to', helper.stateTOposition(*next_block)
return angle_between_position(position,
helper.stateTOposition(*next_block))
w = [[0 for i in xrange(M)] for j in xrange(N)]
t = [[0 for i in xrange(M)] for j in xrange(N)]
for i in xrange(N):
inp = map(int, raw_input().split())
for j in xrange(M):
s[i][j] = inp[3 * j]
w[i][j] = inp[3 * j + 1]
t[i][j] = inp[3 * j + 2] % (s[i][j] + w[i][j])
s.reverse()
w.reverse()
t.reverse()
start = (0, 0)
end = (2 * N - 1, 2 * M - 1)
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
for u, l in getnhd(v):
##print l,u
vuLength = D[v] + l
if u in D:
if vuLength < D[u]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif u not in Q or vuLength < Q[u]:
Q[u] = vuLength
The output is a pair (D,P) where D[v] is the distance
from start to v and P[v] is the predecessor of v along
the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly
when all edge lengths are positive. This code does not
verify this property for all edges (only the edges seen
before the end vertex is reached), but will correctly
compute shortest paths even for some graphs with negative
edges, and will raise an exception if it discovers that
a negative edge has caused it to make a mistake.
"""
D = {} # dictionary of final distances
P = {} # dictionary of predecessors
Q = priorityDictionary() # est.dist. of non-final vert.
Q[start] = 0
for v in Q:
D[v] = Q[v]
if v == end: break
curnode = ncollection.find_one({'name':v})
availablenodes = curnode['availableNodes']
for node in availablenodes:
nnode = ncollection.find_one({'name':node})
vwLength = D[v] + nodeDist(curnode, nnode, time.time()+D[node]
if node in D:
if vwLength < D[node]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
def Dijkstra(graph,start,end=None):
"""
Find shortest paths from the start vertex to all
vertices nearer than or equal to the end.
The input graph G is assumed to have the following
representation: A vertex can be any object that can
be used as an index into a dictionary. G is a
dictionary, indexed by vertices. For any vertex v,
G[v] is itself a dictionary, indexed by the neighbors
of v. For any edge v->w, G[v][w] is the length of
the edge. This is related to the representation in
<http://www.python.org/doc/essays/graphs.html>
where Guido van Rossum suggests representing graphs
as dictionaries mapping vertices to lists of neighbors,
however dictionaries of edges have many advantages
over lists: they can store extra information (here,
the lengths), they support fast existence tests,
and they allow easy modification of the graph by edge
insertion and removal. Such modifications are not
needed here but are important in other graph algorithms.
Since dictionaries obey iterator protocol, a graph
represented as described here could be handed without
modification to an algorithm using Guido's representation.
Of course, G and G[v] need not be Python dict objects;
they can be any other object that obeys dict protocol,
for instance a wrapper in which vertices are URLs
and a call to G[v] loads the web page and finds its links.
The output is a pair (D,P) where D[v] is the distance
from start to v and P[v] is the predecessor of v along
the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly
when all edge lengths are positive. This code does not
verify this property for all edges (only the edges seen
before the end vertex is reached), but will correctly
compute shortest paths even for some graphs with negative
edges, and will raise an exception if it discovers that
a negative edge has caused it to make a mistake.
"""
final_distances = {} # dictionary of final distances
predecessors = {} # dictionary of predecessors
estimated_distances = priorityDictionary() # est.dist. of non-final vert.
estimated_distances[start] = 0
for vertex in estimated_distances:
final_distances[vertex] = estimated_distances[vertex]
if vertex == end: break
for edge in graph[vertex]:
path_distance = final_distances[vertex] + graph[vertex][edge]
if edge in final_distances:
if path_distance < final_distances[edge]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif edge not in estimated_distances or path_distance < estimated_distances[edge]:
estimated_distances[edge] = path_distance
predecessors[edge] = vertex
return (final_distances,predecessors)
def getRoute(self, startEdge, endEdge, algorithm):
# if startEdge is internal
if startEdge[0] == ':':
return None
# applies dijkstra's algorithm
# to find the lightest route (list of edges) between the startEdge to the endEdge
# in the self.__graph network
D = {}
P = {}
Q = priorityDictionary()
Q[startEdge] = 0
# alters the edge weights if the error insertion is the chosen algorithm
alteredEdgeWeights = {}
if algorithm == 3:
for edge, weight in self.__edgeWeights.iteritems():
newWeight = weight * WEIGHT_VARIATION
alteredEdgeWeights[edge] = weight + random.uniform(-newWeight, newWeight)
for v in Q:
D[v] = Q[v]
if v == endEdge:
break
for w in self.__graph[v]:
vwValue = 0
# DistanceDijkstra - algorithm = 0
if algorithm == 0:
vwValue = D[v] + self.__graph[v][w]
# OccupancyDijsktra - algorithm = 1
elif algorithm == 1:
vwValue = D[v] + self.__graph[v][w] + pow((self.__graph[v][w] * self.__edgeWeights[w]), 2)
# RoadBlock - algorithm = 2
elif algorithm == 2:
if(self.__edgeWeights[w]>OCCUPANCY_LIMIT):
break
vwValue = D[v] + self.__graph[v][w]
# ErrorInsertion - algorithm = 3
elif algorithm == 3:
vwValue = D[v] + self.__graph[v][w] + pow((self.__graph[v][w] * alteredEdgeWeights[w]), 2)
if w in D:
if vwValue < D[w]:
raise ValueError, "Dijkstra: found better path to already-final vertex"
elif w not in Q or vwValue < Q[w]:
Q[w] = vwValue
P[w] = v
# computes the route from the dictionary with the preceding edges (P)
try:
route = []
toEdge = endEdge
while endEdge != startEdge:
endEdge = P[endEdge]
route.insert(0, toEdge)
toEdge = endEdge
route.insert(0, toEdge)
except:
# catches the no route found exception
# for example, caused by road blocks when algorithm = 2
return []
return route
def a_star(self, sources, old_target):
'''
Implementation of multisource A* as mentioned in
practical heuristic for minor embedding paper.
multisource A*
Input: graph G, edge weights , heuristic costs , sources
Output: vertex cv such that maxk i=1 d(cv, s(i)) is
minimal among all vertices
'''
# Initializing variables:
d = {}
est = {}
reached = {}
min_est = priorityDictionary()
min_src = {}
path = []
path_cs = []
inf = len(self.original_nodes_list )**2*\
len(self.edges)*10**4 # a very large number
for v in self.nodes:
for i in range(len(sources)):
# d[(v,i)] = inf # best known distance from v to i
# est[(v,i)] = inf # heuristic distance
# reached[(v,i)] = False # node v reached from source i?
d[(v, sources[i])] = inf # best known distance from v to i
est[(v, sources[i])] = inf # heuristic distance
reached[(v,
sources[i])] = False # node v reached from source i?
for v in self.nodes:
min_est[v] = inf # min. dist. to v among all i
for i in range(len(sources)):
# This is dangereous:
d[(sources[i], sources[i])] = 0
min_est[sources[i]] = 0
#min_src[sources[i]] = i #index of source for min_est
# I am trying to use the correct indecies for sources:
min_src[sources[i]] = sources[i] #index of source for min_est
while True:
#current node calculation:
if path != []:
cv = self.neighbours[path[-1]][0]
else:
cv = self.neighbours[sources[0]][0]
for v in self.nodes:
if min_est[v] < min_est[cv]:
cv = v
#current source calculation:
cs = min_src[cv]
reached[(cv, cs)] = True
# Check if all sources have reached cv:
check_counter = 0
for i in range(len(sources)):
if (cv, sources[i]) in reached.keys():
if reached[(cv, sources[i])] == True:
if cv not in sources:
if cv not in self.neighbours[sources[i]]:
check_counter += 1
if check_counter == len(sources):
return cv # all sources have reached cv
# Find new best source for cv:
found_minsrc = sources[0]
for i in range(len(sources)):
if reached[(cv, sources[i])] == False:
if est[(cv, sources[i])] < est[(cv, found_minsrc)]:
found_minsrc = sources[i]
min_src[cv] = found_minsrc
# Find new best distance for cv:
min_est[cv] = est[(v, min_src[cv])]
# update neighbour distances:
for v in self.neighbours[cv]:
# alternate distance to cs
alt = d[(cv, cs)] + self.edge_weights[(v, cv)]
if alt < d[(v, cs)]:
d[(v, cs)] = alt #distance improved
# calculate heuristic cost for vertex v to old target g*:
h_cost_v = len(self.shortestPath(v, old_target))
#est[(v,cs)] = d[(v,cs)] + h_cost[v] #new heuristic distance
est[(v, cs)] = d[(v, cs)] + h_cost_v
if est[(v, cs)] < min_est[v]:
min_est[v] = est[(v, cs)] #improved best distance
min_src[v] = cs
def __init__(self, maxsize, origitems=None):
self.count = 0
self.maxsize = maxsize
self.priority = priorityDictionary()
self.drop_listeners = []
OrderedSet.__init__(self, origitems)
def Dijkstra(graph, start, end=None):
"""
Find shortest paths from the start vertex to all
vertices nearer than or equal to the end.
The input graph G is assumed to have the following
representation: A vertex can be any object that can
be used as an index into a dictionary. G is a
dictionary, indexed by vertices. For any vertex v,
G[v] is itself a dictionary, indexed by the neighbors
of v. For any edge v->w, G[v][w] is the length of
the edge. This is related to the representation in
<http://www.python.org/doc/essays/graphs.html>
where Guido van Rossum suggests representing graphs
as dictionaries mapping vertices to lists of neighbors,
however dictionaries of edges have many advantages
over lists: they can store extra information (here,
the lengths), they support fast existence tests,
and they allow easy modification of the graph by edge
insertion and removal. Such modifications are not
needed here but are important in other graph algorithms.
Since dictionaries obey iterator protocol, a graph
represented as described here could be handed without
modification to an algorithm using Guido's representation.
Of course, G and G[v] need not be Python dict objects;
they can be any other object that obeys dict protocol,
for instance a wrapper in which vertices are URLs
and a call to G[v] loads the web page and finds its links.
The output is a pair (D,P) where D[v] is the distance
from start to v and P[v] is the predecessor of v along
the shortest path from s to v.
Dijkstra's algorithm is only guaranteed to work correctly
when all edge lengths are positive. This code does not
verify this property for all edges (only the edges seen
before the end vertex is reached), but will correctly
compute shortest paths even for some graphs with negative
edges, and will raise an exception if it discovers that
a negative edge has caused it to make a mistake.
"""
final_distances = {} # dictionary of final distances
predecessors = {} # dictionary of predecessors
estimated_distances = priorityDictionary() # est.dist. of non-final vert.
estimated_distances[start] = 0
for vertex in estimated_distances:
final_distances[vertex] = estimated_distances[vertex]
if vertex == end: break
for edge in graph[vertex]:
path_distance = final_distances[vertex] + graph[vertex][edge]
if edge in final_distances:
if path_distance < final_distances[edge]:
raise ValueError, \
"Dijkstra: found better path to already-final vertex"
elif edge not in estimated_distances or path_distance < estimated_distances[
edge]:
estimated_distances[edge] = path_distance
predecessors[edge] = vertex
return (final_distances, predecessors)