class Instance(object):
"""
Instance class representing the TSP Instance
"""
def __init__(self, nodes, alpha, beta, decay, q):
# Make a list of nodes as variables which are JSONifiable
nodes_var = []
for node in nodes:
nodes_var.append(vars(node))
self.nodes = nodes_var
self.alpha = alpha
self.beta = beta
self.decay = decay
self.q = q
self.min_pheromone = 0.01
self.local_deposit = 0.1
self.distances = []
self.pheromones = []
colony = Colony()
self.ants = colony.ants
self.shortest_path = colony.shortest_path
self.min_distance = colony.min_distance
# Initialise the distances between nodes and pheromone trails
for i in range(len(nodes)):
distances = []
pheromones = []
for j in range(len(nodes)):
distances.append(0 if i == j else nodes[i].distance(nodes[j]))
pheromones.append(self.min_pheromone)
self.distances.append(distances)
self.pheromones.append(pheromones)
def get_path_distance(self, path):
"""
Returns the total total distance of a path taken
"""
length = len(path)
distance = 0
for i in range(length):
distance += self.distances[path[i]][path[(i + 1) % length]]
return distance
def update_pheromones(self, colony):
"""
Updates pheromones between nodes globally, a way of letting ants know on future
generations about the strongest paths to take.
"""
# Decay all pheromone trails
for i in range(len(self.nodes)):
for j in range(len(self.nodes)):
self.pheromones[i][j] *=(1- self.decay)
# Add to edge pheromones if edge was part of successful tour
for ant in colony.ants:
distance = self.get_path_distance(ant.path)
if distance <= colony.min_distance:
for i, j in ant.nodes_traversed():
self.pheromones[i][j] += self.q / distance
# Keep pheromone trails greater than or equal to 0.01, so nodes do not become
# completely unviable choices.
for i in range(len(self.nodes)):
for j in range(len(self.nodes)):
self.pheromones[i][j] = max(self.pheromones[i][j], self.min_pheromone)
def aco(self, gens, current_gen, client):
"""
Returns the generation reached and the shortest path found by the aco
algorithm along with its distance
"""
# The time at the start of the algorithm
time_start = time.time()
# Initalise the colony and its parameters
self.colony = Colony()
self.colony.ants = self.ants
self.colony.shortest_path = self.shortest_path
self.colony.min_distance = self.min_distance
# Initialise an array to be append with nodes
shortest_path = []
# Do generations from the current generation to the generation number needed
for i in range(current_gen, gens):
# The current time
time_now = time.time()
time_elapsed = time_now-time_start
# If exectutiion time has reached 25 seconds, return result
if (time_elapsed) > 25:
break
# Ants within colony perform their tours
self.colony.perform_tours(self)
# Get the shortest tour found by the ants
shortest_path = self.colony.shortest_path
# Global update of pheromones
self.update_pheromones(self.colony)
# Generation successful, thus increase the generation reached
gen_reached = i+1
# Update Instance parameters to be returned to client
self.shortest_path = shortest_path
self.min_distance = self.colony.min_distance
msg = "Generation " + str(i) + " distance " + str(round(self.colony.min_distance, 3)) + " path " + str(shortest_path)
# Emit a message using SocketIO for a dynamic console
socketio.emit('my event', msg, room=client)
socketio.sleep(0.00000000001)
return gen_reached, shortest_path, self.colony.min_distance
