info.append({'generation': i,
'agent': 0,
'fitness': explorer.fitness,
'solution': explorer.solution})
# Update pheromones
for a, b in graph.edges():
weight = graph[a][b]['weight']
if (a, b) in explorer or (b, a) in explorer:
increase = explorer.fitness
else:
increase = 0
graph[a][b]['weight'] = ant_system_update(learning_rate, weight, increase)
# Construct solutions
for j, agent in enumerate(population, start=1):
agent.search(graph, max_steps)
agent.evaluate()
info.append({'generation': i,
'agent': j,
'fitness': agent.fitness,
'solution': agent.solution})
# Log generation info
collection_agents.insert(info)
print('# generation {0}'.format(i))
for i in range(generations):
population = [ProportionalAnt(start, end) for x in range(agents)]
info = []
# Construct solutions
for j, agent in enumerate(population):
agent.search(graph, max_steps)
agent.evaluate()
info.append({'generation': i,
'agent': j,
'fitness': agent.fitness,
'solution': agent.solution})
# Log generation info
collection_agents.insert(info)
print('# generation {0}'.format(i))
# Update pheromones
for a, b in graph.edges():
weight = graph[a][b]['weight']
increase = sum([agent.fitness for agent in population
if (a, b) in agent or (b, a) in agent])
graph[a][b]['weight'] = ant_system_update(rho, weight, increase)
# Log pheromone info
collection_pheromones.insert({'generation': i,
'pheromones': graph.edges(data=True)})