def new_active_ant(ant: Ant, vehicle_num: int, local_search: bool,
IN: np.numarray, q0: float, beta: int,
stop_event: Event):
"""
:param ant:
:param vehicle_num:
:param local_search:
:param IN:
:param q0:
:param beta:
:param stop_event:
:return:
"""
# print('[new_active_ant]: start, start_index %d' % ant.travel_path[0])
unused_depot_count = vehicle_num
while not ant.index_to_visit_empty() and unused_depot_count > 0:
if stop_event.is_set():
# print('[new_active_ant]: receive stop event')
return
next_index_meet_constrains = ant.cal_next_index_meet_constrains()
if len(next_index_meet_constrains) == 0:
ant.move_to_next_index(0)
unused_depot_count -= 1
continue
length = len(next_index_meet_constrains)
ready_time = np.zeros(length)
due_time = np.zeros(length)
for i in range(length):
ready_time[i] = ant.graph.nodes[
next_index_meet_constrains[i]].ready_time
due_time[i] = ant.graph.nodes[
next_index_meet_constrains[i]].due_time
delivery_time = np.maximum(
ant.vehicle_travel_time + ant.graph.node_dist_mat[
ant.current_index][next_index_meet_constrains], ready_time)
delta_time = delivery_time - ant.vehicle_travel_time
distance = delta_time * (due_time - ant.vehicle_travel_time)
distance = np.maximum(1.0,
distance - IN[next_index_meet_constrains])
closeness = 1 / distance
transition_prob = ant.graph.pheromone_mat[ant.current_index][next_index_meet_constrains] * \
np.power(closeness, beta)
transition_prob = transition_prob / np.sum(transition_prob)
if np.random.rand() < q0:
max_prob_index = np.argmax(transition_prob)
next_index = next_index_meet_constrains[max_prob_index]
else:
next_index = MultipleAntColonySystem.stochastic_accept(
next_index_meet_constrains, transition_prob)
ant.graph.local_update_pheromone(ant.current_index, next_index)
ant.move_to_next_index(next_index)
if ant.index_to_visit_empty():
ant.graph.local_update_pheromone(ant.current_index, 0)
ant.move_to_next_index(0)
ant.insertion_procedure(stop_event)
if local_search is True and ant.index_to_visit_empty():
ant.local_search_procedure(stop_event)
def acs_vehicle(new_graph: VrptwGraph, vehicle_num: int, ants_num: int,
q0: float, beta: int, global_path_queue: Queue,
path_found_queue: Queue, stop_event: Event):
"""
:param new_graph:
:param vehicle_num:
:param ants_num:
:param q0:
:param beta:
:param global_path_queue:
:param path_found_queue:
:param stop_event:
:return:
"""
print('[acs_vehicle]: start, vehicle_num %d' % vehicle_num)
global_best_path = None
global_best_distance = None
current_path, current_path_distance, _ = new_graph.nearest_neighbor_heuristic(
max_vehicle_num=vehicle_num)
current_index_to_visit = list(range(new_graph.node_num))
for ind in set(current_path):
current_index_to_visit.remove(ind)
ants_pool = ThreadPoolExecutor(ants_num)
ants_thread = []
ants = []
IN = np.zeros(new_graph.node_num)
while True:
print('[acs_vehicle]: new iteration')
if stop_event.is_set():
print('[acs_vehicle]: receive stop event')
return
for k in range(ants_num):
ant = Ant(new_graph, 0)
thread = ants_pool.submit(
MultipleAntColonySystem.new_active_ant, ant, vehicle_num,
False, IN, q0, beta, stop_event)
ants_thread.append(thread)
ants.append(ant)
for thread in ants_thread:
thread.result()
for ant in ants:
if stop_event.is_set():
print('[acs_vehicle]: receive stop event')
return
IN[ant.index_to_visit] = IN[ant.index_to_visit] + 1
if len(ant.index_to_visit) < len(current_index_to_visit):
current_path = copy.deepcopy(ant.travel_path)
current_index_to_visit = copy.deepcopy(ant.index_to_visit)
current_path_distance = ant.total_travel_distance
IN = np.zeros(new_graph.node_num)
if ant.index_to_visit_empty():
print(
'[acs_vehicle]: found a feasible path, send path info to macs'
)
path_found_queue.put(
PathMessage(ant.travel_path,
ant.total_travel_distance))
new_graph.global_update_pheromone(current_path,
current_path_distance)
if not global_path_queue.empty():
info = global_path_queue.get()
while not global_path_queue.empty():
info = global_path_queue.get()
print('[acs_vehicle]: receive global path info')
global_best_path, global_best_distance, global_used_vehicle_num = info.get_path_info(
)
new_graph.global_update_pheromone(global_best_path,
global_best_distance)
ants_thread.clear()
for ant in ants:
ant.clear()
del ant
ants.clear()
def acs_time(new_graph: VrptwGraph, vehicle_num: int, ants_num: int,
q0: float, beta: int, global_path_queue: Queue,
path_found_queue: Queue, stop_event: Event):
"""
:param new_graph:
:param vehicle_num:
:param ants_num:
:param q0:
:param beta:
:param global_path_queue:
:param path_found_queue:
:param stop_event:
:return:
"""
print('[acs_time]: start, vehicle_num %d' % vehicle_num)
global_best_path = None
global_best_distance = None
ants_pool = ThreadPoolExecutor(ants_num)
ants_thread = []
ants = []
while True:
print('[acs_time]: new iteration')
if stop_event.is_set():
print('[acs_time]: receive stop event')
return
for k in range(ants_num):
ant = Ant(new_graph, 0)
thread = ants_pool.submit(
MultipleAntColonySystem.new_active_ant, ant, vehicle_num,
True, np.zeros(new_graph.node_num), q0, beta, stop_event)
ants_thread.append(thread)
ants.append(ant)
for thread in ants_thread:
thread.result()
ant_best_travel_distance = None
ant_best_path = None
for ant in ants:
if stop_event.is_set():
print('[acs_time]: receive stop event')
return
if not global_path_queue.empty():
info = global_path_queue.get()
while not global_path_queue.empty():
info = global_path_queue.get()
print('[acs_time]: receive global path info')
global_best_path, global_best_distance, global_used_vehicle_num = info.get_path_info(
)
if ant.index_to_visit_empty() and (
ant_best_travel_distance is None or
ant.total_travel_distance < ant_best_travel_distance):
ant_best_travel_distance = ant.total_travel_distance
ant_best_path = ant.travel_path
new_graph.global_update_pheromone(global_best_path,
global_best_distance)
if ant_best_travel_distance is not None and ant_best_travel_distance < global_best_distance:
print(
'[acs_time]: ants\' local search found a improved feasible path, send path info to macs'
)
path_found_queue.put(
PathMessage(ant_best_path, ant_best_travel_distance))
ants_thread.clear()
for ant in ants:
ant.clear()
del ant
ants.clear()
def new_active_ant(ant: Ant, vehicle_num: int, local_search: bool, IN: np.numarray, q0: float, beta: int, stop_event: Event):
"""
按照指定的vehicle_num在地图上进行探索，所使用的vehicle num不能多于指定的数量，acs_time和acs_vehicle都会使用到这个方法
对于acs_time来说，需要访问完所有的结点（路径是可行的），尽量找到travel distance更短的路径
对于acs_vehicle来说，所使用的vehicle num会比当前所找到的best path所使用的车辆数少一辆，要使用更少的车辆，尽量去访问结点，如果访问完了所有的结点（路径是可行的），就将通知macs
:param ant:
:param vehicle_num:
:param local_search:
:param IN:
:param q0:
:param beta:
:param stop_event:
:return:
"""
# print('[new_active_ant]: start, start_index %d' % ant.travel_path[0])
# 在new_active_ant中，最多可以使用vehicle_num个车，即最多可以包含vehicle_num+1个depot结点，由于出发结点用掉了一个，所以只剩下vehicle个depot
unused_depot_count = vehicle_num
# 如果还有未访问的结点，并且还可以回到depot中
while not ant.index_to_visit_empty() and unused_depot_count > 0:
if stop_event.is_set():
# print('[new_active_ant]: receive stop event')
return
# 计算所有满足载重等限制的下一个结点
next_index_meet_constrains = ant.cal_next_index_meet_constrains()
# 如果没有满足限制的下一个结点，则回到depot中
if len(next_index_meet_constrains) == 0:
ant.move_to_next_index(0)
unused_depot_count -= 1
continue
# 开始计算满足限制的下一个结点，选择各个结点的概率
length = len(next_index_meet_constrains)
ready_time = np.zeros(length)
due_time = np.zeros(length)
for i in range(length):
ready_time[i] = ant.graph.nodes[next_index_meet_constrains[i]].ready_time
due_time[i] = ant.graph.nodes[next_index_meet_constrains[i]].due_time
delivery_time = np.maximum(ant.vehicle_travel_time + ant.graph.node_dist_mat[ant.current_index][next_index_meet_constrains], ready_time)
delta_time = delivery_time - ant.vehicle_travel_time
distance = delta_time * (due_time - ant.vehicle_travel_time)
distance = np.maximum(1.0, distance-IN[next_index_meet_constrains])
closeness = 1/distance
transition_prob = ant.graph.pheromone_mat[ant.current_index][next_index_meet_constrains] * \
np.power(closeness, beta)
transition_prob = transition_prob / np.sum(transition_prob)
# 按照概率直接选择closeness最大的结点
if np.random.rand() < q0:
max_prob_index = np.argmax(transition_prob)
next_index = next_index_meet_constrains[max_prob_index]
else:
# 使用轮盘赌算法
next_index = MultipleAntColonySystem.stochastic_accept(next_index_meet_constrains, transition_prob)
# 更新信息素矩阵
ant.graph.local_update_pheromone(ant.current_index, next_index)
ant.move_to_next_index(next_index)
# 如果走完所有的点了，需要回到depot
if ant.index_to_visit_empty():
ant.graph.local_update_pheromone(ant.current_index, 0)
ant.move_to_next_index(0)
# 对未访问的点进行插入，保证path是可行的
ant.insertion_procedure(stop_event)
# ant.index_to_visit_empty()==True就是feasible的意思
if local_search is True and ant.index_to_visit_empty():
ant.local_search_procedure(stop_event)
def acs_vehicle(new_graph: VrptwGraph, vehicle_num: int, ants_num: int, q0: float, beta: int,
global_path_queue: Queue, path_found_queue: Queue, stop_event: Event):
"""
对于acs_vehicle来说，所使用的vehicle num会比当前所找到的best path所使用的车辆数少一辆，要使用更少的车辆，尽量去访问结点，如果访问完了所有的结点（路径是可行的），就将通知macs
:param new_graph:
:param vehicle_num:
:param ants_num:
:param q0:
:param beta:
:param global_path_queue:
:param path_found_queue:
:param stop_event:
:return:
"""
# vehicle_num设置为比当前的best_path少一个
print('[acs_vehicle]: start, vehicle_num %d' % vehicle_num)
global_best_path = None
global_best_distance = None
# 使用nearest_neighbor_heuristic算法初始化path 和distance
current_path, current_path_distance, _ = new_graph.nearest_neighbor_heuristic(max_vehicle_num=vehicle_num)
# 找出当前path中未访问的结点
current_index_to_visit = list(range(new_graph.node_num))
for ind in set(current_path):
current_index_to_visit.remove(ind)
ants_pool = ThreadPoolExecutor(ants_num)
ants_thread = []
ants = []
IN = np.zeros(new_graph.node_num)
while True:
print('[acs_vehicle]: new iteration')
if stop_event.is_set():
print('[acs_vehicle]: receive stop event')
return
for k in range(ants_num):
ant = Ant(new_graph, 0)
thread = ants_pool.submit(MultipleAntColonySystem.new_active_ant, ant, vehicle_num, False, IN, q0,
beta, stop_event)
ants_thread.append(thread)
ants.append(ant)
# 这里可以使用result方法，等待线程跑完
for thread in ants_thread:
thread.result()
for ant in ants:
if stop_event.is_set():
print('[acs_vehicle]: receive stop event')
return
IN[ant.index_to_visit] = IN[ant.index_to_visit]+1
# 蚂蚁找出来的路径与current_path进行比较，是否能使用vehicle_num辆车访问到更多的结点
if len(ant.index_to_visit) < len(current_index_to_visit):
current_path = copy.deepcopy(ant.travel_path)
current_index_to_visit = copy.deepcopy(ant.index_to_visit)
current_path_distance = ant.total_travel_distance
# 并且将IN设置为0
IN = np.zeros(new_graph.node_num)
# 如果这一条路径是feasible的话，就要发到macs_vrptw中
if ant.index_to_visit_empty():
print('[acs_vehicle]: found a feasible path, send path info to macs')
path_found_queue.put(PathMessage(ant.travel_path, ant.total_travel_distance))
# 更新new_graph中的信息素，global
new_graph.global_update_pheromone(current_path, current_path_distance)
if not global_path_queue.empty():
info = global_path_queue.get()
while not global_path_queue.empty():
info = global_path_queue.get()
print('[acs_vehicle]: receive global path info')
global_best_path, global_best_distance, global_used_vehicle_num = info.get_path_info()
new_graph.global_update_pheromone(global_best_path, global_best_distance)
ants_thread.clear()
for ant in ants:
ant.clear()
del ant
ants.clear()
def acs_time(new_graph: VrptwGraph, vehicle_num: int, ants_num: int, q0: float, beta: int,
global_path_queue: Queue, path_found_queue: Queue, stop_event: Event):
"""
对于acs_time来说，需要访问完所有的结点（路径是可行的），尽量找到travel distance更短的路径
:param new_graph:
:param vehicle_num:
:param ants_num:
:param q0:
:param beta:
:param global_path_queue:
:param path_found_queue:
:param stop_event:
:return:
"""
# 最多可以使用vehicle_num辆车，即在path中最多包含vehicle_num+1个depot中，找到路程最短的路径，
# vehicle_num设置为与当前的best_path一致
print('[acs_time]: start, vehicle_num %d' % vehicle_num)
# 初始化信息素矩阵
global_best_path = None
global_best_distance = None
ants_pool = ThreadPoolExecutor(ants_num)
ants_thread = []
ants = []
while True:
print('[acs_time]: new iteration')
if stop_event.is_set():
print('[acs_time]: receive stop event')
return
for k in range(ants_num):
ant = Ant(new_graph, 0)
thread = ants_pool.submit(MultipleAntColonySystem.new_active_ant, ant, vehicle_num, True,
np.zeros(new_graph.node_num), q0, beta, stop_event)
ants_thread.append(thread)
ants.append(ant)
# 这里可以使用result方法，等待线程跑完
for thread in ants_thread:
thread.result()
ant_best_travel_distance = None
ant_best_path = None
# 判断蚂蚁找出来的路径是否是feasible的，并且比全局的路径要好
for ant in ants:
if stop_event.is_set():
print('[acs_time]: receive stop event')
return
# 获取当前的best path
if not global_path_queue.empty():
info = global_path_queue.get()
while not global_path_queue.empty():
info = global_path_queue.get()
print('[acs_time]: receive global path info')
global_best_path, global_best_distance, global_used_vehicle_num = info.get_path_info()
# 路径蚂蚁计算得到的最短路径
if ant.index_to_visit_empty() and (ant_best_travel_distance is None or ant.total_travel_distance < ant_best_travel_distance):
ant_best_travel_distance = ant.total_travel_distance
ant_best_path = ant.travel_path
# 在这里执行信息素的全局更新
new_graph.global_update_pheromone(global_best_path, global_best_distance)
# 向macs发送计算得到的当前的最佳路径
if ant_best_travel_distance is not None and ant_best_travel_distance < global_best_distance:
print('[acs_time]: ants\' local search found a improved feasible path, send path info to macs')
path_found_queue.put(PathMessage(ant_best_path, ant_best_travel_distance))
ants_thread.clear()
for ant in ants:
ant.clear()
del ant
ants.clear()