import random
import pandas as pd
import time
# 这里定义一个全局变量，用来表示最大的货架数——即批次种类限制
N = 200
# 然后是写一个函数，实现只存储同一订单对应的货品种类
def big_and_small(choose0, order):
"""
该函数用于在未保留的订单中划分大订单和小订单，大订单占1/3
:param choose0:
:param order:
:return:
"""
big1 = []
small0 = []
choose0 = dict(choose0)
# 选取的订单字典，键是订单号，值是种类数
choose_order = {}
for i1 in choose0:
if choose0[i1] == 0:
choose_order[i1] = len(order[i1])
# 将未被选的订单字典排序
choose_order = sorted(choose_order.items(), key=lambda x: x[1])
new_len = len(choose_order)
# 计算大小分界的索引
mid_index = new_len * 2 // 3
index0 = 0
for i1 in choose_order:
if index0 < mid_index:
small0.append(i1[0])
else:
big1.append(i1[0])
index0 += 1
return big1, small0
# 写一个函数判断是否在当前批次
def is_in(batch, order):
"""
其实应该是得出新增订单的新增货品种类数
:param batch: 当前批订单, 内容是包含的货品
:param order: 当前订单，也是货品
:return: 新增货品种类数
"""
# 将批次和订单都列表化
batch = list(batch)
order = list(order)
ans = 0
# 循环累加不在当前批次的货品种类数
for o0 in order:
if o0 not in batch:
ans += 1
return ans
# 写一个选取函数
def choose_batch(orders1, batch01, batch02, keys0):
"""
该函数用于在大订单集或者小订单集将这些订单实现分批
引入洗牌算法，将选取的订单的顺序进行打乱随机
:param orders1: 订单
:param batch01: 批次列表1，存储批次对应的订单
:param batch02: 批次列表2，存储批次对应的货品种类
:param keys0: 选取订单的列表, 内容是订单号
:return: 返回选取加入后的批次1和2
"""
# 写一个洗牌算法
# 先获取选取订单的长度
# keys_len = len(keys0)
# # 逆序遍历选取列表
# # 通过交换随机下标来实现将选取的订单的顺序洗牌
# for k in range(keys_len - 1, -1, -1):
# # 获取随机下标
# rindex 跟单网gendan5.com= random.randint(0, k)
# # 交换位置
# temp = keys0[rindex]
# keys0[rindex] = keys0[k]
# keys0[k] = temp
batch1 = dict(batch01)
batch2 = dict(batch02)
# 获取到当前最大批次数对应的下标
i0 = len(batch1) - 1
# 如果此时没有，这需要置为0
if i0 < 0:
i0 = 0
for key0 in keys0:
# 初始时需要创建批次
if i0 not in batch1:
batch1[i0] = []
batch2[i0] = []
add_type = -1
index0 = -1
flag = False
for j in range(i0 + 1):
if key0 in batch1[j]:
flag = True
break
at = is_in(batch2[j], orders[key0])
# 完全重合直接退出循环
if at == 0:
index0 = j
break
if (add_type < 0 or add_type >= at) and len(batch2[j]) + at <= N:
add_type = at
index0 = j
if flag:
continue
if index0 >= 0:
batch1[index0].append(key0)
for good in orders1[key0]:
if good not in batch2[index0]:
batch2[index0].append(good)
else:
i0 += 1
batch1[i0] = []
batch1[i0].append(key0)
batch2[i0] = orders1[key0]
return batch1, batch2
# 然后接下来写一个从小订单中选取
def choose_small(orders1, batch1, batch2, small1):
orders1 = dict(orders1)
keys = list(small1)
batch1 = dict(batch1)
batch2 = dict(batch2)
batch1, batch2 = choose_batch(orders1, batch1, batch2, keys)
return batch1, batch2
# 这个是在大的数组中重新选取
def choose_big(big1, orders1, batch, batch2):
orders1 = dict(orders1)
keys = list(big1)
# 批次数组有两种，一种是存储订单，一种是存储货品
batch1 = dict(batch)
batch2 = dict(batch2)
batch1, batch2 = choose_batch(orders1, batch1, batch2, keys)
return batch1, batch2
# 然后写一个迭代函数
def gen_ite(order, btch_1, btch_2):
"""
参考遗传算法的精英基因保留的思想，我们在迭代的过程中将批次中订单数较多的保留，这里每次选取10%的批次
然后需要将这些批次标记为已经选择然后重新在为选取的订单集划分大小订单
选取的批次可以直接新增在下一代的开头
:param order: 订单字典，订单号对应相应的货品种类
:param btch_1:
:param btch_2:
:return: 最终多次迭代后的结果
"""
btch_1 = dict(btch_1)
btch_2 = dict(btch_2)
for i0 in range(1):
# 先对批次的订单数排序 —— 降序排序
btch = sorted(btch_1.items(), key=lambda x: len(x[1]), reverse=True)
btch_len = len(btch)
index0 = btch_len // 5
# 生成两个新的字典，和btch1，2类似
bt1 = {}
bt2 = {}
# 先将精英保留
# 选择字典
choose_0 = {}
for j in order.keys():
choose_0[j] = 0
for k in range(index0):
for o0 in btch_1[k][1]:
choose_0[o0] = 1
bt1[k] = btch_1[btch[k][0]]
bt2[k] = btch_2[btch[k][0]]
keys = []
for m in order.keys():
if choose[m] == 0:
keys.append(m)
# # 接下来划分大小订单
# big1, small1 = big_and_small(choose, orders)
# print('big: %d small: %d' % (len(big1), len(small1)))
# # 之后根据划分的大小订单在得到新的分批
# bt1, bt2 = choose_big(big1, orders, bt1, bt2)
# print(len(bt1), len(bt2))
# # bt1, bt2 = choose_small(orders, bt1, bt2, small1)
# # print(len(bt1), len(bt2))
bt1, bt2 = choose_batch(order, bt1, bt2, keys)
# 然后比较新旧分批的批数来更新批次
if len(bt1) < len(btch_1):
btch_1 = bt1
btch_2 = bt2
print('第%d次, 批次数为%d' % (i0 + 1, len(btch_1)))
le = 0
for bi in btch_1:
le += len(btch_1[bi])
print(le)
return btch_1, btch_2
# 写一个计算距离的函数
def compute_dis(order, batch10, goods):
"""
:param order: 其实就是订单集合，通过订单可以知道需要的货品
:param batch10: 批次下的订单列表
:param goods: 当前批次下的货物关于货物列表的索引
:return: 返回距离和
"""
order = dict(order)
batch10 = list(batch10)
goods = dict(goods)
dis = 0
for bat in batch10:
index_max = -1
index_min = -1
for good in order[bat]:
if index_min == -1 and index_max == -1:
index_max = index_min = goods[good]
else:
index_max = max(index_max, goods[good])
index_min = min(index_min, goods[good])
dis += index_max - index_min
return dis
# 写一个第二问的解决函数
def solution2(order, batch2, batch1):
batch1 = dict(batch1)
batch2 = dict(batch2)
order = dict(order)
pre_total = 0
ans = []
p = 0
total = 0
start0 = time.time()
for batch in batch2.items():
# 现在是对每一个批次
# 先建立货物种类到下标映射, 和下标到货物的映射
good_to_index = {}
index_to_good = batch[1]
for index1 in range(len(index_to_good)):
good = index_to_good[index1]
good_to_index[good] = index1
dis = compute_dis(order, batch1[batch[0]], good_to_index)
print('当前第%d批原始距离为 %d' % (p + 1, dis))
pre_total += dis
for i0 in range(200000):
j = random.randint(0, len(index_to_good) - 1)
k = random.randint(0, len(index_to_good) - 1)
if j == k:
continue
# 获取货物信息
good1, good2 = index_to_good[j], index_to_good[k]
# 先交换
temp = good_to_index[good1]
good_to_index[good1] = good_to_index[good2]
good_to_index[good2] = temp
if dis > compute_dis(order, batch1[batch[0]], good_to_index):
index_to_good[j], index_to_good[k] = good2, good1
dis = compute_dis(order, batch1[batch[0]], good_to_index)
else:
good_to_index[good1], good_to_index[good2] = good_to_index[good2], good_to_index[good1]
print('迭代后 第%d批原始距离为 %d' % (p + 1, dis))
total += dis
ans.append(index_to_good)
p += 1
print('原始距离', pre_total)
print('最终距离', total)
end0 = time.time()
print('时间消耗', end0 - start0)
return ans
# 找到订单的起始点
def find_min_and_max_index(order_good, good_index):
order_good = list(order_good)
good_index = dict(good_index)
min_index = max_index = -1
for good_name in order_good:
if min_index == -1 and max_index == -1:
min_index = max_index = good_index[good_name]
continue
min_index = min(min_index, good_index[good_name])
max_index = max(max_index, good_index[good_name])
return min_index, min_index
def right_find(order, batch1, good_index, pos):
order = dict(order)
batch1 = list(batch1)
good_index = dict(good_index)
ans = ''
j = int(pos)
tar = j
for name in batch1:
i1, i2 = find_min_and_max_index(order[name], good_index)
if i1 >= pos and (j > i1 or j == pos):
j = i1
ans = name
tar = i2
return ans, tar
def left_find(order, batch1, good_index, pos):
order = dict(order)
batch1 = list(batch1)
good_index = dict(good_index)
ans = ''
j = pos
tar = j
for name in batch1:
i1, i2 = find_min_and_max_index(order[name], good_index)
if i1 <= pos and (j < i1 or j == pos):
j = i2
ans = name
tar = i1
return ans, tar
def find_people(distance):
"""
找工人函数
我们根据方向和工人所处的位置决定，选取哪个工人
:param distance: 距离数组
:return: 返回工人的下标
"""
# 答案初始化为下标0
ans = 0
distance = list(distance)
for i1 in range(1, len(distance)):
if distance[i1] < distance[ans]:
ans = i1
return ans
def solution3(order, batch1, batch2, n):
"""
该函数主要用于求解第三问
:param order: 订单对应的货品，字典对象
:param batch1: 分批好的订单，批号对应的订单
:param batch2: 分批好的订单对应的货品种类，批号对应的种类
:param n: 工人的数量
:return: 返回第三问需要的结果
"""
# 先将这些字典对象示例化，实现可变参数在函数内和外的分离
batch1 = dict(batch1)
order = dict(order)
batch2 = dict(batch2)
# 答案，用列表存储，每一个其实也是列表，分别对应答案需要的答案
ans = []
peo_dis1 = [] # 用来计算每一批员工的总路程
total_dis0 = [] # 总距离
for i1 in range(n):
total_dis0.append(0)
# 我们遍历批次
for i1 in batch1.keys():
# 实例化批次对应的订单和货品种类，这里其实还是将其变为数组的形式
bat_or = list(batch1[i1])
bat_go = list(batch2[i1])
# 货品在该批次中对应的位置， 即货架的位置
good_index = {}
# 遍历一遍订单和货品种类，这里的顺序其实就是第二问求解的摆放顺序
for ind in range(len(bat_go)):
# 先获取当前的货物
good = bat_go[ind]
# 利用字典的映射关系，设置货物对应的摆放位置的关系
good_index[good] = ind
people = [] # 工人列表，用来存储工人的位置
task = [] # 任务列表，用来存储工人当前批干的第几个订单
peo_dis0 = [] # 当前批的距离
# 先将这两个列表都初始化为0
for p in range(n):
people.append(0)
task.append(0)
peo_dis0.append(0)
# 当批次中的订单还有的时候，需要进行分配
while len(bat_or) > 0:
# 我们可以先进行正向的寻找, 先找当前最小的下标，同级优先按工人编号寻找
p = find_people(peo_dis0)
# 获取当前工人的位置
pos = people[p]
# 然后在以当前工人位置为起点，向右寻找最近的订单，并返回订单号和订单完成时工人的位置
name, next_pos = right_find(order, bat_or, good_index, pos)
# 如果找不到订单，说明右边没有订单了，需要逆向寻找
if name == '':
# 向当前位置最左边找离最近的订单
name, next_pos = left_find(order, bat_or, good_index, pos)
task[p] += 1 # 当前工人的任务数 + 1
temp = [name, i1 + 1, p + 1, task[p]] # 当前行的答案
ans.append(temp) # 将这一行的答案添加到最终答案中
bat_or.remove(name) # 将这个订单从列表中移除
print(next_pos)
# 计算移动距离
dis = abs(pos - next_pos)
# 加和
peo_dis0[p] += dis
people[p] = next_pos
peo_dis1.append(peo_dis0)
for ik in range(n):
total_dis0[ik] += peo_dis0[ik]
# 返回答案
return ans, peo_dis1, total_dis0
if __name__ == "__main__":
start = time.time()
# 先写一个文件路径
filepath = '附件1：订单信息.csv'
# 先读取文件
# 使用 pandas 内置的读取方式
df = pd.read_csv(filepath) # 将文件读取为一个dataframe格式
con = pd.DataFrame(df, columns=['OrderNo', 'ItemNo'])
orders = {}
for line in con.index:
if not con.iloc[line]['OrderNo'] in orders:
orders[con.iloc[line]['OrderNo']] = []
orders[con.iloc[line]['OrderNo']].append(con.iloc[line]['ItemNo'])
s = set(orders.keys())
print(len(s))
# 选取字典，表示对应订单是否被选取
choose = {}
# 初始化为未被选取
for key in orders.keys():
choose[key] = 0
big, small = big_and_small(choose, orders)
print(len(big), len(small))
b1, b2 = choose_big(big, orders, {}, {})
b1, b2 = choose_small(orders, b1, b2, small)
end = time.time()
print(end - start)
print(len(b1), len(b2))
btch1, btch2 = dict(b1), dict(b2)
file = open('result1.csv', 'w+')
result1 = 'OrderNo,GroupNo\n'
for b in btch1.keys():
for o in btch1[b]:
result1 += o + ',' + str(b + 1) + '\n'
print(result1, file=file)
file.close()
ans2 = solution2(orders, batch1=btch1, batch2=btch2)
result2 = 'ItemNo,GroupNo,ShelfNo\n'
for index in btch2.keys():
btch2[index] = ans2[index]
for _i in range(len(ans2[index])):
result2 += ans2[index][_i] + ',' + str(index + 1) + ',' + str(_i + 1) + '\n'
file = open('result2.csv', 'w+')
print(result2, file=file)
file.close()
ans3, peo_dis, total_dis = solution3(orders, batch2=btch2, batch1=btch1, n=5)
result3 = 'OrderNo,GroupNo,WorkerNo,TaskNo\n'
for a in ans3:
for index in range(len(a)):
result3 += str(a[index])
if index + 1 < len(a):
result3 += ','
else:
result3 += '\n'
file = open('result3.csv', 'w+')
print(result3, file=file)
file.close()
result4 = ''
for b in peo_dis:
for index in range(len(b)):
result4 += str(b[index])
if index + 1 < len(b):
result4 += ','
else:
result4 += '\n'
file = open('员工路程，单批.csv', 'w+')
print(result4, file=file)
file.close()
file = open('员工路径总和.csv', 'w+')
result4 = ''
for i in range(len(total_dis)):
result4 += str(total_dis[i])
if i + 1 < len(total_dis):
result4 += ','
else:
result4 += '\n'
print(result4, file=file)
file.close()