def GetOUTPUTDATA(ind,
M_max=4,
c_op=c_op,
t_convey=t_convey,
t_pack=t_pack,
c_od=c_od,
lm=lm):
"""
ind:例如[0,3,19,8,7,0,10,6,4,14,12,9,16,0,13,11,18,20,2,0,17,1,15,5,0]
M_max:分区数
"""
batch, batch_item = getBatch(ind)
batch_Quantity = [len(i) for i in batch_item]
data = getBatchPicktime(batch_item)
batch_packingtime = [t_pack * i for i in batch_Quantity]
ODtime = ODtimeTable(np.array(batch)) #每个客户点交付时间
ODtime = [i[-1] * 10 for i in ODtime] #每个批次最后一个客户点交付时间*10
op_duetime = [(lm - i) for i in ODtime]
"""求解BSP"""
N_max = data.shape[1]
index = list(range(1, N_max + 1))
#SPT
data1 = data[1:, :].sum(axis=0)
tup = list(zip(index, data1))
tup.sort(key=lambda x: x[1])
Seq_SPT = np.array([i[0] for i in tup])
data_best_all_SPT = data[:, Seq_SPT - 1]
TPicktime_SPT = makespan(data_best_all_SPT)[-1, :].sum()
#LDT
tup = list(zip(index, ODtime))
tup.sort(key=lambda x: x[1], reverse=0)
Seq_LDT = np.array([i[0] for i in tup])
data_best_all_LDT = data[:, Seq_LDT - 1]
TPicktime_LDT = makespan(data_best_all_LDT)[-1, :].sum()
#择优
if TPicktime_SPT <= TPicktime_LDT:
Seq = Seq_SPT
data_best_all = data_best_all_SPT
else:
Seq = Seq_LDT
data_best_all = data_best_all_LDT
order = np.array(data_best_all[0, :]) - 1
return Seq, data_best_all, batch_packingtime, order, op_duetime, ODtime
def fitness(self, pop, c_op, op_duetime, c_od, batch_packingtime,
t_convey):
fitness = np.zeros([self.pop_size, 1])
for i in range(self.pop_size):
# fitness[i] = 1 / makespan_value(self.data[:, pop[i]]) #最大完成时间
# fitness[i] = 1 / sum(makespan(self.data[:, pop[i]])[-1, :]) #合计完成时间
# fitness[i] = 1 / (sum(makespan(self.data[:, pop[i]])[-1, :])*c_op) # 合计完成时间*单位成本
######目标：合计订单拣选成本+违约成本##########
#各批次订单抵达后台时间
batch_arrive_time = np.array([
i + t_convey
for i in (makespan(self.data[:, pop[i]], t_convey)[-1, :])
])
order = pop[i]
# 各批次订单开始配送时间
batch_delivery_start_time = batch_arrive_time + np.array(
[batch_packingtime[i] for i in order])
#违约时间
batch_over_duetime = []
for j in list(
np.array(batch_delivery_start_time) -
np.array([op_duetime[i] for i in order])):
if j < 0:
batch_over_duetime.append(0)
else:
batch_over_duetime.append(j)
# 总订单拣选成本+违约成本=合计订单开始配送时间*单位时间成本+合计违约时间*单位违约时间成本
TC = ((sum(batch_delivery_start_time) / 10 * c_op) +
sum(batch_over_duetime) / 10 * c_od)
fitness[i] = 1 / (TC)
######合计加权时间##########
# W=np.array([1,1,1,1]) #各工件权重
# Value_all=np.array(makespan(self.data[:, pop[i]])[-1, :])
# Seq_temp=makespan(self.data[:, pop[i]])[0,:]
# W_new=W[Seq_temp-1] #工件权重调整
# fitness[i] = 1 / sum(np.multiply(Value_all, W_new)) #加权合计完成时间
################################
return fitness
def calOPCost(ind,
M_max=4,
c_op=c_op,
t_convey=t_convey,
t_pack=t_pack,
c_od=c_od,
lm=lm):
"""
ind:例如[0,3,19,8,7,0,10,6,4,14,12,9,16,0,13,11,18,20,2,0,17,1,15,5,0]
M_max:分区数
"""
batch, batch_item = getBatch(ind)
batch_Quantity = [len(i) for i in batch_item]
data = getBatchPicktime(batch_item)
batch_packingtime = [t_pack * i for i in batch_Quantity] #b属于B
ODtime = ODtimeTable(np.array(batch))
ODtime = [i[-1] * 10 for i in ODtime] #每个批次最后一个客户点交付时间*10
op_duetime = [(lm - i) for i in ODtime]
"""求解BSP"""
N_max = data.shape[1]
index = list(range(1, N_max + 1))
#SPT
data1 = data[1:, :].sum(axis=0)
tup = list(zip(index, data1))
tup.sort(key=lambda x: x[1])
Seq_SPT = np.array([i[0] for i in tup])
data_best_all_SPT = data[:, Seq_SPT - 1]
TPicktime_SPT = makespan(data_best_all_SPT)[-1, :].sum()
#LDT
tup = list(zip(index, op_duetime))
tup.sort(key=lambda x: x[1])
Seq_LDT = np.array([i[0] for i in tup])
data_best_all_LDT = data[:, Seq_LDT - 1]
TPicktime_LDT = makespan(data_best_all_LDT)[-1, :].sum()
#择优
if TPicktime_SPT <= TPicktime_LDT:
Seq = Seq_SPT
data_best_all = data_best_all_SPT
else:
Seq = Seq_LDT
data_best_all = data_best_all_LDT
order = np.array(data_best_all[0, :]) - 1
"""时间类"""
#各分区订单抵达后台时间
batch_arrive_time = np.array([
i + t_convey for i in (makespan(data_best_all, t_convey)[-1, :])
]) #b 属于pi，拣货顺序批次
# 批次订单开始配送时间=抵达后台时间+打包分类时间
batch_delivery_start_time = batch_arrive_time + np.array(
[batch_packingtime[i] for i in order])
#批次订单违约时间
batch_over_duetime = []
for j in list(
np.array(batch_delivery_start_time) -
np.array([op_duetime[i] for i in order])):
if j < 0:
batch_over_duetime.append(0)
else:
batch_over_duetime.append(j)
"""成本类：将时间/10"""
#各批次订单拣选成本
batch_op_cost = np.array(batch_delivery_start_time) / 10 * c_op
#各批次订单违约成本
batch_overduetime_cost = np.array(batch_over_duetime) / 10 * c_od
#订单总拣选成本+违约成本
TC = batch_op_cost.sum() + batch_overduetime_cost.sum()
return TC
]
# 如果最迟到达时间大于实际到达客户的时间，则延迟为0，否则延迟设为实际到达时间与最迟到达时间之差
timeDelay = [
max(timeArrangement[i] - desiredTime[i], 0)
for i in range(len(bestInd))
]
C_od_d = sum(timeDelay) * c_od_d
"""计算订单拣选延迟时间及成本"""
Seq, data_best_all, batch_packingtime, order, op_duetime, ODtime = GetOUTPUTDATA(
bestInd)
#拣货开始时间
op_start_time = makespan_left(data_best_all)
#各分区订单抵达后台时间
batch_arrive_time = np.array([
i + t_convey for i in (makespan(data_best_all, t_convey)[-1, :])
]) #b 属于pi，拣货顺序批次
# 批次订单开始配送时间=抵达后台时间+打包分类时间
batch_delivery_start_time = batch_arrive_time + np.array(
[batch_packingtime[i] for i in order])
#批次订单违约时间
batch_over_duetime = []
for j in list(
np.array(batch_delivery_start_time) -
np.array([op_duetime[i] for i in order])):
if j < 0:
batch_over_duetime.append(0)
else:
batch_over_duetime.append(j)
C_od_op = sum(batch_over_duetime) / 10 * c_od
def evaluation(self, df):
size = len(df.index)
for i in range(size):
df.loc[i, 'makespans'] = tool.makespan(np.array(self.span), df['jobs'][i])
return df
print("Best Seq:", Seq)
data_best_all = data[:, Seq - 1]
order = array(data_best_all[0, :]) - 1
"""绘制甘特图"""
Draw = 0
if Draw == 1:
plt.figure()
gatt(data_best_all, t_convey)
plt.show()
"""计算各指标"""
"""时间类"""
#拣货开始时间
op_start_time = makespan_left(data_best_all)
#拣货完成时间
makespan_time = makespan(data_best_all)
#各分区订单抵达后台时间
batch_arrive_time = array(
[i + t_convey for i in (makespan(data_best_all)[-1, :])])
#批次订单开始配送时间=抵达后台时间+打包分类时间
batch_delivery_start_time = batch_arrive_time + array(
[batch_packingtime[i] for i in order])
#批次订单违约时间
batch_over_duetime = []
for j in list(
array(batch_delivery_start_time) -
array([op_duetime[i] for i in order])):
if j < 0:
batch_over_duetime.append(0)
else:
batch_Quantity=[len(i) for i in batch_item]
data=getBatchPicktime(batch_item)
batch_packingtime=[t_pack*i for i in batch_Quantity]
op_duetime=[lm-i for i in ODtime]
"""bsp"""
N_max=data.shape[1]
index=list(range(1,N_max+1))
data1=data[1:,:].sum(axis=0)
tup=list(zip(index,data1))
tup.sort(key=lambda x:x[1])
Seq_SPT=array([i[0] for i in tup])
data_best_all=data[:,Seq_SPT-1]
TPicktime_SPT=makespan(data_best_all)[-1,:].sum()
tup=list(zip(index,C_d_Bat))
tup.sort(key=lambda x:x[1],reverse=True)
Seq_LDT=array([i[0] for i in tup])
data_best_all=data[:,Seq_LDT-1]
TPicktime_LDT=makespan(data_best_all)[-1,:].sum()
if TPicktime_SPT<=TPicktime_LDT:
Seq=Seq_SPT
else:
Seq=Seq_LDT
