决策变量的累积和作为预算约束（纸浆）

Question

我的程序没有按要求接受我的约束。

这是我的全部代码：

    import sys
!{sys.executable} -m pip install pulp
import pulp
import pandas as pd

dates = [d.strftime('%Y-%m-%d') for d in pd.date_range("2020-01-01","2020-12-31")]


days = range(367)
preise = range(367)
capacity = 100000

#Decision variable
volumes = pulp.LpVariable.dicts("volumes", days,
                                 lowBound=-2400, 
                                 upBound=1500, 
                                 cat=pulp.LpContinuous)

#Initialize Problem
storage = pulp.LpProblem("Storage Valuation", pulp.LpMaximize)

#Objective Function
storage += pulp.lpSum([volumes[(i)] * preise[(i)] 
                       for i in days]), "objective function"

for j in days:
    storage += sum(volumes) <= capacity, "capacity constraint {}".format(j)


storage.solve()
print("Status : ", pulp.LpStatus[storage.status])
print("Result : ", pulp.value(storage.objective))
data = []
for v in storage.variables():
    data.append({'wert': v.varValue})
result = pd.DataFrame (data)


result.insert(1, "Storage level", result['wert'].cumsum(), True)
result.insert(2, "Date", dates, True)

with pd.option_context('display.max_rows', None, 'display.max_columns', None):  
    print(result)

我的决策变量的累计和每天都不允许超过一定的容量。

对于约束，我也试过类似的东西：

    for j in days:
    storage += pulp.lpSum([volumes[i][j] for i in days) <= capacity, "Capacity Limit {}".format(j)

我认为 lpSum 语句有问题。也许我的程序需要类似计数变量的东西。如果是这样，我该如何实施？

谢谢！

Answer 1

以下更改给出了结果（虽然我不太确定这是你需要的）：

#days = range(367)
#preise = range(367)
days = range(366)
preise = range(366)

更改约束条件（限制累计和）：

#for j in days:
#    storage += sum(volumes) <= capacity, "capacity constraint {}".format(j)    
for j in days:
    storage += pulp.lpSum([volumes[(i)] for i in range(j+1)]) <= capacity

不要遍历字典，变量的顺序无法保证。相反：

#for v in storage.variables():
#    data.append({'wert': v.varValue})
for j in days:
    data.append({'wert': volumes[(j)].varValue})

决策变量的累积和作为预算约束（纸浆）

Cumulative Sum of decision variable as budget constraint (pulp)

python

optimization

constraints

linear-programming

pulp