Модель пульпы неосуществима, но дает осуществимые результаты ⇐ Python

[Anonymous]

В настоящее время я работаю над кодом, который сможет решить наш график распределения каждый день, без необходимости заставлять кого-то в компании делать это каждое утро. В настоящее время у меня есть следующий код:

Код: Выделить всё

data = [
{
"Day": "2024-07-10",
"Location": "Depo",
"Pick-up": "A",
"Boxes": 11,
"Weekday": "Wednesday",
"Time to pickup": "12:30:00"
},
{
"Day": "2024-07-10",
"Location": "Depo",
"Pick-up": "B",
"Boxes": 2,
"Weekday": "Wednesday",
"Time to pickup": "12:30:00"
},
{
"Day": "2024-07-10",
"Location": "Depo",
"Pick-up": "C",
"Boxes": 5,
"Weekday": "Wednesday",
"Time to pickup": "12:00:00"
}
]

df_stint = pd.DataFrame(data)

# Parameters
bus_capacity = 10
num_buses = 6

model = pulp.LpProblem("Bus_Optimization", pulp.LpMinimize)

# Decision variables
pickup_vars = pulp.LpVariable.dicts("Pickup",
[(i, j) for i in range(num_buses) for j in range(len(df_stint))],
cat = 'Binary')

dropoff_vars = pulp.LpVariable.dicts("Dropoff",
[(i, k) for i in range(num_buses) for k in df_stint['Location'].unique()],
cat='Binary')

# Split variables
split_vars = pulp.LpVariable.dicts("Split",
[(i, j) for i in range(num_buses) for j in range(len(df_stint))],
lowBound=0,
cat='Integer')

# Objective: Minimize the number of buses used
model += pulp.lpSum([dropoff_vars[i, k] for i in range(num_buses) for k in df_stint['Location'].unique()])

# Constraint 1: each stint is picked up exactly once
for j in range(len(df_stint)):
model += pulp.lpSum(pickup_vars[i, j] for i in range(num_buses)) == 1

# Constraint 2: capacity constraint
for i in range(num_buses):
model += pulp.lpSum(df_stint.iloc[j]['Boxes'] * pickup_vars[i, j] for j in range(len(df_stint))) 

Подробнее здесь: [url]https://stackoverflow.com/questions/78737600/pulp-model-is-infeasible-but-gives-feasible-results[/url]