Linear Programming (PuLP) - loop for constraints

[littleangel83]

Hi,

I am new to Python and am trying to get it to solve an optimisation problem. I know how to use Excel to do this, but my problem is too big (too many decision variables and too many constraints) for Excel to manage. So I am trying to teach myself a bit of Python to solve the problem.

Basically I am trying to maximise the profit made across up to 2444 transactions throughout a year (attached to a date and hour of purchase - buy_date and buy_hour, as well as a date and hour of resale - sell_date and sell_hour). In each transaction I can purchase between 0 and 3 units (doesn't even have to be an integer value) and later sell between 0 and 3 units to make a profit. But in each of the 8760 hours in the year I can only store up to 5 units. So obviously I have 2444 decision variables which must be between 0 and 3, and 8760 constraints which must be between 0 and 5.

Please note, the 'constraints' list I have created is just a list of 1's to be used for the lpSum to be used when creating the loop for the 8760 constraints.

This is the code I have written which works:

#import data and store in lists

import csv

filename = 'OneDay3MWBattery5MWhStorage.csv'
with open(filename) as f:
    reader = csv.reader(f)
    header_row = next(reader)

    ID = []
    buy_date = []
    buy_hour = []
    sell_date = []
    sell_hour = []
    profit_unit = []
        
    from datetime import datetime

    for row in reader:
        ID.append(row[0])
        buy_date.append(datetime.strptime(row[1], "%d-%m-%y"))
        buy_hour.append(int(row[2]))
        sell_date.append(datetime.strptime(row[3], "%d-%m-%y"))
        sell_hour.append(int(row[4]))
        profit_unit.append(float(row[5]))
        
filename = 'dates2018.csv'
with open(filename) as f:
    reader = csv.reader(f)
    header_row = next(reader)
    
    date_solver = []
    time_solver = []
    constraints = []
    
    from datetime import datetime

    for row in reader:
        date_solver.append(datetime.strptime(row[0], "%d-%m-%y"))
        time_solver.append(int(row[1]))
        constraints.append(int(row[2]))
  
#store lists in dictionaries

buy_date_dict = {ID[i]: buy_date[i] for i in range(len(ID))} 
buy_hour_dict = {ID[i]: buy_hour[i] for i in range(len(ID))} 
sell_date_dict = {ID[i]: sell_date[i] for i in range(len(ID))} 
sell_hour_dict = {ID[i]: sell_hour[i] for i in range(len(ID))} 
profit_unit_dict = {ID[i]: profit_unit[i] for i in range(len(ID))} 

#set up linear programming problem

prob = LpProblem("Battery One Day",LpMaximize)

ID_vars = LpVariable.dicts("ID",ID,lowBound=0,upBound=3,cat='Continuous')

prob += lpSum([profit_unit_dict[i]*ID_vars[i] for i in ID])
       

This is the code which crashes:

#set up constraints


for j in range(len(date_solver)):
    for i in range(len(ID)):
        if buy_date[i] <= date_solver[j] <= sell_date[i] and buy_hour[i] <= time_solver[j] <= sell_hour[i]:
            prob += lpSum(constraints[j]*ID_vars[i]) >= 0
            prob += lpSum(constraints[j]*ID_vars[i]) <= 5

This is the error message I am getting:

Error:
KeyError                                  Traceback (most recent call last)
<ipython-input-84-c4a318de252e> in <module>
      5     for i in range(len(ID)):
      6         if buy_date[i] <= date_solver[j] <= sell_date[i] and buy_hour[i] <= time_solver[j] <= sell_hour[i]:
----> 7             prob += lpSum(constraints[j]*ID_vars[i]) >= 0
      8             prob += lpSum(constraints[j]*ID_vars[i]) <= 5

KeyError: 1665

Is there anyone who can please advise me on if, and how, this problem can maybe be fixed?

Thank you very much.

[littleangel83]

I have solved this myself. Turns out that the IDs were strings and not integers, so I converted them into integers when adding to the list. There were other problems in my code but I have fixed all the problems and the code seems to be doing what I wanted it to do. Yay!