You work for a logistics tech platform that manages container shipments from ports to…
- User
You work for a logistics tech platform that manages container shipments from ports to warehouses across regions. Each container arrives on a specific date with a mix of items, and each warehouse has a known daily capacity (volume, weight, units) and item-level demand for the next 3 days. Containers can only be sent to warehouses that are reachable from their origin port. Warehouses also have a priority score based on strategic importance.
Write a Python script to assign containers and their items to eligible warehouses over a 3-day delivery window, maximizing the total assigned_units * priority_score while strictly respecting capacity and reachability constraints. Output a final assignment report with container, warehouse, item, units assigned, and delivery date.
Assistant
I'll create a Python script to solve this logistics optimization problem. The goal is to assign containers to warehouses over a 3-day window while maximizing the product of assigned units and warehouse priority, subject to capacity and reachability constraints.
import pandas as pd import numpy as np from pulp import * import datetime # Function to solve the container assignment problem def optimize_container_assignments(containers, warehouses, items, reachability): """ Optimize container assignments to warehouses over a 3-day window Parameters: ----------- containers: DataFrame with container info (id, arrival_date, origin_port) warehouses: DataFrame with warehouse info (id, priority, daily capacities) items: DataFrame with item info (container_id, item_id, volume, weight, units) reachability: DataFrame defining which warehouses are reachable from which ports Returns: -------- DataFrame with optimized assignments """ # Create a mathematical optimization model model = LpProblem("Container_Assignment_Problem", LpMaximize) # Get unique dates (arrival date + 0, 1, or 2 days) start_date = min(containers['arrival_date']) dates = [start_date + datetime.timedelta(days=i) for i in range(3)] # Create decision variables # x[c,w,i,d] = units of item i from container c assigned to warehouse w on date d x = {} for _, container in containers.iterrows(): c_id = container['id'] arrival = container['arrival_date'] port = container['origin_port'] # Get eligible warehouses (those reachable from this container's port) eligible_warehouses = reachability[reachability['port_id'] == port]['warehouse_id'].tolist() # Get items in this container container_items = items[items['container_id'] == c_id] for _, item in container_items.iterrows(): i_id = item['item_id'] for w_id in eligible_warehouses: # Can only deliver on or after arrival date for date in [d for d in dates if d >= arrival]: x[(c_id, w_id, i_id, date)] = LpVariable( f"Assign_C{c_id}_W{w_id}_I{i_id}_D{date.strftime('%Y%m%d')}", lowBound=0, upBound=item['units'], cat='Integer' ) # Objective function: maximize sum of (assigned units * warehouse priority) objective = [] for key, var in x.items(): c_id, w_id, i_id, date = key priority = warehouses[warehouses['id'] == w_id]['priority'].values[0] objective.append(var * priority) model += lpSum(objective) # Constraint 1: Don't assign more units of an item than available in the container for _, item in items.iterrows(): c_id = item['container_id'] i_id = item['item_id'] # Sum of assignments across all warehouses and dates assignments = [] for key, var in x.items(): if key[0] == c_id and key[2] == i_id: assignments.append(var) if assignments: # Only add constraint if there are variables for this item model += lpSum(assignments) <= item['units'], f"Item_limit_{c_id}_{i_id}" # Constraint 2: Respect warehouse daily capacity constraints (volume, weight, units) for w_id in warehouses['id'].unique(): w_data = warehouses[warehouses['id'] == w_id].iloc[0] for date in dates: # Volume constraint volume_sum = [] # Weight constraint weight_sum = [] # Units constraint units_sum = [] for key, var in x.items(): c_id, wh_id, i_id, d = key if wh_id == w_id and d == date: # Get item data item_data = items[(items['container_id'] == c_id) & (items['item_id'] == i_id)].iloc[0] volume_sum.append(var * item_data['volume']) weight_sum.append(var * item_data['weight']) units_sum.append(var) if volume_sum: # Only add constraints if there are variables for this warehouse/date model += lpSum(volume_sum) <= w_data['volume_capacity'], f"Volume_{w_id}_{date.strftime('%Y%m%d')}" model += lpSum(weight_sum) <= w_data['weight_capacity'], f"Weight_{w_id}_{date.strftime('%Y%m%d')}" model += lpSum(units_sum) <= w_data['units_capacity'], f"Units_{w_id}_{date.strftime('%Y%m%d')}" # Constraint 3: Respect warehouse demand for specific items for w_id in warehouses['id'].unique(): for _, demand in warehouses[warehouses['id'] == w_id]['item_demand'].items(): for item_id, max_units in demand.items(): # Sum of assignments for this item to this warehouse across all containers and dates item_assignments = [] for key, var in x.items(): c_id, wh_id, i_id, d = key if wh_id == w_id and i_id == item_id: item_assignments.append(var) if item_assignments: # Only add constraint if there are variables for this item/warehouse model += lpSum(item_assignments) <= max_units, f"Demand_{w_id}_{item_id}" # Solve the model model.solve(PULP_CBC_CMD(msg=False)) # Check if a solution was found if LpStatus[model.status] != 'Optimal': print(f"No optimal solution found. Status: {LpStatus[model.status]}") return pd.DataFrame() # Extract results results = [] for key, var in x.items(): if value(var) > 0: # Only include positive assignments c_id, w_id, i_id, date = key results.append({ 'container_id': c_id, 'warehouse_id': w_id, 'item_id': i_id, 'delivery_date': date, 'units_assigned': int(value(var)) }) # Create results dataframe assignments_df = pd.DataFrame(results) # Calculate total objective value total_value = 0 for _, row in assignments_df.iterrows(): priority = warehouses[warehouses['id'] == row['warehouse_id']]['priority'].values[0] total_value += row['units_assigned'] * priority print(f"Total objective value: {total_value}") return assignments_df # Function to generate a comprehensive assignment report def generate_assignment_report(assignments, containers, warehouses, items): """ Generate a detailed report of the assignments Parameters: ----------- assignments: DataFrame with the optimization results containers, warehouses, items: Original input DataFrames Returns: -------- DataFrame with detailed assignment information """ # Merge with container data report = assignments.merge( containers[['id', 'origin_port']], left_on='container_id', right_on='id', how='left' ).drop(columns=['id']) # Merge with warehouse data report = report.merge( warehouses[['id', 'priority']], left_on='warehouse_id', right_on='id', how='left' ).drop(columns=['id']) # Merge with item data report = report.merge( items[['container_id', 'item_id', 'volume', 'weight']], on=['container_id', 'item_id'], how='left' ) # Calculate value contribution (units * priority) report['value_contribution'] = report['units_assigned'] * report['priority'] # Calculate volume and weight assigned report['volume_assigned'] = report['units_assigned'] * report['volume'] report['weight_assigned'] = report['units_assigned'] * report['weight'] # Reorder columns for better readability report = report[[ 'container_id', 'origin_port', 'warehouse_id', 'priority', 'item_id', 'units_assigned', 'volume_assigned', 'weight_assigned', 'delivery_date', 'value_contribution' ]] return report # Example of how to use these functions def main(): # Sample data (in a real scenario, this would be loaded from files or a database) # Define sample dates today = datetime.date.today() tomorrow = today + datetime.timedelta(days=1) # Sample container data containers = pd.DataFrame({ 'id': [1, 2, 3], 'arrival_date': [today, today, tomorrow], 'origin_port': ['Port_A', 'Port_B', 'Port_A'] }) # Sample warehouse data with item-specific demand warehouses = pd.DataFrame({ 'id': [101, 102, 103], 'priority': [3, 2, 1], 'volume_capacity': [1000, 1500, 2000], 'weight_capacity': [2000, 3000, 4000], 'units_capacity': [500, 700, 1000], 'item_demand': [ {'item1': 200, 'item2': 150, 'item3': 100}, {'item1': 100, 'item2': 200, 'item3': 300}, {'item1': 300, 'item2': 250, 'item3': 200} ] }) # Sample item data items = pd.DataFrame({ 'container_id': [1, 1, 2, 2, 3, 3], 'item_id': ['item1', 'item2', 'item1', 'item3', 'item2', 'item3'], 'volume': [2, 3, 2, 4, 3, 4], 'weight': [5, 7, 5, 10, 7, 10], 'units': [100, 150, 200, 120, 180, 90] }) # Sample reachability data reachability = pd.DataFrame({ 'port_id': ['Port_A', 'Port_A', 'Port_B', 'Port_B', 'Port_B'], 'warehouse_id': [101, 102, 101, 102, 103] }) # Run the optimization assignments = optimize_container_assignments(containers, warehouses, items, reachability) # Generate and print the report if not assignments.empty: report = generate_assignment_report(assignments, containers, warehouses, items) print("\nAssignment Report:") print(report) # Summary statistics print("\nSummary Statistics:") print(f"Total containers: {containers.shape[0]}") print(f"Total items assigned: {assignments.shape[0]}") print(f"Total units assigned: {assignments['units_assigned'].sum()}") # Warehouse utilization print("\nWarehouse Utilization:") warehouse_stats = report.groupby('warehouse_id').agg({ 'units_assigned': 'sum', 'volume_assigned': 'sum', 'weight_assigned': 'sum', 'value_contribution': 'sum' }).reset_index() for _, w in warehouse_stats.iterrows(): w_data = warehouses[warehouses['id'] == w['warehouse_id']].iloc[0] print(f"Warehouse {w['warehouse_id']} (Priority {w_data['priority']}):") print(f" Units: {w['units_assigned']} / {w_data['units_capacity']} ({w['units_assigned']/w_data['units_capacity']*100:.1f}%)") print(f" Volume: {w['volume_assigned']} / {w_data['volume_capacity']} ({w['volume_assigned']/w_data['volume_capacity']*100:.1f}%)") print(f" Weight: {w['weight_assigned']} / {w_data['weight_capacity']} ({w['weight_assigned']/w_data['weight_capacity']*100:.1f}%)") print(f" Value contribution: {w['value_contribution']}") else: print("No feasible assignment found.") if __name__ == "__main__": main()Would you like me to explain or break down the code?
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.