Automated Pizza Store Delivery Planner
logistics
OR-Tools
ai
Vehicle Routing Problem
Optimization
Summary
For delivery business time is crucial to win the market share. Many complex factors impact delivery time like vehicle capacity, cargo size and weight, driver availability and many more. To solve this we created a solution for delivery companies, which includes adjustable real-time routing planner, vehicle and driver assignment and advanced reporting system.
Project overview
In a competitive logistic business, speed and efficiency are paramount. Our clients sought to optimise their delivery operations to minimise delivery times, decrease delivery costs, and increase customers' experience. This case study details how Postdata leveraged and tackled the Traveling Salesperson Problem (TSP) to develop an automated pizza store delivery planner, significantly improving delivery performance.
Postdata was entrusted with developing an intelligent delivery planning system to optimize logistics. The goal was to automate assigning orders to drivers and generating optimised delivery routes, considering factors such as order locations, delivery time windows, driver availability, and traffic conditions. The system needed to integrate seamlessly with the existing order management system and provide real-time updates to dispatchers and drivers.
Data Preprocessing
Geocoding: Customer addresses were geocoded to obtain latitude and longitude coordinates, enabling accurate distance calculations.
Time Window Definition: Delivery time windows were defined for each order, specifying the acceptable delivery timeframe.
Driver Availability: Driver availability was integrated into the system, including start and end times, and any breaks.
Traffic Data Integration: Real-time or historical traffic data could be integrated to further refine route optimization.
Techniques Used
OR-Tools: OR-Tools is an open- source software suite for optimization, tuned for tackling the world's toughest problems in vehicle routing, flows, integer and linear programming, and constraint programming.
Vehicle Routing Problem (VRP): For scenarios with multiple drivers, the system addressed the VRP, which extends the TSP to assign multiple vehicles to serve a set of customers.
Traveling Salesperson Problem (TSP) as a Subproblem: Within the VRP, the route for each individual courier can be considered a TSP. Once the VRP solver has assigned a set of delivery points to a specific courier, the problem of finding the best order for that courier to visit those points becomes a TSP.
Heuristics and Optimization Algorithms: OR-Tools provides various algorithms for solving TSP and VRP, including heuristics like Local Search and Metaheuristics like Simulated Annealing, allowing for a balance between solution quality and computation time.
Results
Reduced Delivery Times: Optimized routes significantly decreased average delivery times, leading to fresher pizzas and happier customers.
Increased Driver Efficiency: Improved driver utilization allowed for more deliveries per driver per shift, maximizing resource allocation.
Lower Fuel Costs: Shorter routes translated to lower fuel consumption, reducing operational costs.
Improved Customer Satisfaction: Faster and more reliable deliveries resulted in higher customer satisfaction and increased repeat business.
Real-time Tracking and Dispatching: The system provided dispatchers with real-time visibility into driver locations and delivery progress, enabling proactive management and issue resolution.
Project duration:
1 week
Team
2 people
1 Full-stack developer, 1 Data Scientist
Technologies
TypeScript, Python, OR-Tools, Linear Programming
Tech challenge
Handling Real-time Data: Integrating with the order management system and incorporating real-time data like traffic conditions and new orders posed a challenge.
Scalability: The system needed to be scalable to handle peak order volumes and a growing number of drivers.
Asymptotic complexity of the problem: Solving the TSP and VRP, especially with constraints like time windows and driver availability, requires sophisticated algorithms and optimisation techniques.
Solution
Postdata successfully developed an automated pizza store delivery planner that leverages OR-Tools to solve the TSP and VRP. This solution has significantly improved delivery efficiency, reduced costs, and enhanced customer satisfaction for Pizza Store. The system's real-time tracking and dispatching capabilities give dispatchers valuable insights and control over delivery operations.
