Online Grocery Shopping

Challenge

In an online grocery shopping application, people order groceries online and pickers go to the shopping center to pick up the requested items and drive them to the customer's house. The project focused only on operations inside the shopping center. The main objective was to optimize the routing and picking methods of items.

Solution

A simulation model was built in AnyLogic using discrete-events. Real data of past orders and the store layout were used to validate the model, with output validation done using Python. The optimization used the concept of genetic algorithms.

Optimization Concepts Tested

In addition to the single order optimization, two other concepts were tested:

• Batching optimization: Batches of orders are collected before picking, so the picker picks several orders at the same time in one store visit
• Zoning optimization: The store layout is divided into zones, with items picked per zone. Optimization is done on 2 levels: optimizing item picking order within a zone and optimizing the order of zones

Finally, a combination of both concepts was tested.

Part of the work also involved data cleaning, analysis, and input calibration, as the real data had many problems. All optimization analysis was conducted using Python.

Outcome

The analysis showed that the mixed method between batching and zoning leads to the minimum mean picking time per item. The model itself can also be used as a testing tool to optimize future orders and find the best picking route.

Project Features

• Industry: Commercial
• Model: Discrete Events
• Duration: 2 weeks