Packaging Material Recommendation Algorithms for E‑commerce Fulfillment

In e‑commerce order fulfillment, packaging (boxing or bagging) is essential but often leads to material waste and low space utilization.

The article outlines a packaging material recommendation system that selects the optimal packaging type and size based on product dimensions and business rules.

Algorithm architecture : two rule‑based filters (material type and size constraints) followed by a size‑matching algorithm that chooses the smallest box or bag capable of containing all items.

Algorithm principles : the problem is modeled as a three‑dimensional bin‑packing (NP‑hard) task. Exact solutions use integer programming with constraints to avoid overlap and ensure items stay within the container. Heuristic methods include tree‑search with rank evaluation, Largest Area Fit First, and brute‑force enumeration, offering faster but approximate results.

Derived applications : (1) Box‑cutting – splitting large orders into multiple parcels; (2) Merging – combining orders for the same address; (3) Box‑type design – re‑optimizing box dimensions from billions of historical orders; (4) Packaging‑scheme recommendation – using multimodal AI to match product characteristics (fragile, flammable, high‑value) with suitable protective packaging.

The system aims to reduce packaging cost, improve space utilization, and lower damage rates, while relying on accurate data, rule maintenance, and continuous performance monitoring.