AI summary · arXiv: 2202.04277v1 · cs.LG

A decision-tree framework to select optimal box-sizes for product shipments

Karthik S. Gurumoorthy, Abhiraj Hinge
Original paper Submitted to arXiv on 09 Feb. 2022 License: CC BY-NC-ND 4.0

How to read this summary? Illustrated guide

The essentials

  • Use of improperly sized boxes in package-handling facilities can lead to wastage and increased shipping costs
  • E-commerce companies face the challenge of determining the optimal number and dimensions of cuboidal boxes to minimize wastage and shipping costs
  • Proposed solution focuses on single-count shipments with one product per box
  • Step 1: Reduce problem to a clustering problem in a three-dimensional space defined by length, width, and height
  • Step 2: Present an efficient forward-backward decision tree based clustering method to obtain desired number of clusters and corresponding box dimensions
  • Proposed algorithm allows for incremental generation of clusters without discarding current solution
  • Simulations using actual single-count shipments transported by Amazon showed promising results:
  • Modification of existing box dimensions led to a reduction of 4.4% in shipment volume and decrease in non-utilized air volume space by 2.2%
  • Introduction of four additional boxes improved shipment volume reduction to 10.3% and air volume improvement to 6.1%

Key points written by AI: they may contain errors, check them against the paper.

AI-generated keywords: E-commerce Box Sizing Clustering Decision Tree Optimization

Read further

  • Comprehensive Summary
  • Layman's Summary
  • Blog article
  • Original abstract
In package-handling facilities, the use of boxes with varying sizes to ship products can lead to wastage and increased shipping costs when improperly sized boxes are used. This issue becomes even more challenging for e-commerce companies that deal with a large number of products. The main question they face is how many cuboidal boxes should be manufactured and what should be their dimensions in order to minimize wastage and shipping costs. To address this problem, the authors propose a solution specifically designed for single-count shipments where each box contains only one product. The solution consists of two steps: 1. The first step involves reducing the problem to a clustering problem in a three-dimensional space defined by length, width, and height. Each cluster represents a group of products that will be shipped in a particular size variant. 2. In the second step, the authors present an efficient forward-backward decision tree based clustering method with low computational complexity on N (the total number of products) and K (the desired number of clusters). This method allows them to obtain K clusters along with their corresponding box dimensions. The proposed algorithm is designed to generate clusters incrementally without discarding the current solution. This means that adding or deleting a size variant can be easily achieved by stopping the backward pass early or executing it for one more iteration. To evaluate the efficacy of their approach, the authors conducted simulations using actual single-count shipments transported by Amazon over a month. They used the proposed box dimensions and compared the results with existing box dimensions without introducing any new size variants. The simulations showed promising results; by simply modifying the existing box dimensions they achieved a reduction of 4.4% in shipment volume leading to a decrease in non-utilized air volume space by 2.2%. When four additional boxes were introduced, both shipment volume reduction and air volume improvement improved significantly to 10.3% and 6.1%, respectively.

Was this synthesis useful?

Score: 0

Why do we need votes?

Votes are used to determine whether we need to re-run our summarizing tools. If the count reaches -10, our tools can be restarted.

Figures and plots

Ask about a figure by number or use its question button to request visual analysis. Up to two figures and four image panels per question. Values read from plots are approximate; check the original data.

AI assistant instructions?

Tools for this paper

Explain it differently 1 free rewrite left today
Save to a collection Group papers in collections to compare them, chat with them and export them. Continue with Google Create a free account Log in

Advanced tools Premium

For researchers who follow a field closely: citation alerts, journal-club slides, version comparison and Zotero export. Discover Premium

Zotero Zotero guide
Citation alerts Citation alerts guide
Who works on it?

Among the 23 papers around this one, the name that comes back most: S. Y. Leung (2 papers)

2 more people sign several of them. The full map adds their institutions and approaches, from the papers of the last years.

Continue with Google Create a free account: your first map is free
Weekly letter The latest letter of its field, "Artificial intelligence this week" of 16 November 2026: Artificial intelligence this week: What impressive scores leave out · Receive one by email for free

Papers around this one

See them on a map

Ranked by how close their title and abstract are to this paper’s. Select some to read them together.

Closest papers

  1. Very close
    A Machine Learning Approach to Solve the E-commerce Box-Sizing Problem
    S. Kandula, Debjit Roy, K. Akartunalı · 2024 · Production and operations management · cited 4 times · Not summarized yet
    Abstract

    E-commerce packages are notorious for their inefficient usage of space. More than one-quarter volume of a typical e-commerce package comprises air and filler material. The inefficient usage of space significantly reduces the transportation and distribution capacity increasing the operational costs. Therefore, designing an optimal set of packaging box sizes is crucial for improving efficiency. We present the first learning-based framework to determine the optimal packaging box sizes. In particular, we propose a three-stage optimization framework that combines unsupervised learning, reinforcement learning, and tree search to design box sizes. The package optimization problem is formulated into…

  2. Close Cited by this paper
    A Machine Learning Approach to Shipping Box Design
    Guang Yang, Cun Mu · 2018 · Intelligent Systems with Applications · cited 4 times · Not summarized yet
    Abstract

    Having the right assortment of shipping boxes in the fulfillment warehouse to pack and ship customer's online orders is an indispensable and integral part of nowadays eCommerce business, as it will not only help maintain a profitable business but also create great experiences for customers. However, it is an extremely challenging operations task to strategically select the best combination of tens of box sizes from thousands of feasible ones to be responsible for hundreds of thousands of orders daily placed on millions of inventory products. In this paper, we present a machine learning approach to tackle the task by formulating the box design problem prescriptively as a generalized version o…

  3. Close
    A Branch-and-Repair Method for Three-Dimensional Bin Selection and Packing in E-Commerce
    P. Fontaine, S. Minner · 2022 · Operational Research · cited 25 times · Not summarized yet
    Abstract

    Selecting a Parcel Type Portfolio to Reduce Unused Space in Transportation Wrongly sized parcels lead to unused space and inefficient transportation. With continuously increasing e-commerce and last-mile delivery volumes, available parcel types at a warehouse can significantly impact unused space that is transported. In “A Branch-and-Repair Method for Three-Dimensional Bin Selection and Packing in E-Commerce,” Fontaine and Minner solve the trade-off between cost of unused space and cost of parcel variety through optimizing the portfolio of available parcel types. To solve large instances with millions of binary decision variables, the authors develop an exact decomposition method that allows…

  4. Close* Cited by this paper
    Carton Set Optimization in E‐commerce Warehouses: A Case Study
    Manjeet Singh, Ehsan Ardjmand · 2020 · Journal of Business Logistics · cited 14 times · Not summarized yet
  5. Close
    Optimizing Clustering Techniques for Retail Industry: One-Dimensional Time Warping Method for Demand Uncertainty
    Mochamad Nizar Palefi Ma'ady, P. Meilanitasari, Nisa Isrofi et al. · 2023 · International Conference on Sustainable Information Engineering and Technology · cited 1 time · Not summarized yet
    Abstract

    Product availability is a crucial factor in maintaining customer satisfaction and securing revenue streams for retailers and suppliers. By clustering products based on demand patterns and inventory turnover, retailers can enhance efficiency and improve customer satisfaction. This study proposes optimizing product clustering by comparing dynamic time warping and k-means methods to find the best clustering solution. The results show that dynamic time warping outperforms the k-means algorithm, with a higher Silhouette score of 0.8406 compared to 0.7578. This indicates that our proposed method is more efficient in optimizing product clustering. Furthermore, the time complexity is faster compared…

  6. Close
    Predicting Packaging Sizes Using Machine Learning
    Michael Heininger, R. Ortner · 2022 · Operations Research Forum · cited 5 times · Not summarized yet
    Abstract

    The increasing rate of e-commerce orders necessitates a faster packaging process, challenging warehouse employees to correctly choose the size of the package needed to pack each order. To speed up the packing process in the Austrian e-commerce company niceshops GmbH, we propose a machine learning approach that uses historical data from past deliveries to predict suitable package sizes for new orders. Although for most products no information regarding the volume is available, using an approximate volume computed from the chosen packages of previous orders can be shown to significantly increase the performance of a random forest algorithm. The respective learned model has been implemented int…

  7. Close
    Automated Clustering and Vehicle Dispatching Service for Bulky Product Delivery
    R. R. Asri, Aria Bisma Wahyutama, Mintae Hwang et al. · 2026 · 2026 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology (IAICT) · Not summarized yet
    Abstract

    This study proposes a web-based decision-support system for automated clustering and dispatching in bulky product delivery planning. In practice, delivery preparation processes, including order grouping, demand aggregation, vehicle dispatch planning, and driver coordination, are still performed manually, which creates operational inefficiencies and limits scalability as order volumes continue to grow. The proposed system integrates geocoding-based distance clustering using the Haversine formula with rule-based vehicle dispatching approach to generate feasible delivery plans under operational constraints. The system is implemented using Python and PostgreSQL with a web-based interface for dat…

  8. Close
    Implementasi Algoritma K-Means Clustering untuk Pengelompokan Produk E-Commerce Berdasarkan Harga, Diskon, dan Total Revenue
    R. Pasaribu, Saidi Ramadan Siregar · 2026 · TIN: Terapan Informatika Nusantara · Not summarized yet
    Abstract

    The rapid growth of e-commerce has generated a large volume of transactional data; however, this data has not been fully utilized to support strategic decision-making, particularly in product segmentation. The main problem addressed in this study is the absence of a systematic product grouping approach based on key attributes such as price, discount, and revenue, which leads to less effective pricing and promotional strategies. Therefore, this study aims to analyze product sales patterns and cluster e-commerce products based on the characteristics of price, discount_percent, and total_revenue. The dataset used is an Amazon-style e-commerce dataset consisting of 50,000 transaction records and…

Prior work it builds on

  1. Related* Used in depth by this paper A method it uses
    Carton Box Optimization Problem of VMI-based Apparel Supply Chain
    W. Wong, S. Y. Leung · 2006 · 2006 IEEE International Conference on Management of Innovation and Technology · cited 3 times · Not summarized yet
  2. Related* A method it uses
    Optimal Size Selection in Standardization: A Case Study
    C. Bongers · 1982 · Journal of the Operational Research Society · cited 15 times · Not summarized yet
  3. Related* A method it uses
    An Integer Programming Approach to the Apparel Sizing Problem
    Peter Tryfos · 1986 · Journal of the Operational Research Society · cited 53 times · Not summarized yet
  4. Related* A method it uses
    On the Optimal Sizing Problem
    R. V. V. Vidal · 1994 · Journal of the Operational Research Society · cited 3 times · Not summarized yet
  5. Related
    A Branch-and-Cut Algorithm for the Latent Class Logit Assortment Problem
    I. Méndez-Díaz, J. Miranda-Bront, Gustavo J. Vulcano et al. · 2010 · Electron. Notes Discret. Math. · cited 95 times · Not summarized yet
    Abstract

    Abstract We study the product assortment problem of a retail operation that faces a stream of customers heterogeneous with respect to preferences. Upon arrival, each customer checks the offer set displayed by the firm, and either chooses one of those products according to a multinomial-logit (MNL) criterion or quits without purchasing. The firm's goal is to maximize the expected revenue extracted from each customer. The general version of the logit assortment problem is known to be NP-Hard. In this paper, we analyze uncapacitated and capacitated (i.e., with a limited number of products to display) versions of it, and propose a branch-and-cut algorithm that is computationally feasible and lea…

How are these papers found?

We compare the title and abstract of this paper with those of every paper summarized on this site, and with the papers Semantic Scholar recommends, those it cites and those citing it (for a much-cited paper, its most recent citations and, from OpenAlex, the most cited ones). Among the prior work and the follow-ups, the papers that use each other in depth, a method it uses and the much-cited follow-ups come a little higher. The levels (very close, close, related) come from the similarity of the texts, not from a judgment of quality; * marks a paper without an abstract, compared by its title. The sentence under a paper is written by an AI from the two abstracts: check it in the papers. A paper marked not relevant by three readers leaves the list.

Updated on 4 November 2026.

The illustrated guide to the papers around a paper

The previous summary was created more than a year ago and can be re-run (if necessary) by clicking on the Run button below.

Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.