IIMA-Flipkart team solve eCommerce packaging problem, pick up award
The award recognizes the creative and unique application of a combination of analytical techniques in a new area
Researchers and other working professionals who had teamed up from the Indian Institute of Management Ahmedabad and Flipkart Internet Private Limited have recently secured a win at the 2022 Innovative Applications in Analytics Award (eCommerceIAAA) for their work entitled "A Data-driven Optimization Approach to Solve the E-commerce Packaging Problem."
The award, sponsored by the Analytics Society of INFORMS, Kinaxis and Adelphi University, recognizes creative and unique application of a combination of analytical techniques in a new area. The prize also promotes the awareness and value of the creative combination of analytics techniques in unusual applications to provide insights and business value.
Debjit Roy, associate professor of production and quantitative methods at the Indian Institute of Management Ahmedabad, presented the team's work during the 2022 INFORMS Business Analytics Conference in Houston, which was held from April 3-5.
The project developed approaches to determine the optimal packaging box assortment by employing a mixed-integer linear programming formulation that provides substantial gains in reducing the box assortment size, improving volumetric efficiency by 12%, reducing CO2 emissions by 8,000 metric tonnes, and resulted in a cost savings of about $3.7mn USD accrued over two years. Next, the team developed an innovative hybrid optimization framework combining unsupervised learning, reinforcement learning, and tree-search. This approach promises a further improvement of 5% in volumetric efficiency.
The winning team included faculty members Debjit Roy, Shanthan Kandula, and Srikumar Krishnamoorthy from IIM Ahmedabad, Sharvendu Bhushan, Himanshu Gupta, Chandrasekhar K, Rohan Nanaware, and Sandeep Sangwan from Flipkart Internet Private Limited.
"Flipkart carries a large assortment of SKUs, around 200 thousand+. So identifying the optimal assortment of packaging boxes for these SKUs is a big challenge. Using machine learning algorithms, we were able to design the box sizes that improve the packaging factor and reduce costs," IIMA faculty Debjit Roy told ITLN.
"Determining the optimal packaging box assortment is crucial for e-commerce platforms. Typically, an assortment of fewer than 50 boxes is used to cover a few hundred thousand SKUs. Since the number of boxes is far less than the number of SKUs, boxes rarely fit the SKUs accurately, resulting in empty space. Typically, the one-third volume of a typical package comprises air and filler material. The inefficient usage of space significantly increases operational costs and carbon footprint. In this project, we develop approaches to determine the optimal packaging box assortment in two phases," cited an abstract of their work.