Optimizing product choice for consumers

How can we present consumers with an optimal choice of products, both online and offline? A novel interdisciplinary research partnership that involves artificial intelligence and operations research, points to some effective solutions.



The success of a retail business is directly dependent on a limited number of product ranges on display in stores. But optimizing these ranges could prove to be a long and costly exercise. Moreover, introducing a new item with no previous sales track-record in the store could also be a real headache. To date, such challenges have been of primary concern to experts in operations research.


Online merchants have a huge variety of products, but customized recommender systems only propose a limited number of choices to the buyer. These suggestions may often lack complementarity and diversity. Also, if the supply in stock is significant, some relevant articles may not “stand out” effectively enough from the inventories. Recommender systems, such as those of Amazon or Netflix, were primarily developed by researchers with expertise in machine learning.


Laurent Charlin

“We innovate by exploring the synergies between machine learning and operations research, which were always seen as being two separate entities.”

Laurent Charlin, Assistant Professor, Department of Decision Sciences


Share this page:

Research > Digital transformation > Achievements > Optimizing product choice for consumers