Assortment Optimization
Overview
Assortment optimization studies which products to offer, and sometimes where to place them, when customers choose among the offer and an outside option. Methods model choice with multinomial logit, ranking-based, and Markov-chain models, then use combinatorial optimization to handle inventory, capacity, diversity, and placement constraints. Online bandit methods learn customer preferences while balancing exploration and revenue, including settings with contextual choices, pricing, and changing arrivals. A central challenge is making optimization scalable while accounting for imperfect choice models, heterogeneous markets, and the cost of learning through real decisions.
Latest papers
15 papers0 papers in the last four weeks, against 2 the four weeks before. 0.0% of all new papers.