MNL-Bandit: A Dynamic Learning Approach to Assortment Selection

We consider a dynamic assortment selection problem where in every round the retailer offers a subset (assortment) of N substitutable products to a consumer, who selects one of these products accord...