Assortment Optimization for Boundedly Rational Customers

Abstract

This paper presents an assortment optimization model for boundedly rational customers. The problem has application in designing the travel menu for on-demand mobility services. We present the customer behavior using the Random Regret Minimization (RRM) choice model, considering the reference-dependency and choice-set dependency of preferences as strong violations of perfect rationality premises. We propose an efficient algorithm to find the optimal assortment when customers’ behavior follows RRM. We have tested our algorithm for micromobility services. The results show that our proposed algorithm can find the optimal solution for all studied instances. Moreover, we compare the planned assortments against the widely used multinomial logit model (based on the premise of full rationality) to examine the effects of reference-dependency and choice set-dependency on the assortment decisions. Our results indicate that these behavioral phenomena have significant impacts on the optimal choice set, so they need to be taken into account by those who want to offer a menu of options to their customers.