A branch-and-cut algorithm for the latent-class logit assortment problem

摘要: We study the product assortment problem of a retail operation that faces stream customers who are heterogeneous with respect to preferences. Each customer belongs market segment characterized by consideration set includes alternatives viewed as options, and preference weights assigns each those alternatives. Upon arrival, he checks offer displayed firm, either chooses one products or quits without purchasing according multinomial-logit (MNL) criterion. The firm's goal is maximize expected revenue extracted during fixed time horizon. This also arises in growing area choice-based, network management, where computational speed critical factor for practical viability solution approach. so-called latent-class, logit known be NP-Hard. In this paper, we analyze unconstrained constrained (i.e., limited number display) versions it, propose branch-and-cut algorithm computationally fast leads (nearly) optimal solutions.