A DUAL RELAXATION BOUND TO A MIXED LOGIT PRICING MODEL

초록

Determining optimal pricing for multiple substitutable products across mixed customer segments is a fundamental objective for businesses aiming to sustain competitiveness. This paper focuses on the study of optimal pricing strategies for multiple substitutable products within a fixed time horizon, while accounting for the heterogeneity of customer preferences across different segments. To model this heterogeneity, a mixed logit model is utilized. Pricing optimization with the overlapping multiple segments is a difficult nonconvex maximization problem. We develop a binary fractional programming model that computes relaxation upper bound for the mixed logit pricing problem by disaggregating and discretizing price vectors. We next develop an effective Lagrangian dual formulation that uses simple Newton search to compute dual function. For randomly generated problem instances, we observed that the proposed model achieves within one percent of the optimal value for small sized problem, and it consistently generates tight upper bound for bigger sized problems within 500 CPU seconds. ICIC International © 2024.

키워드