Bi-objective assortment optimization under a ranking-based choice model: formulation and solution approach using NSGA-II

Assortment planning is a critical component of revenue management in the retail sector. This study delves into a bi-objective assortment optimization issue within a ranking-based customer choice model. We introduce an integer programming formulation designed to optimize both expected revenue and customer satisfaction. Given the NP-hard classification of the problem, a multi-objective optimization method, namely, a fast, non-dominated sorting genetic algorithm (NSGA-II), is employed to find the Pareto-optimal front for large-sized problems. Our validation process, comprising extensive numerical experiments, underscores the algorithm's efficacy and the model's robustness. The results confirm our model's capability to enhance retailers' revenue prospects while fulfilling customer satisfaction.