Bundling retailing under stochastic market

Abstract

This research studies bundling retailing problem under stochastic market. Compared to the conventional bundling study, the incorporation of inventory issue becomes significant when demands are uncertain. First, a two-stage stochastic programming model involving both pricing and ordering decisions has been built for two-product mixed bundling strategy under stochastic market. Reservation price model is adopted to demonstrate the relationship between prices and demands. Two different policies, namely non-sharing policy and sharing policy are proposed and compared. In the latter model, concavity in order quantities has been proved. The algorithm with Downhill Simplex Method has been proposed to search for the prices. Considerable numerical analysis was carried out to examine the effects of relevant factors, such as cost structure and demand variation, on the performance of mixed bundling. These results can serve as guidelines for practitioners who face particular market conditions. Second, based on the first work, many assumptions are relaxed and more realistic conditions are considered, including Multinomial Model Logit Model for customer choice behavior, joint reservation price distribution and different types of product pairs. Sample Average Approximation with IPA (Infinitesimal Perturbation Analysis) gradient estimator is used to solve this model, with extensive numerical studies for various parameters. The results provide more managerial insights regarding several important factors like correlation coefficient between the reservation prices for the individual products and degree of contingency. Third, a special type of bundling is analyzed using dynamic pricing. We defined the bundling consisting of an advertising component and a main component in terms of pricing effect. Closed-form results have been obtained and comparisons with some heuristics have been conducted.