Rough approximation‐based inventory optimization with random prices for B2C companies during the promotion period

Abstract

Most B2C e-commerce giants such as Amazon, Bestbuy, Alibaba, periodically or occasionally conduct promotional activities that put significant stress on their inventory systems. The sharp demand and price fluctuations caused by these promotional activities distress stocks over an inventory period if customer satisfaction is maintained. This article develops a chance-constrained multiobjective programming model to obtain optimal order quantities in an e-commerce inventory system with stochastic demand and prices. Traditional techniques for coping with the stochastic coefficients using an expected value operator or probability measures are usually inefficient in estimating quantities. With rational assumptions regarding market behavior, stochastic demand and price are characterized with a normal distribution based on historical data. A novel method called rough approximation is designed to flexibly formulate the feasible region, from which a crisp model is obtained that has no uncertain information. A proxy ideal point-based genetic algorithm (PIP-GA) is then developed to solve the model, allowing decision makers to fix the optimal order quantities for different promotional activity scales. Finally, the results are illustrated using a practical example from the Chinese B2C e-commerce giant “JD.COM.” Some special features are discussed to show the availability of the proposed model and the solution methodology.