A data-driven optimization model for e-commerce based on hybrid ALNS-PSO algorithms

In order to gain a competitive advantage among emerging peers, e-commerce companies must not only design new products favored by consumers, but also consider product supply chain services to achieve better business results. This paper develops a data-driven optimization model based on logistic regression and multi-objective programming to provide a practical low-cost solution for e-commerce supply chain optimization. As decision maker, e-retailers, the key to decision-making lies in choosing a supplier with a suitable geographical location and low production cost from a large number of homogeneous suppliers. This paper analyzes consumer transaction data, obtains the correlation between various types of products and geographic districts, and matches suitable suppliers for different regions to reduce costs. Combining adaptive large neighborhood search (ALNS) and particle swarm optimization (PSO) algorithm, a hybrid ALNS-PSO is designed to solve the proposed data-driven model. A practical case proved the effectiveness and practicality of the proposed model.