Geographical relevance-based multi-period optimization for e-commerce supply chain resilience strategies under disruption risks

With the fierce competition in e-commerce, e-tailers are required to rapid responses to a variety of customised orders with multiple frequencies and strict delivery times. The delay or insufficient supply caused by disruptions might result in lost sales during long-term processes. To address this problem, a two-stage stochastic programming model considering profits, consumer service level (CSL) as well as market priorities is developed to manage long-term disruptions. We analyse multi-period consumer transaction data and formulate geographical relevance (GR) to link each marketplace with historical data in related regions and then prioritise market segments. A GR-based two-stage stochastic programming with multi-period is proposed, which (1) considers both proactive mitigation decisions before disruption and reactive recovery plans after disruption; (2) collaborates three resilience strategies; (3) optimises the e-tailer's profits considering market priorities during long-term disruptions. Using a real case of Chinese e-commerce under the COVID-19 pandemic, it is illustrated (1) the applicability and performance of the proposed GR-based model for multi-period resilience optimisation improving both the CSL and the total profit; (2) the efficiency and robustness of the developed sequential particle swarm optimisation with social structures algorithm. The proposed method could optimise e-tailers' response strategies for managing long-term disruptions in practice.