Abstract: Nowadays, in the era of economic globalization, people attach importance to the concept of lean production, coupled with the frequent occurrence of natural disasters such as earthquakes, all of which bring great pressure to the nodes of the supply chain and exacerbate the vulnerability of the supply chain. Since 2019, the outbreak of COVID-19 has been a major public health event whose future long-term nature and severity cannot be predicted, causing ripple effects along the supply chain, leading to disruptions at both upstream and downstream nodes. Therefore, in this context, without proper planning, enterprises will bear high maintenance costs and loss costs, and even face serious consequences such as reduced customer satisfaction and distrust. Therefore, under this condition, resilient supply chain emerges. Supply chain resilience is defined as a dynamic ability to respond to and recover from disruptions, so that the supply chain can effectively adapt to the risks of disruptions, thus increasing the competitive advantage of enterprises. Therefore, on the one hand, designing a resilient supply chain can effectively alleviate the risks caused by supply chain disruption and ensure the sustainable development of enterprises in the future. On the other hand, studying the optimal resilient supply chain for the normal production operation of enterprises can reduce unnecessary redundancy, which is conducive to reducing the cost pressure of the overall operation of enterprises and supply chain. Therefore, it is of research significance for enterprises to establish resilient supply chain to reduce cost risk.
Therefore, this paper considers the construction of resilient supply chain through proactive investment and post-recovery measures. It discusses demand fluctuations and different customer types when designing supply chain scenarios. A constraint programming model is established to minimize the total cost of supply chain. The effectiveness coefficient of resilient supply chain is designed according to the time of interruption and the total cost of supply chain. Whether the established resilient supply chain is reasonable under different scenarios is evaluated. A resilient index is designed to evaluate the resilient capacity of different resilient supply chains in the face of disruptions by combining the disruption time and disruption degree. Through the analysis of these two coefficients, the state of resilient supply chain under different scenarios is comprehensively evaluated. In the decision-making process, different demand fluctuations and different customer types will affect the total cost of the supply chain, and different customer types and different resilience measures will affect the selection of supply chain nodes, which provides a theoretical reference for enterprises at the risk of the epidemic to establish a resilient supply chain.
We use genetic algorithm to solve the model by referring to supply chain related article data for example analysis. The results of example analysis show that, at the risk of interruption of the supply chain network, although the resilience of the supply chain will be the strongest when new customers are selected with a high probability, the cost of maintaining the normal operation of the supply chain is also the most. Although the loss cost of choosing old customers is greater than that of choosing new customers, by comparing these three scenarioscomprehensively, the total cost and λ of choosing old customers with high probability are the smallest. Therefore, in order to maximize the economic benefits of the supply chain, it is the best choice to supply old customers with a high probability. On this basis, customer types are further subdivided according to acceptable interruption time and shortage penalty cost, and the cost changes in different customer types and different demand fluctuations are analyzed. And the resilient coefficient and the effectiveness coefficient of resilient supply chain are analyzed to further analyze the influence of different customer types on the target. The results of the example show that when the demand fluctuation is small, it will be of the most economic value to choose the customer with small fixed cost and low penalty cost. Moreover, the resilient supply chain of this kind of customers is also more effective. However, the acceptable interruption time of this kind of customers is short, and the resilience of the supply chain is poor, which cannot meet the supply in a short time. In the early stage of interruption, it can have higher economic benefits, but it is not conducive to the long-term development of supply chain.
In future studies, more uncertain factors can be considered based on the realistic background, and the algorithm in this paper can be further optimized to get more accurate results in more complex supply chain network resilience design studies.

Key words: resilient supply chain, multi-level interruption, demand fluctuation, customer type

摘要： 新冠疫情在全球范围内蔓延阻碍了节点供应正常运转,造成供应链发生多级中断,由此导致产品的供货量、产品质量、经济目标等与预定目标发生偏差,对企业造成严重影响。为了有效应对供应中断问题,考虑通过事前主动投资和事后恢复措施构建弹性供应链,在供应链场景设计时考虑需求的波动、不同的客户类型,以最小化供应链总成本为目标建立约束规划模型,根据中断时间设计建立弹性供应链的有效性系数,在不同情景下评估建立弹性供应链的有效性。结合中断恢复时间和中断程度设计了一个弹性指标,用来评估不同弹性供应链面对中断时的弹性能力。通过对比分析这两个系数,综合评价不同情景下弹性供应链的状态。最后进行算例分析,通过遗传算法进行求解,算例分析结果显示,在决策时考虑不同顾客类型和多种修复措施对处于疫情风险下的企业建立弹性供应链具有重要现实意义。

关键词: 弹性供应链, 多级中断, 需求波动, 顾客类型