Multi-objective robust optimization for facility location problem of personalized medicine supply chain

The network design of a personalized medicine supply chain (PMSC) is more complex than that of the traditional medical supply chain due to its special characteristics, including shorter shelf life, personalized manufacturing, etc. Despite growing interest in PMSCs, the integration of multi-objective optimization under uncertainties is often overlooked. In this work, a multi-objective multi-period optimization framework under uncertainties is proposed to address the facility location problem of PMSCs considering key characteristics, like failure rate and production mode, with waiting time, cost and coverage as objectives. To deal with the uncertainties in transportation time and operating costs, a robust optimization method is adopted, with illustrative examples used to validate it. Then, two tailored algorithms are developed for large-scale cases, which facilitate computation and obtain better solutions than other baseline algorithms. The results show the trade-off between coverage, waiting time and cost. Lower waiting time can lead to a greater reduction in cost at the same coverage. Under the influence of uncertainty, conservative decisions made by decision makers can result in a greater fluctuation in waiting time than other objectives. The findings assist healthcare policy makers and PMSC managers to deal with multiple objectives and uncertainties in practice.