Supplier Selection and Order Allocation for Assembly Products using Multi-Stage Stochastic Bi-Objective Approach

This study addresses supplier selection and order allocation (SS&OA) strategy for multi-component assembly products considering demand uncertainties and disruption risks. We propose a novel multi-stage stochastic bi-objective mixed-integer linear programming (MILP) model leveraging the augmented- $\varepsilon$-constraint 2 method (AUGMECON2) to optimize total cost and purchasing value. The model incorporates disruptions, allows flexible order fulfilment, and integrates a comprehensive supplier evaluation process considering qualitative and quantitative criteria. Furthermore, an optimisation-based simulation procedure, based on a rolling planning horizon framework, is used to approximate the solution. The findings reveal diminishing returns on purchasing value with increasing costs. The optimal sourcing strategy depends on minimum order quantity (MOQ). Multiple sourcing excels under MOQ due to demand responsiveness, while dual sourcing offers superior cost efficiency without MOQ. Additionally, multiple sourcing outperforms in managing inventory levels. Sensitivity analysis confirms these trends and the impact of Bill-of-Materials complexity, supplier count, and demand variability on cost and service level. Finally, this research provides valuable insights for manufacturers in dynamic environments, enabling them to optimize SS&OA decisions and achieve improved supply chain performance.