Bilevel joint optimisation for product family architecting considering make-or-buy decisions
针对产品族架构中制造商与供应商的决策冲突，提出将产品族架构与自制或外购决策建模为斯塔克尔伯格博弈，并用嵌套遗传算法求解双层优化模型，案例验证了可行性。
Product family architecting (PFA) aims at identification of common modules and selective modules to enable product family configuration for mass customisation. Due to nowadays manufacturers moving more towards assembly-to-order production throughout a distributed supply chain, the common practice of outsourcing of certain modules entails make-or-buy (MOB) decisions that must be taken into account in PFA. While the PFA and MOB decisions are enacted for different concerns of the manufacturer and the suppliers, it is important to deal with joint optimisation of the PFA and MOB problems. The prevailing decision models for joint optimisation are mainly originated from an ‘all-in-one’ approach that assumes both PFA and MOB decisions can be integrated into one single-level optimisation problem. Such an assumption neglects the complex trade-offs underlying two different decision-making problems and fails to reveal the inherent coupling of PFA and MOB decisions. This paper proposes to formulate joint optimisation of the PFA and MOB problems as a Stackelberg game, in which a bilevel decision mechanism model is deployed to reveal the inherent coupling and hierarchical relationships between PFA and MOB decisions. A nonlinear bilevel optimisation model is developed with the PFA problem acting as the leader and each MOB problem performing as a follower. A nested genetic algorithm is developed to solve the bilevel optimisation model. A case study of power transformer PFA subject to MOB considerations is presented to illustrate the feasibility and effectiveness of bilevel joint optimisation.