Sequencing mixed-model assembly lines with risk-averse stochastic mixed-integer programming
研究了大型车辆制造商中因零件交付和质量不可靠导致的排序问题，用随机优化找到提升准时性能的策略，并针对不同风险偏好提出规避最差情景的规划方法。
Sequencing decisions in mixed-model assembly lines are complicated by various uncertainty factors. This paper addresses a real-life uncertainty factor identified in a manufacturer of large vehicles, by modelling unreliable part delivery and quality. Stochastic optimisation is applied to find sequencing policies that improve the on-time performance of its mixed-model assembly lines. As schedulers have different levels of risk aversion, a risk-averse programme is further presented to protect against the decision maker’s chosen fraction of worst scenarios. Computational studies with Progressive Hedging as the solution method, and its lower bounding approach, demonstrate the high quality of resulting sequencing decisions and the time efficiency of the solution method.