Bad-scenario Set Based Risk-resisting Robust Scheduling Model
基于坏场景集的抗风险鲁棒调度模型

We discuss robust scheduling models under uncertain environments described by scenario approach. Using the insights revealed by the analysis of traditional uncertain scheduling models involving the conflicting and balancing twofold relevance, which are the motivation of pursuing better performance and the conservatism of resisting risk, we establish a kind of new robust scheduling model. The optimization objective combines expected performance and robustness measure with a balance factor. A risk-resisting robustness measure is defined based on the concept of bad-scenario set, in which the number of bad scenarios can be adjusted by a standard performance. Thus, a set of robust scheduling models is established as the balance factor or the standard performance varies. A series of theorems reveal the relationship among the set of new models proposed in this paper and traditional uncertain scheduling models. And the condition of effectiveness of robustness for the set of new models is proposed as a theorem. Furthermore, an extensive experiment was conducted for job-shop scheduling problems with uncertain processing time. The computational results provide evidence that the set of new models is more comprehensive and more integrated in terms of pursuing better statistic performance and resisting the risk of performance deterioration. Thus, the new model can realize better balance between expected performance and risk-resisting robustness, as comparied against existing uncertain scheduling models.